NAEMSP ABSTRACTS

NAEMSP ABSTRACTS

ABSTRACTS FOR THE 2013 NAEMSP SCIENTIFIC ASSEMBLY

These are the abstracts for the National Asso-ciation of EMS Physicians Scientific Assembly,Bonita Springs, Florida, January 10–12, 2013.Key words: EMS; out-of-hospital; emergencymedical services; abstracts; National Associa-tion of EMS Physicians

1. SYSTEMWIDE REGIONALIZATION OF EMS ANDHOSPITAL CARE FOR OUT-OF-HOSPITALCARDIAC ARREST: ASSOCIATION WITH IMPROVEDSURVIVAL AND NEUROLOGIC OUTCOMES

Daniel Spaite, Bentley Bobrow, Uwe Stolz,Vatsal Chikani, Will Humble, Terry Mullins,Margaret Mullins, J. Steven Stapczynski,Karl Kern, Gordon Ewy, Arizona EmergencyMedicine Research Center

Background. Some specialized postarrest in-terventions for out-of-hospital cardiac arrest(OHCA) have been shown to improve out-comes. We evaluated the impact of imple-menting a voluntary, comprehensive, statewidesystem of designated cardiac receiving cen-ters (CRCs) and prehospital (emergency med-ical services [EMS]) triage/bypass protocols onOHCA outcomes. We previously reported pre-liminary results, but this report includes hospi-tals from across the state, much larger numbers,and neurologic outcomes. Methods. In Decem-ber 2007, the Arizona Department of HealthServices initiated a program of designating hos-pitals as CRCs based on ability to provideAmerican Heart Association (AHA) Guidelinepostarrest therapy and 24-hours-a-day/seven-days-a-week (24/7) percutaneous coronary in-terventions. In addition, the State EMS Coun-cil approved protocols allowing EMS to bypasslocal hospitals to take patients with return ofspontaneous circulation (ROSC) to CRCs if theexpected increase in transport interval was <15minutes. Design: Prospective before/after ob-servational study comparing OHCA patientsadmitted to CRCs during the six months priorto implementation (Phase 1 [P1]) with thosebrought after (P2). Outcomes: Survival to hos-pital discharge (cohorts: all rhythms, shockablerhythms) and neurologic status (good = cere-bral performance category of 1 or 2). Statistics:Fisher’s exact test, multiple logistic regression,α < 0.05. Results. Over 150 EMS agencies par-ticipated and 31 hospitals were designated asCRCs (December 2007 to December 2010; serv-ing ∼80% of Arizona’s population). Four hun-dred forty OHCA patients (age ≥18 years) weretransported to CRCs in P1 and 1,734 in P2. All-rhythm survival increased from 8.9% to 14.4%(adjusted odds ratio [aOR] = 2.14 [95% confi-dence interval (CI): 1.43, 3.21; p < 0.01]). Sur-vival with CPC 1/2 increased from 5.9% to

PREHOSPITAL EMERGENCY CARE 2013;17:103–134

doi: 10.3109/10903127.2012.733061

8.9% (aOR 2.12; 1.30, 3.45; p < 0.01). For shock-able rhythms, survival increased from 19.1%to 33.0% (aOR 2.40; 1.48, 3.90; p < 0.01) andCPC 1/2 increased from 15.0% to 22.8% (aOR2.12; 1.22, 3.69; p < 0.01). Conclusions. Im-plementation of a statewide system of CRCsand EMS bypass was independently associ-ated with significant increases in rates of over-all survival (61.8% relative increase) and goodneurologic status (+50.8%). In addition, sur-vival among those with shockable rhythms im-proved (+72.8%), as did their rate of good neu-rologic outcome (+52.0%). Implementation ofEMS bypass protocols and regionalization ofpostarrest care across a vast demography is fea-sible and, in this study, was associated withdramatic improvements in survival and neuro-logic outcome.

2. WHAT IS THE OPTIMAL CHEST COMPRESSIONDEPTH DURING OUT-OF-HOSPITAL CARDIACARREST RESUSCITATION OF ADULT PATIENTS?

Ian Stiell, Siobhan Brown, Clifton Callaway,Graham Nichol, Tom Aufderheide, SheldonCheskes, Christian Vaillancourt, DavidHostler, Daniel Davis, Ahamed Idris, JamesChristenson, Laurie Morrison, John Stouffer,Cliff Free, University of Ottawa

Background. The 2010 American Heart Associ-ation/International Liaison Committee on Re-suscitation (AHA/ILCOR) recommendationssuggested an increase in cardiopulmonaryresuscitation (CPR) compression depth foradults, with a target >50 mm and no up-per limit. This target is based on limited ev-idence and, hence, we sought to determinethe optimal compression depth range for adultpatients. Methods. We studied emergencymedical services–treated out-of-hospital car-diac arrest patients from the Resuscitation Out-comes Consortium (ROC) PRIMED clinical trialand Epistry–Cardiac Arrest database for whomelectronic CPR compression depth data wereavailable, from June 2007 to December 2010. Wecalculated anterior chest wall depression in mil-limeters for each minute of CPR. We controlledfor 10 covariates, including compression rateand calculated adjusted odds ratios for survivalto hospital discharge, 24-hour survival, andany return of spontaneous circulation (ROSC).Smoothing splines were used to explore the re-lationship between average compression depthand outcome. Results. We included 9,142 adultpatients from nine U.S. and Canadian citieswith these characteristics: mean age 67.5 years;male 64%; bystander witnessed 44%; bystanderCPR 42%; initial rhythms ventricular fibril-lation/ventricular tachycardia 24%, pulselesselectrical activity 20%, asystole 49%, and othernonshockable rhythm 6%; and outcomes ROSC31.3%, one-day survival 22.8%, and survival tohospital discharge 7.3%. For all patients, themean compression rate was 108 per minute;the mean compression fraction was 0.68; and

the mean compression depth was 41.9 mm,with the following ranges: <38 mm 37%,38–51 mm 45%, and >51 mm 18%. The adjustedodds ratios for survival to discharge, withdepth >51 mm as reference, were <38 mm, 0.69(95% confidence interval [CI] 0.53, 0.90) and38–51 mm, 1.03 (0.81, 1.30). Results were simi-lar for the intermediate outcomes of ROSC andone-day survival. Covariate-adjusted splinecurves revealed that the maximum survivalwas associated with a depth of 45.8 mm, fol-lowed by a decline in survival by 50 mm (op-timal interval 44–49 mm). We also found nodifferences in the spline curves between malesand females. Conclusions. This study foundthat more than one-third of patients receivedvery low compression depth. The optimal CPRcompression depth for survival appears to be46 mm (44–49) for both male and female adults,but falls off after 50 mm. These findings con-flict with the 2010 international guideline rec-ommendations.

3. OUTCOMES OF MILD THERAPEUTICHYPOTHERMIA IN OUT-OF-HOSPITAL CARDIACARREST: A NATIONWIDE RETROSPECTIVEANALYSIS

Won Pyo Hong, Sang Do Shin, Eui Jung Lee,Young Sun Ro, Joo Yeong Kim, Chang BaePark, Kyoung Jun Song, Dae Han Wi, SeoulNational University College of Medicine

Background. Although heavily emphasized,the effect of therapeutic hypothermia (TH)has been evaluated only in well-controlledhospital-based trials. This study aimed to de-termine whether mild TH was associated withimproved outcomes in out-of-hospital cardiacarrest (OHCA) on a nationwide scale. Meth-ods. We used a national OHCA cohort databasefrom January 2008 to December 2010; we in-cluded all emergency medical services (EMS)-treated adult patients with OHCA of presumedcardiac etiology. Multivariable logistic regres-sion analysis was used to determine the ef-fect of TH. Subgroup analyses were performedregarding initial presenting electrocardiogramrhythm (shockable vs. nonshockable). We ex-tracted propensity-matched samples to controlfor selection bias, and the propensity-matchedcohort was also analyzed by multivariablelogistic regression. Results. Among 64,155EMS-assessed OHCA patients with availableoutcome data, 44,794 (69.8%) were adults withpresumably cardiac etiology; 4,557 (10.2%) ofthese patients survived to admission and wereselected for the final analysis. The survival-to-discharge rate was higher in the TH groupthan in the non-TH group (53.6% vs. 33.4%,p < 0.001). A good neurologic outcome wasalso higher in the TH group than the non-TH group (18.3% vs. 10.3%, p < 0.001). Theadjusted odds ratios (ORs) of the TH ver-sus the non-TH group were 1.78 (95% confi-dence interval [CI] 1.41–2.26) for survival to

103

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.

104 PREHOSPITAL EMERGENCY CARE JANUARY/MARCH 2013 VOLUME 17 / NUMBER 1

discharge and 1.31 (95% CI 0.94–1.82) for goodneurologic outcome. In subgroup analysis,the adjusted ORs of TH for survival to dis-charge were 2.00 (95% CI 1.13–3.55) and 1.60(95% CI 1.24-2.07) for shockable rhythm andnonshockable rhythm, respectively. The ad-justed ORs of TH for good neurologic out-come were 1.81 (95% CI 1.04–3.13) and 0.89(95% CI 0.59–1.33), respectively. In the propen-sity score–matched cohort, the adjusted ORsof hypothermia were 1.83 (95% CI 1.33–2.50)for survival to discharge and 1.20 (95% CI0.76–1.88) for good neurologic outcome. Con-clusions. Mild therapeutic hypothermia im-proved survival to discharge and neurologicoutcomes on a nationwide scale. The effectwas increased when the initial rhythm wasshockable.

4. DISPATCHER-ASSISTED BYSTANDERCARDIOPULMONARY RESUSCITATION IN AMETROPOLITAN CITY: A BEFORE-AND-AFTERPOPULATION STUDY

Sang Do Shin, Joo Yeong Kim, Chang BaePark, Kyoung Jun Song, Seoul NationalUniversity Hospital

Background. The goal of this study wasto determine the effect of dispatcher-assistedbystander cardiopulmonary resuscitation (D-CPR) on the outcomes of out-of-hospital car-diac arrest (OHCA). Methods. This study wasperformed in an emergency medical services(EMS) system with a single-tiered basic to in-termediate service level and with about 65 des-tination hospitals in a metropolitan city with apopulation of 10 million. All EMS runs are dis-patched by a single centralized and physician-supervised center. An OHCA database includ-ing outcomes and demographic, Utstein, EMS,and hospital factors was populated from a dis-patch center registry and EMS run sheets, andfollowed by medical record review for outcomesurvey during 2009 to 2011. Cases with un-known outcome, patient age less than 15 years,and noncardiac causes were excluded. The in-tervention was the novel D-CPR method (inthe 2010 American Heart Association guide-line), which was implemented in January 2011.The primary and secondary end points weresurvival to discharge and good neurologic out-come (cerebral performance category 1 to 2).Adjusted odds ratios (ORs) and 95% confidenceintervals (95% CIs) for estimating the effect sizeof the D-CPR period compared with the pre-vious two years were calculated for outcomes,adjusting for potential predictors (age, gen-der, witness, response time, transport time, ini-tial electrocardiogram, and emergency depart-ment level [1 to 4]). Results. There were 8,143EMS-assessed adult OHCAs with presumedcardiac cause. Of these, bystander CPR wasperformed for 5.7% of the patients (148/2,600)in 2009, 6.7% (190/2,857) in 2010, and 37.8%(1,016/2,686) in 2011 (p < 0.001). The survivalto discharge was 7.1% in 2009, 7.1% in 2010,and 9.5% in 2011 (p = 0.001). Good neuro-logic outcome was 2.1% in 2009, 2.0% in 2010,and 3.6% in 2011 (p < 0.001). The adjustedORs (95% CIs) for survival to discharge for2011 and 2010 compared with 2009 were 1.53(1.24–1.90) in 2011 and 1.16 (0.93–1.45) in 2010.The adjusted ORs (95% CIs) for good neu-rologic outcome for 2011 and 2010 comparedwith 2009 were 2.12 (1.47–3.05) in 2011 and1.21 (0.81–1.80) in 2010. Conclusions. In thismetropolitan before-and-after study, an EMSintervention of a D-CPR protocol showed sig-nificant increase of bystander CPR and re-sulted in improved survival and neurologicoutcome.

5. COMPARISON OF SUCCESS RATES BETWEENTWO VIDEO LARYNGOSCOPE SYSTEMS USED INA PREHOSPITAL CLINICAL TRIAL

Aaron Burnett, Ralph Frascone, SandiWewerka, Samantha Kealey, Zabrina Evens,Kent Griffith, Joshua Salzman, RegionsHospital EMS

Background. To date, no clinical trial has com-pared the paramedic placement success ratesof two different video laryngoscope (VL) sys-tems in a nonsimulation setting. Methods. Thisinstitutional review board–approved, multia-gency, prospective, prehospital, nonrandom-ized, crossover trial compared the successrates and complications for two VL systems(Storz CMAC R©, Macintosh #4 blade; KingVISIONTM, Size 3). The providers completedinitial didactic and hands-on training, weregrouped by agency into two treatment armsbased on call volume, and were randomly as-signed the initial treating VL system. Patientinclusion criteria were: 1) ≥18 years old, 2) re-quired advanced airway management per stan-dardized patient care guidelines, and 3) treat-ing provider was trained in the use of VL. Pa-tients with supraglottic airway placement priorto arrival were excluded. After six months oftelephone data collection (available 24 hoursa day/seven days a week) following a place-ment attempt, the VL systems were crossedover for the remaining four months. Patient(age, gender, body mass index [BMI], race, eth-nicity, primary impression, call type, and diffi-cult airway) and provider (age, gender, yearsof experience, >1 VL placement during thestudy period, agency, and phase) demograph-ics were compared between groups. Overallsuccess rates (successful placement/number ofpatients), success rates by attempt (success-ful placement/total number of attempts), andcomplications were compared between treat-ment groups using exact logistic regression.The Cormack-Lehane (CL) score for each at-tempt was also analyzed for impact on successrates using a generalized linear mixed-effectsmodel. Results. Between October 2011 and Au-gust 2012, a total of 97 patients were treatedby 55 of 186 (30%) trained providers. Therewere no differences in patient or provider de-mographics between groups. The overall suc-cess rate was significantly higher for the CMACgroup (79% vs. 53%, odds ratio [OR] = 3.39;p = 0.008), as was the success by attempt rate(58.4% vs. 32.8%, OR = 2.88; p = 0.004). Theprovider-reported rates and absolute numbersof complications were similar between treat-ment groups (p = 0.16; p = 0.12). The mostfrequent complication reported was vomitingduring insertion (7.2%). Higher CL scores pre-dicted lower odds of success (OR = 0.379,p < 0.0001). Conclusions. Airway manage-ment with the CMAC resulted in a significantlyhigher success rate, although the rate is similarto previously published ETI success rates.6. EVALUATION OF THE IMPLEMENTATION OF THEREVISED ACUTE STROKE MEDICAL REDIRECTPARAMEDIC PROTOCOL (ASMRPP) IN URBANAND RURAL SETTINGS

Ian Stiell, Kristy Smaggus, CatherineClement, Michael Sharma, Doug Socha,Marco Sivilotti, Albert Jin, Jeffrey Perry, JimLumsden, Cally Martin, Mark Froats, RichardDionne, John Trickett, University of Ottawa

Background. The regional Acute Stroke Med-ical Redirect Paramedic Protocol (ASMRPP)stroke-redirect guidelines were recently revisedto allow emergency medical services (EMS)to bypass the nearest hospital for designatedstroke centers if total transport time would

be <2 hours and total time from symptomonset <3.5 hours. We sought to evaluate theimpact and effectiveness of implementing therevised ASMRPP within a large urban and ru-ral region. Methods. We conducted a 12-monthmulticenter, prospective cohort study involvingall prehospital patients presenting with possi-ble acute stroke. Participating were 1,000 ba-sic life support and 300 advanced life supportparamedics of nine land EMS agencies, oper-ating in a catchment area of 10 rural coun-ties and five cities (total population of 1.7 mil-lion, total area 15,000 square miles), with 22acute care hospitals and two university hos-pital stroke centers. Paramedics completed arecord form for each case and, initially, a secondparamedic independently completed the form.Outcomes and data analyses included redirectsensitivity and specificity, patient outcomes,adverse events, interrater reliability with thekappa statistic, and EMS and hospital impact.Results. We enrolled 987 eligible patients: male50.7%, mean age 73.8 years (range 16–101), metredirect criteria 52.9%, mean total prehospitaltime 43.2 minutes (range 14–165), and prehos-pital adverse events 12.6%. Of the 503 whomet redirect criteria and were transported to astroke center, 76.7% had a stroke code activated,51.9% had a final diagnosis of stroke, 21.9% re-ceived thrombolysis, 7.2% had adverse eventsin the emergency department (ED), 2.6% wererepatriated to their local hospital from the ED,and 89.7% survived to discharge. Their meanNational Institutes of Health (NIH) Stroke Scalescore was 8.2. For all 562 patients transportedto a stroke center, the paramedics were 98.0%sensitive and 30.4% specific for stroke code ac-tivation in the ED. The paramedics had ex-cellent interrater agreement, with kappa val-ues ranging from 0.59 to 0.90 for redirect cri-teria, from 0.84 to 1.0 for contraindication cri-teria, and 0.94 for need to transport to a strokecenter. Conclusions. In this large urban/ruralassessment of the ASMRPP stroke-redirectguidelines, paramedics were highly accurate inidentifying patients who needed urgent strokecenter transport and showed excellent inter-rater agreement for criteria. This protocol bene-fits patients without unduly burdening EMS orhospitals.

7. MODIFIED RANKIN SCORE AT HOSPITALDISCHARGE IS PREDICTIVE OF ONE-YEARNEUROLOGIC FUNCTION IN SURVIVORS AFTERCARDIAC ARREST

Marvin Wayne, David Tupper, TomAufderheide, Brian Mahoney, Robert Swor,Robert Domeier, Michael Olinger, RichardHolcomb, Cindy Setum, Ralph Frascone,Whatcom County EMS

Background. A randomized multicenterclinical trial compared standard cardiopul-monary resuscitation (S-CPR) with activecompression–decompression CPR plus aninspiratory impedance threshold device(ACD+ITD) in patients with nontraumatic,out-of-hospital cardiac arrest (OHCA) from apresumed cardiac cause. Survival to hospitaldischarge (HD) with favorable neurologicfunction, defined as a modified Rankin score(MRS) ≤3, and one-year survival were greaterin the ACD+ITD group. We evaluated theconcordance of the primary end point, HD withMRS ≤3, with multiple secondary neurologicend points assessed at one year. Methods. Atotal of 813 patients were enrolled in the S-CPRgroup and 842 in the ACD+ITD group. TheMRS at HD, and the Cerebral PerformanceCategory Scale (CPC), Overall PerformanceCategory (OPC), Health Utilities Index (HUI),and Cognitive Abilities Screening Instrument

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.

NAEMSP 2013 ANNUAL MEETING ABSTRACTS 105

(CASI) assessments were based on responsesfrom consented survivors and known deathsat one year. Neurologic assessments wereadministered by research staff blinded to theCPR treatment. Survival data were availablefor 98% of subjects. Fisher’s exact test, Pearsonchi-square test, Mann-Whitney U test, andt-test for equality of means were used, asapplicable, for comparisons. All statisticaltests were two-sided, and p-values <0.05were regarded as significant. Results. TheMRS (≤3 vs. >3) neurologic assessment at HDwas highly predictive of whether a patientwould be alive with favorable neurologicfunction in both study groups, assessed usingthe CPC score (>3 vs. ≤3) at one year: 35 of37 (94.6%) subjects in the S-CPR group and50 of 54 (92.6%) subjects in the ACD+ITDgroup with MRS ≤3 at HD had CPC >3 atone year (98.0% observed agreement, kappa= 0.800, p < 0.001). Similar concordance wasalso shown with overall survival, OPC, HUI,and CASI. Conclusion. Neurologic status atthe time of HD, as measured by MRS, is highlypredictive of long-term neurologic function atone year. This is the first time that the MRS atthe time of HD has demonstrated the ability topredict long-term outcomes for patients withOHCA.

8. SURVIVAL FROM HOSPITAL DISCHARGE TO ONEYEAR AFTER OUT-OF-HOSPITAL CARDIACARREST: A COMPARISON OF STANDARD CPRVERSUS ACTIVE COMPRESSION–DECOMPRESSIONCPR PLUS AN IMPEDANCE THRESHOLD DEVICE

Robert Swor, Richard Holcomb, RalphFrascone, Brian Mahoney, Marvin Wayne,Robert Domeier, Michael Olinger, DavidTupper, Demetris Yannopoulos, TomAufderheide, William Beaumont MedicalCenter

Background. Little is known about the impactof the method of cardiopulmonary resuscita-tion (CPR) on long-term survival followingout-of-hospital cardiac arrest. A recent NationalInstitutes of Health–funded multicenterprospective randomized clinical trial, theResQTrial, compared standard CPR (S-CPR)versus active compression–decompressionCPR plus the use of an impedance thresholddevice (ACD+ITD). ACD+ITD was associatedwith a relative 53% increase in survival tohospital discharge with favorable neurologicfunction for subjects with a cardiac arrest ofpresumed cardiac etiology, compared with S-CPR. Using data from all patients randomizedto one of these two methods of CPR from thistrial, we tested the hypothesis that ACD+ITDwould improve the likelihood of survival fromthe time of hospital discharge to one year aftercardiac arrest. Methods. A total of 1,335 adultpatients with nontraumatic out-of-hospitalcardiac arrest were enrolled in the S-CPR groupand 1,403 in the ACD+ITD group; 134 patientsversus 165 patients, respectively, survived tohospital discharge. A Kaplan-Meier analysiswas performed for all patients known to bedischarged alive from the hospital. Data up toone year after the cardiac arrest were obtainedfrom patient records, patient interviews, andpublic records. Results. Fewer patients in theS-CPR group survived to hospital dischargewith modified Rankin score (MRS) ≤3 (5.7%vs. 7.9%, p = 0.03), but the demographiccharacteristics of survivors were similarbetween groups. Starting with 100% survivalat hospital discharge, survival decreased inboth groups over time, but more notably inthe S-CPR group. Six months after cardiacarrest, survival rates were 77% in the S-CPRgroup and 88% in the ACD+ITD group. After

365 days, survival rates were 72% in the S-CPRgroup and 83% in the ACD+ITD group (logbase rank p-value = 0.014). Conclusions. Thesurvivors to hospital discharge in the ResQTrialwho were treated with ACD+ITD CPR had anabsolute 11% greater likelihood of survivingto 365 days after cardiac arrest comparedwith the patients treated with S-CPR. Thesedata support the hypothesis that increasedperfusion during CPR, obtained with the useof ACD+ITD CPR, results in a significantlyhigher likelihood of long-term survival, re-gardless of etiology of the nontraumatic cardiacarrest.

9. ASSESSING THE ACCURACY OF COMPUTERECG INTERPRETATION FOR IDENTIFYING ACUTEMYOCARDIAL INFARCTION

Amy Kule, Stephanie Hang, Kelly Sawyer,Alfred Burris, Justin Trivax, Robert Swor,Oakland University William Beaumont Schoolof Medicine

Background. Computer interpretation of elec-trocardiograms (CI-ECG) has assisted in iden-tifying potential ST-segment elevation myocar-dial infarction (STEMI) to shorten triage timeand decrease door-to-balloon time for patientswith acute myocardial infarction (AMI). How-ever, literature evaluating computer algorithmaccuracy only compares CI-ECG with physi-cian interpretation for STEMI, which providesan imperfect “gold standard” for AMI diag-nosis. Our objective was to evaluate the per-formance characteristics of CI-ECG for coro-nary artery occlusion using angiography as thegold standard. Methods. We examined a ret-rospective cohort of CI-ECGs obtained fromemergency medical services (EMS)-transportedadult patients admitted to the catheterizationlaboratory for suspicion of AMI at a single,large, academic community emergency depart-ment. EMS ECGs from January 2006 to Febru-ary 2011 were reviewed and dichotomized aseither “CI-ECG+,” defined as “AMI suspected”by computer printout, or “CI-ECG–.” Patientdemographics and relevant time intervals, in-cluding time from EMS ECG until reperfu-sion (minutes), were assessed for impact oncomputer accuracy of diagnosis. The primaryoutcome of “definite AMI” was based on an-giographic evidence for acute coronary vesselocclusion or presence of thrombus in a cul-prit vessel and confirmed by an independentcardiologist. Sensitivity, specificity, and like-lihood ratios (LRs) were calculated as mea-sures to evaluate CI-ECG. Results. A total of173 patients were identified, of whom 54%were male, the mean age (range) was 65.7(35–94) years, and the mean time (n = 134)from ECG to reperfusion was 81.9 (standarddeviation ±25.6) minutes. Overall, 73 (42.2%)“CI-ECGs+” and 43 (24.9%) “CI-ECGs–” hadcoronary occlusion on angiography. Computeraccuracy did not differ between groups bygender or time interval from ECG to reper-fusion; however, those with CI-ECG+ hada trend toward younger age (mean differ-ence 4.1 years, 95% confidence interval [CI]0.08–8.06, p-value 0.046). Performance charac-teristics of CI-ECG for definite AMI revealedsensitivity 62.9% (95% CI 53.4–71.6); specificity36.8% (95% CI 24.8–50.7); positive likelihood ra-tio 0.99 (95% CI 0.78–1.27); and negative likeli-hood ratio 1.01 (95% CI 0.75–1.35). Conclusion.Using coronary angiography as a gold stan-dard, computer interpretation of ECGs poorlyidentifies coronary artery occlusion. Furtherwork is needed to understand the relativevalue of CI-ECG versus clinician interpreta-tion of ECG in the diagnosis of coronary arteryocclusion.

10. COMPARISON OF THE EFFECTS OF CHESTCOMPRESSION BOUT LENGTH ON CORONARYPERFUSION PRESSURE DURINGCARDIOPULMONARY RESUSCITATION

David Salcido, Joshua Reynolds, JamesMenegazzi, University of Pittsburgh School ofMedicine

Background. Periodic reduction in blood flowand pressure associated with interruptions inventilations during cardiopulmonary resusci-tation (CPR) may be detrimental. Continu-ous or prolonged delivery of chest compres-sions may maintain higher blood flow andpressure longer, offsetting the physiologic ef-fects of protocolized pauses. Objective. Toexamine the relationship between coronaryperfusion pressure (CPP) and duration of un-interrupted chest compressions in a controlledresuscitation model. Methods. Five juvenile,mixed-breed domestic swine (28.1 kg ± 5.4)were sedated with ketamine and xylazine, in-tubated, and ventilated mechanically. The an-imals were then anesthetized and paralyzed,and micromanometer-tipped catheters were in-troduced by cutdown through the right femoralartery and vein into the aorta and right atrium,respectively. Aortic and right atrial pressureswere recorded continuously, and the diastolicend-point pressure differential between the twowas taken as the CPP. Ventricular fibrillationwas induced with a 3-second 100-mA transtho-racic shock, followed by 8 minutes withouttreatment, and then 30:2 mechanical CPR (LU-CAS2, Jolife) at a rate of 100 per minute was ini-tiated. Epinephrine and vasopressin were ad-ministered after 2 minutes of CPR. At 1, 3,5, and 7 minutes of CPR, compressions wereallowed to go on for one bout of 60 con-tinuous chest compressions, followed by tworescue breaths. The CPP characteristics werecompared between each 60-compression boutand the immediate preceding bout. Character-istics included: CPP at compression 30, CPPat the final compression, maximum CPP, meanCPP, CPP difference between the first com-pression and compression 30, and CPP differ-ence between the first compression and the fi-nal compression. The CPP characteristics werecompared between 30- and 60-compressionbouts over time with repeated-measures anal-ysis of variance using an alpha of 0.05. Results.Mean CPP 30-compression values versus meanCPP 60-compression values over all time pointswere as follows: (6.5, 13.6, 36.1, 28.7) vs. (8.8,14.2, 34.1, 26.1). Mean CPP characteristics gen-erally increased over time relative to the firstminute of CPR (p < 0.001). The CPP charac-teristics did not differ overall between the 30-and 60-compression bouts. Conclusion. In acontrolled laboratory resuscitation model, dou-bling the number of uninterrupted compres-sions from 30 to 60 does not beneficially changeCPP characteristics.

11. FREQUENCY OF MANUSCRIPT PUBLICATIONFOLLOWING PRESENTATION OF EMS ABSTRACTSAT NATIONAL MEETINGS

Brian Clemency, Heather Lindstrom, StevenGurien, Berly Jaison, Jeffrey Thompson,University at Buffalo

Background. Specialized knowledge and a sci-entific body of literature are the foundationof emergency medical services’ (EMS’s) recog-nition as a subspecialty within emergencymedicine. EMS research is often presented atnational meetings and published in abstractform, but full publication occurs less frequently.The primary goal of our study was to deter-mine the rate at which EMS-related researchpresented at selected conferences went on to

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


publication. A secondary goal was the determi-nation of the time to manuscript publication.Methods. We conducted a retrospective re-view of published abstracts from the 2003–2005national meetings of the American Collegeof Emergency Physicians (ACEP), the Societyfor Academic Emergency Medicine (SAEM),the National Association of EMS Physicians(NAEMSP), the Association of Air MedicalServices (AAMS), and the National Associa-tion of EMS Educators (NAEMSE) to iden-tify EMS-related abstracts. We then searchedPubMed (www.pubmed.gov) using abstract ti-tle key words and authors’ names to de-termine if the study had been published ina PubMed-indexed journal in the time sincepresentation and abstract publication. Results.Abstracts for the five conferences were re-viewed for 2003–2005. Six abstracts were ex-cluded because of manuscript publication priorto presentation. Six hundred thirty-five EMS-related abstracts were identified. The totalnumber of EMS abstracts presented and thepercentage subsequently published as fullmanuscripts were: ACEP 128, 48.4%; AAMS66, 33.3%; NAEMSE 24, 16.7%; NAEMSP 282,42.9%; and SAEM 135, 53.3%. The overall rateof publication was 44.3%. The average time topublication was 22.2 months (standard devia-tion 16.5, range 0 to 94 months). Conclusion.Fewer than half of the EMS abstracts went on tomanuscript publication. This represents missedopportunities for the growth of EMS as a sub-specialty.

12. BIOMECHANICAL ANALYSIS OF SPINALIMMOBILIZATION DURING PREHOSPITALEXTRICATION: A PROOF-OF-CONCEPT STUDY

Mark Dixon, Joseph O’Halloran, NiamhCummins, University College Dublin

Background. In most countries, road traffic col-lisions (RTCs) are the main cause of cervicalspine injuries. There are several techniques inuse for spinal immobilization during prehospi-tal extrication; however, the evidence for theseis poor. The aim of this study was to estab-lish which rescue technique provides the min-imal deviation of the cervical spine from theneutral inline position during the extrication ofthe RTC patient using biomechanical analysistechniques. Methods. A simulated male patient(weight 80 kg, height 180 cm) was fitted with acervical collar and extricated from a preparedmotor vehicle with roof removed and standardemergency medical services safety measures inplace. A rescue crew of four firefighter firstresponders and two paramedics performedeight different extrication techniques. The pa-tient was marked with biomechanical sensorsin the midline and in two horizontal planes atthe level of the forehead and clavicles, respec-tively. Relative movement between the sen-sors was captured via 12 infrared high-speedmotion-analysis cameras recording at 200 Hz. Avirtual three-dimensional mathematical modelwas developed from the recorded movement.Results. Control measurements were takenfrom the patient during self-extrication underverbal instruction and movement was recordedof 4.194◦ left of midline (LOM) to 2.408◦ rightof midline (ROM), resulting in total movementof 6.602◦. In comparison, the minimum de-viation recorded during equipment-aided ex-trication (long spinal board and/or extrica-tion device) was movement of 3.365◦ LOMand 8.352◦ ROM, resulting in total movementof 11.717◦. The maximum deviation recordedduring equipment-aided extrication was move-ment of 1.588◦ LOM and 24.498◦ ROM, result-ing in total movement of 26.086◦. Conclusions.Standard extrication techniques cause up tofour times more cervical spine movement dur-ing extrication than controlled self-extrication.

This pilot demonstrates the need for furthermulticenter evaluation of current rescue tech-niques and the requirement to investigate theclinical significance of such movement.

13. DEGRADATION OF BENZODIAZEPINES AFTER120 DAYS OF EMS DEPLOYMENT

Jason McMullan, Elizabeth Jones, KurtDenninghoff, Daniel Spaite, Erin Zaleski,Robert Silbergleit, University of Cincinnati

Introduction. Emergency medical services(EMS) treatment of status epilepticus withbenzodiazepines improves outcomes, butit is unclear which benzodiazepine is bestsuited for use in the EMS environment. Thereis little evidence published regarding theheat stability of benzodiazepines deployedon active EMS units.This study’s objectivewas to describe the degradation of diazepam,lorazepam, and midazolam as a function oftemperature exposure and time over 120 daysof storage on active EMS units. Methods. Vialsof diazepam, lorazepam, and midazolam weredistributed to four active EMS units in eachof two EMS systems in the U.S. Southwestduring the summer of 2011. Medications wereplaced in study boxes that logged temperatureevery minute and were stored in EMS unitsper local agency policy. Two vials of eachdrug were removed from each box at 30-dayintervals and underwent high-performanceliquid chromatography in a central laboratoryto determine drug concentration, which wascompared with labeled concentration. Meankinetic temperature (MKT) exposure wasderived for each sample. Concentrationswere analyzed as means with 95% confidenceintervals (CIs), and groups were comparedwith repeated-measures analysis of variance(ANOVA). Results. One hundred ninety-twototal samples were collected (two samples perfour units per city at four time points for eachdrug). Diazepam and midazolam experiencedminimal degradation at each time point. At120 days, the mean relative concentration (95%CI) of diazepam was 97.0% (95.7–98.2%) andthat of midazolam was 99.0% (97.7–100.2%).Lorazepam experienced significant degrada-tion by 60 days (95.6%; 91.6–99.5%), with halfof all samples <95% of labeled concentration.The relative concentration of lorazepam was90.3% (85.2–95.4%) at 90 days and 86.5%(80.7–92.3%) at 120 days. The groups weredifferent at every time point (ANOVA p <0.007). The mean MKT was 30.2◦C (95% CI28.5–31.0◦C). Increasing MKT was associatedwith greater degradation of lorazepam, butnot midazolam or diazepam. Conclusions.Midazolam and diazepam show, respectively,no or little degradation related to time ortemperature over 120 days of deployment inthe field in active EMS units. Lorazepam expe-riences significant and progressive degradationwith time at 60, 90, and 120 days, and withincreasing MKT exposure.

14. THE INFLUENCE OF PREHOSPITALHYPOTENSION AND HYPOXIA ON OUTCOMES INPATIENTS WITH MAJOR TRAUMATIC BRAININJURY

Daniel Spaite, Vatsal Chikani, BentleyBobrow, Michael Sotelo, Bruce Barnhart,Kurt Denninghoff, Joshua Gaither, ChadViscusi, David Adelson, Duane Sherrill,David Hardin, Uwe Stolz, ArizonaEmergency Medicine Research Center

Background. For major traumatic brain injury(TBI), few studies have evaluated the impact ofprehospital hypotension and hypoxia on out-comes other than mortality, and even thesehave generally been small. We evaluated theimpact of prehospital hypotension/hypoxia

on multiple outcomes in TBI patients in astatewide trauma system. Methods. The Ari-zona State Trauma Registry contains EMS andtrauma center (TC) data from all trauma pa-tients transported by ∼300 emergency medicalservices (EMS) agencies to eight level I or II TCsin Arizona. Prehospital hypotension (bloodpressure [BP] <90 mmHg in adults/childrenaged ≥10 years) and/or hypoxia (oxygen satu-ration [O2 sat] <90%) and various TC outcomesin all moderate/severe TBI cases (Centers forDisease Control and Prevention [CDC] BarellMatrix Type 1) from January 1, 2007, to Decem-ber 31, 2011, were evaluated (exclusions: trans-fers; missing EMS O2 sat or BP data). Four co-horts were established by presence of hypoxiaand/or hypotension: neither, hypoxia only, hy-potension only, or both. We compared survival,TC length of stay (LOS), intensive care unit(ICU) LOS, TC charges ($), and final dispositionacross cohorts. Results. Of 12,475 cases meet-ing inclusion criteria, 4,757 (38.1%) were miss-ing O2 sat or BP data, leaving 7,718 cases (nei-ther, 87.4%; hypotension only, 3.6%; hypoxiaonly, 6.0%; both, 3.0%). The median patient agewas 44 years (interquartile range [IQR]: 27, 60;70.2% male). Mortality was as follows: overall,12.6%; neither, 7.1%; hypotension only, 24.6%;hypoxia only, 32.5%; both, 82.6% (p < 0.00, allcomparisons). The adjusted odds ratios (aORs)for death (reference = neither) were as follows:hypotension only, 2.32 (95% confidence interval[CI]: 1.66, 3.23); hypoxia only, 3.02 (2.36, 3.87);both, 29.5 (19.7, 44.2). The LOS (median; IQR)was as follows: neither, 4 days (2; 9); hypoten-sion only, 8 (4; 19); hypoxia only, 12 (5; 20);both, 13 (8; 18); p < 0.0001. The ICU LOS wasas follows: neither, 2 days (1; 5); hypotensiononly, 4 (2; 13); hypoxia only, 9 (3; 14); both, 10(4; 16); p < 0.0001. Charges were as follows:neither, $50,370 (27,645; 110,480); hypoten-sion only, $125,558 (54,695; 306,506); hypoxiaonly, $171,831 (65,179; 285,103); both, $182,659(59,749; 369,332); p < 0.0001. Patients with hy-potension and/or hypoxia were much morelikely to be discharged to rehabilitation/long-term care than those who had neither (aOR =1.90; 95% CI 1.56, 2.32). Conclusion. In thisstatewide, multisystem analysis, prehospitalhypotension/hypoxia had a profound impacton mortality, even after controlling for in-jury severity, age, and prehospital intubation.Hypotensive/hypoxic patients also had muchhigher LOS, ICU LOS, and inpatient chargesand were much more likely to be dischargedto long-term care. Implementation of the EMSTBI treatment guidelines targeting these issuesis likely to have a major impact on outcomes.

15. EVALUATION OF END-TIDAL CARBONDIOXIDE LEVELS BEFORE, DURING, AND AFTERRETURN OF SPONTANEOUS CIRCULATION INOUT-OF-HOSPITAL CARDIAC ARREST

Ryan Murphy, Daniel Spaite, Uwe Stolz,Gary Smith, Annemarie Silver, MadalynKaramooz, Margaret Mullins, John Tobin,Bentley Bobrow, University of ArizonaCollege of Medicine

Background. In out-of-hospital cardiac arrest(OHCA), little is known about the changes incirculatory parameters that occur during thetransition from ongoing cardiopulmonary re-suscitation (CPR) to return of spontaneous cir-culation (ROSC). Previous work has showngood correlation between chest compression(CC) metrics (rate, depth, recoil, fraction) andend-tidal carbon dioxide (ETCO2) levels dueto associated variations in blood flow. Thus,changes in ETCO2 levels occurring aroundROSC may be clinically important. Methods.Data from an Utstein-compliant registry alongwith electronic CC metrics and ETCO2 datawere collected by two emergency medical

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


services (EMS) agencies on consecutive in-tubated, adult, nontraumatic OHCA patientswho achieved ROSC (October 8, 2008, toSeptember 9, 2011). CC quality/event datawere collected using accelerometer-equippeddefibrillators (E Series, ZOLL Medical). CCmetrics and ETCO2 data were compiled foreach minute. ROSC was determined frompatient care reports and defibrillator-verifiedCC cessation. Statistics: one-way analysis ofvariance, Kruskal-Wallis, α = 0.05. Results.Twenty-eight cases of ROSC had availabledata (age 64 ± 16 years, 64% male). Initialrhythm: ventricular fibrillation, 14; asystole, 8;and pulseless electrical activity, 6. Seven pa-tients (25%) survived to discharge. The medianETCO2 level increased gradually before ROSC(p = 0.02; –4 minutes = 22.2 mmHg [interquar-tile range (IQR): 19.7–37.5], –3 minutes = 23.6[16.9–37.7], –2 minutes = 33.5 [17.1–48.4], –1minute = 36.2 [22.0–54.0]) to 43.3 mmHg inthe minute of ROSC [30.3–99.0] and stayed rel-atively constant (p = 0.99) after ROSC (+1minute = 46.1 mmHg [30.0–64.0], +2 minutes =41.6 [20.9–51.6], +3 minutes = 43.2 [15.6–48.4],+4 minutes = 42.6 [13.2–47.3]). There were noclinically relevant changes in CC or respira-tory metrics preceding ROSC. A subset of pa-tients with ETCO2 <30 mmHg at ROSC (n =4) did not show increased ETCO2 (p = 0.9)in the minutes preceding ROSC (–4 minutes= 20.0 mmHg [19.7–20.2], –3 minutes = 20.6[16.3–37.7], –2 minutes = 17.1 [7.9–35.9], –1minute = 20.2 [10.6–27.7], minute of ROSC =17.9 mmHg [9.5–26.5]), and ETCO2 stayed loweven after EMS-declared ROSC (+1 minute =20.6 [14.2–22.2], +2 minutes = 20.7 [13.1–23.2],+3 minutes = 15.6 [15.1–24.1], +4 minutes =13.2 [13.2–13.2]; p = 0.97). In this low-ETCO2cohort, three of four patients (75%) experi-enced rearrest within 2.5 minutes of ROSC,whereas only one of 24 (4%) with ETCO2 >30mmHg experienced rearrest within 2.5 min-utes. Conclusion. Prior to ROSC, the ETCO2level increases significantly. Thus, consistentETCO2 increases during resuscitation may her-ald impending ROSC and warrant height-ened attentiveness by EMS providers. In thispreliminary study, ETCO2 level >30 mmHgwas associated with longer sustained ROSC.Future study should evaluate whether it is ap-propriate to continue CPR in ROSC patientswith low ETCO2 level, since flow may be min-imal and the risk of immediate rearrest may bevery high.

16. TRANSFER OF NONURGENT EMERGENCYMEDICAL SERVICES 9-1-1 CALLERS TO ATELEPHONE NURSE ASSESSMENT SERVICE(TNAS)

Ian Blanchard, Greg Vogelaar, Lois Andruski,Lara Osterreicher, Jane Huang, Jim Trumbley,Wadhah Almansoori, Tyler Williamson,Andrew Anton, Alberta Health Services EMSand the University of Calgary

Background. Emergency medical services(EMS) systems have been challenged tomaintain present service levels by factors suchas offload delay and increasing call volumes.Some EMS systems have tried to leverageexisting health care programs by transferringnonurgent EMS 9-1-1 callers to a telephonenurse assessment service (TNAS) insteadof EMS response. Objective. To assess thepotential impact to the quality of patient careof transferring low-priority 9-1-1 callers to aTNAS instead of EMS response. Methods.Prospective modified single-subject designon a sample of adult 9-1-1 callers in an urbancenter who received an Alpha or Omegaresponse as determined by the Medical PriorityDispatch System (MPDS). These patientsreceived routine EMS response, but were also

transferred to a TNAS for assessment prior toEMS arrival; the TNAS did not provide adviceto the patient. The actual disposition from theEMS system was compared with the theoreticaldisposition from the TNAS. Impact to thequality of care if EMS treatment had not beenprovided was assessed by modified nominalgroup technique to achieve expert panelconsensus. Results. A total of 405 patientsmet the inclusion criteria, the median agewas 61 years (interquartile range 39, 79), and56% were female. The TNAS identified that98 patients (24%; 95% confidence interval [CI]20%, 28%) required immediate EMS response,137 patients (34%; 95% CI 29%, 39%) did notrequire EMS but were advised to visit theemergency department (ED), and 170 patients(42%; 95% CI 37%, 47%) required neitherimmediate EMS response nor an ED visit. Theexpert panel reviewed the group of 170 patientsand concluded that in 14 (8%) cases, EMStreatment may have impacted mortality (n = 1)or morbidity (n = 13), and in 61 (36%) cases, itmay have impacted comfort. Quality assurancereview of the 14 mortality/morbidity cases re-vealed nine cases of errors in the MPDS/TNASassessments and six cases of additional patientinformation/condition changes during EMSassessment. Conclusion. In this sample oflow-priority 9-1-1 callers, transfer to a TNASinstead of EMS response may impact thequality of patient care by introducing a patientsafety risk and the potential for delayed com-fort measures. Further research determiningwho can be safely transferred to a TNAS andhow to identify these patients is required.

17. VOLUME SHIFTS SIGNIFICANTLY IMPACTCHEST COMPRESSION–GENERATED BLOOD FLOW

Joshua Lampe, Josiah Garcia, Tai Yin, GeorgeBratinov, Christopher Kaufman, LanceBecker,University of Pennsylvania

Background. The existence of a blood volumeshift during resuscitation has been a hypothet-ical explanation of the observed reduction inchest compression (CC) efficacy as a function oftime. However, central blood flows, and there-fore volumes, have not been thoroughly in-vestigated during prolonged cardiopulmonaryresuscitation (CPR). Methods. CPR hemody-namics in nine domestic swine (∼30 kg) werestudied using standard physiologic monitor-ing. Flow and pressure sensors were placedon the abdominal aorta (AA) and the inferiorvena cava (IVC) slightly inferior to the kid-neys. Ventricular fibrillation (VF) was electri-cally induced. Mechanical CCs were started af-ter 10 minutes of untreated VF and contin-ued for 54 minutes. Results. Hemodynamicdata indicate that the study animals separatedinto two groups depending on the directionof the final net IVC flow. At the start of me-chanical CC, there were no significant differ-ences in IVC and AA flows between the twogroups. At the end of resuscitation, the IVC andAA flows were significantly different betweenanimals, with net forward IVC flow (IVCpos;49.9 ± 13.9 mL/min) and net negative IVCflow (IVCneg; –51.14 ± 22.4 mL/min). Surpris-ingly, the IVCneg animals had higher forwardAA flows (8.0 ± 2.0 mL/min vs. –2.1 ± 1.8,p = 0.006). As a result, the IVCneg animalswere adding blood volume to the tissue be-low the flow probes, whereas the IVCpos an-imals were removing blood volume from thetissue below the flow probes at the end of re-suscitation. Conclusions. Both the IVCneg andthe IVCpos groups experienced significant vol-ume shifts during the resuscitation. However,the volume shifts appear to be in opposite di-rections. The volume shifts between these reser-voirs have a profound impact on the distribu-

tion of CC-generated blood flow. These obser-vations require further investigation into thehemodynamics of volume shifts during resus-citation.

18. DOES HEALTH STATUS INFLUENCE THEWILLINGNESS TO PROVIDE INFORMED CONSENT?RESULTS FROM A CARDIAC ARREST TRIALCONDUCTED UNDER WAIVER OF INFORMEDCONSENT

Ralph Frascone, Joshua Salzman, DemetrisYannopoulos, Brian Mahoney, Robert Swor,Robert Domeier, Marvin Wayne, TomAufderheide, Michael Olinger, SandiWewerka, David Tupper, Richard Holcomb,Regions Hospital EMS

Background. In a recent out-of-hospital cardiacarrest (OHCA) trial conducted under an ini-tial waiver of informed consent (IC) (21 § CFR50.24), data from public records were collectedthat shed light on whether the willingness toprovide subsequent IC was associated withstudy outcomes. We hypothesized that IC wasless likely to be obtained in subjects with signif-icantly compromised health status. Methods.A post hoc analysis was conducted using datafrom a National Institutes of Health–fundedrandomized, controlled OHCA clinical trialcomparing active compression–decompressioncardiopulmonary resuscitation (CPR) plus animpedance threshold device (ACD+ITD) withstandard CPR. The primary end point was sur-vival to hospital discharge (HD) with favorableneurologic function [Modified Rankin Scalescore (MRS) ≤3]. The status of the consent pro-cess was tabulated for all subjects who survivedto hospital admission. Unadjusted Fisher’s ex-act test and associated odds ratios were used tocompare the MRS at HD by IC status. Results.Among a total study population of 1,655 sub-jects, 457 survivors were admitted to the hos-pital, and 440 had known HD status: 320 gaveIC, 46 were unable to complete the IC process(the instititional review board allowed medicalrecord review), and 74 denied IC. Survival withan MRS ≤3 was significantly higher in subjectswhere IC was given: 35.0% vs. 4.1%, p < 0.001.Sixteen of the 17 cases with missing MRS out-comes were in the IC-denial group. Even if allwere considered to have favorable outcomes,the resulting rate (21.1%) was less than thatseen among the subjects with IC given (p =0.015). Conclusion. The subjects who denied ICwere significantly less likely to have favorableoutcomes. These findings suggest that some re-suscitation trials may unknowingly underrep-resent those subjects with the worst prognosesin a target study population despite prespeci-fied inclusion and exclusion criteria, because ofthe unwillingness or inability of the subjects ortheir families to provide IC.

19. CORRELATIONS BETWEEN CPR QUALITYMETRICS AND END-TIDAL CO2 INOUT-OF-HOSPITAL CARDIAC ARREST

Daniel Spaite, Uwe Stolz, Ryan Murphy,Madalyn Karamooz, Annemarie Silver, JohnTobin, Terry Mullins, Gordon Ewy, BentleyBobrow, Arizona Emergency MedicineResearch Center

Background. Limited data from human car-diopulmonary resuscitation (CPR) studies in-dicate a possible relationship between chestcompression (CC) quality metrics and end-tidalcarbon dioxide (ETCO2) levels due to associ-ated variations in blood flow. We evaluated thecorrelations between CC metrics and ETCO2during out-of-hospital cardiac arrest (OHCA).Methods. Data from an Utstein-compliant reg-istry along with electronic CC metrics andETCO2 data were collected on consecutive

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


intubated, adult, nontraumatic OHCA patientstreated by two emergency medical servicesagencies (September 2008 to September 2010).CC quality/event data were collected usingaccelerometer-equipped defibrillators (E Series,ZOLL Medical) and were reviewed using CodeReview software. CC metrics (rate, depth, re-coil, fraction) and mean ETCO2 data were com-piled for each minute (first minute excludedbecause of artificially high values) when pa-tients had no spontaneous circulation (absenceof documented return of spontaneous circu-lation [ROSC] or ETCO2 >50 mmHg). Multi-variable regression was used to quantify thecorrelations between CC quality metrics andETCO2. Results. Among 586 OHCA patients,416 had defibrillator files, of which 123 wereintubated patients with ETCO2 data duringCPR (study population). One hundred forty-two (34%) patients were not intubated; 112(27%) were intubated but had no capnographydata; 39 (9%) were intubated but had insuffi-cient ETCO2 data or ETCO2 only after ROSC.Mean CC depth, mean CC release velocity (“re-coil”), and CC fraction were each significantlyrelated to ETCO2 in univariate analysis (r2 =0.11, 0.09, 0.05, respectively, p < 0.05), while CCrate and length of preshock/postshock pausewere not. Ventilation rate was inversely relatedto ETCO2 (r2 = 0.10, –0.80 mmHg/breath/min[95% confidence interval: –1.23, –0.37], p <0.001) and CC depth and recoil were sig-nificantly related to ETCO2 in multivariableanalysis (controlling for bystander CPR andventilation rate). Adjusted coefficients wereas follows: CC depth: 7.46 mmHg/inch (3.16,11.76); recoil: 13.7 mmHg/500 milli-inches/sec(5.8, 21.7). Conclusion. The ETCO2 level is sig-nificantly correlated with several CC metricsduring CPR and might provide a meaningfulmeasure of blood flow generated by compres-sions. Intraresuscitation evaluation of ETCO2levels may be helpful in improving blood flowand holds promise for impacting outcomes.

20. A PILOT STUDY EVALUATING THE USE OFDOUBLE-SEQUENTIAL EXTERNALDEFIBRILLATION IN OUT-OF-HOSPITALREFRACTORY VENTRICULAR FIBRILLATION

Brent Myers, Jose Cabanas, Valerie De Maio,Ryan Lewis, Joseph Zalkin,WakeMed Health & Hospitals

Background. Ventricular fibrillation (VF) isconsidered the out-of-hospital cardiac arrest(OOHCA) rhythm with the highest likelihoodof neurologically intact survival. Unfortunately,there are occasions where VF does not re-spond to standard defibrillatory shocks. Cur-rent American Heart Association guidelinesacknowledge there are insufficient data to de-termine the optimal pad placement, waveform,or energy level that produces the best con-version rates from OOHCA with VF. The ob-jective of this study was to describe a tech-nique of double-sequential external defibrilla-tion (DSED) for cases of refractory VF dur-ing OOHCA resuscitation. Methods. A ret-rospective case series was performed in anurban/suburban emergency medical services(EMS) system with advanced life support care(population 900,000). Included were adult pa-tients with OOHCA having refractory VF dur-ing resuscitation efforts by EMS providers. Re-fractory VF was defined as persistent VF fol-lowing at least five unsuccessful single shockswith anterior-lateral pad placement and a doseof antiarrhythmic medication without change.Once in refractory VF, EMS personnel appliedpads with an anterior-posterior pad placementand utilized a second defibrillator to attemptsingle defibrillation with the new monitor/pad

placement. If VF continued, the EMS person-nel then utilized the original and second mon-itor/defibrillator (two sets of pads with twomonitors) charged to maximum energy, andshocks were delivered from both machines atthe same time. Data were collected from elec-tronic dispatch and patient care reports fordescriptive analysis. Results. From January 7,2008, to December 31, 2010, a total of 10 patientswere treated with DSED. The median age was76.5 years (interquartile range [IQR]: 65–82),with a median resuscitation time of 51 min-utes (IQR: 45–62). The median number of singleshocks was 6.5 (IQR: 6–11), with a median of 2(IQR: 1–3) DSED shocks delivered. VF broke af-ter DSED in seven cases (70%). Only three pa-tients (30%) had return of spontaneous circula-tion in the field and none survived to discharge.There were no adverse events or cardiac mon-itor failures reported. Conclusion. This pilotstudy demonstrates that the DSED technique isfeasible in the out-of-hospital setting. In this se-ries, refractory VF was terminated 70% of thetime, but no patient survived to discharge. Fur-ther research is needed to better understand thecharacteristics and treatment strategies of re-fractory VF.

21. AN EPIDEMIOLOGIC PROFILE OFRESUSCITATION SYSTEMS-OF-CAREPERFORMANCE IN OUT-OF-HOSPITAL CARDIACARREST

Zach Dewar, Andrew Travers, Jan Jensen,Alix Carter,Emergency Health Services, NovaScotia/Dalhousie University

Background. The study objective was to linkstructure, process, system, and outcome (SPSO)resuscitation measures from community, emer-gency medical services (EMS), and in-hospitalsystems of care (SOCs), to identify factors sig-nificantly associated with positive outcomes byexamining the system as a whole. Methods. Ina provincial EMS system responding to 920,000residents, SPSO data were collected from first-responder charts, EMS dispatch, electronic pa-tient care records, and hospital charts for all2011 out-of-hospital cardiac arrests (OOHCAs).Data were linked using deterministic linkage.Descriptive and chi-square analyses were per-formed on cases of Utstein-defined cardiac eti-ology. The primary outcome was survival todischarge. The secondary outcome was EMSperformance, measured by sustained return ofspontaneous circulation (ROSC) on emergencydepartment (ED) arrival. Factors for analy-sis were determined a priori by consensus,grounded in literature. Significance adjustedto p < 0.01 for multiple comparisons. Re-sults. EMS responded to 1,514 cases of possi-ble OOHCA in 2011. Excluded cases included:189 (12.5%) overtriage, no patient, missing data,duplicate entries, in-hospital cases, and interfa-cility transfers; 68 (4.5%) traumatic OOHCAs;714 (47%) no attempted resuscitation; and 27(1.8%) Utstein noncardiac etiology. Five hun-dred sixteen of 1,514 (34%) cardiac etiologycases were analyzed. No unlinkable recordswere present. The mean age was 66 years (stan-dard deviation = 17.5), and 353 were male(68.4%). In 222 of 516 (43%) cases, resuscitationwas terminated in the field; 294 of 516 patients(57%) were transported; 122 of 516 patients(23.6%) had sustained ROSC; 36 of 122 patients(29.5%) received ED targeted therapeutic hy-pothermia (ED TTH); 24 of 122 patients (19.7%)were transferred to higher-level care; and 32of 122 patients (26.2%) had percutaneous coro-nary interventions (PCIs). Of patients with sus-tained ROSC, 42 of 122 (34.4%) survived to dis-charge with a median cerebral performance cat-egory of 1 (90th percentile: 2). Forty-two of 516

(8.1%) survived to discharge. Factors signifi-cantly associated with survival included wit-nessed OOHCA (p = 0.005), first-responderCPR (p = 0.008), epinephrine (p < 0.001), in-tubation (p = 0.001), sustained ROSC (p <0.001), ED TTH (p < 0.001), transfer to higher-level care (p < 0.001), and PCI (p < 0.001).Limitations include collection from adminis-trative databases, limited first-responder data,and limited availability of postarrest measures.Conclusions. Factors from each SOC were as-sociated with survival. These findings haveimplications for prospective research, clinicalpractice, and policy to integrate SOCs to im-prove OOHCA outcomes.

22. DETERMINANTS OF VENTRICULARFIBRILLATION INCIDENCE AS FIRST-RECORDEDRHYTHM DURING OUT-OF-HOSPITAL CARDIACARREST AND ASSOCIATION WITH LONG-TERMNEUROLOGIC OUTCOMES

Demetris Yannopoulos, Richard Holcomb,Ralph Frascone, Brian Mahoney, MarvinWayne, Robert Swor, Robert Domeier,Michael Olinger, David Tupper, TomAufderheide,University of Minnesota Medical Center

Background. We sought to identify the fac-tors that were associated with higher in-cidence of ventricular fibrillation and sur-vival with good neurologic function in theResQTrial, which compared standard car-diopulmonary resuscitation (S-CPR) versus ac-tive compression–decompression CPR withan inspiratory impedance threshold device(ACD+ITD) in patients with out-of-hospitalcardiac arrest (OHCA). Methods. A retrospec-tive analysis of a randomized multicenter clin-ical study of 1,655 patients with OHCA. 88.3%(106/120) of the patients discharged with goodneurologic function (modified Rankin score[MRS] ≤3) had a first recorded rhythm of ven-tricular fibrillation/pulseless ventricular tachy-cardia (VF). The first rhythm was recorded in99.4% (1,645/1,655) of the cases about 9.5 min-utes after the 9-1-1 call, on average 3 minutesafter the arrival of EMS on the scene and afterCPR was performed for at least 2 minutes. Re-sults. A total of 32.8% of the patients had VF asthe presenting rhythm and 42.8% received by-stander CPR. The presence of bystander CPRwas associated with a higher VF incidence onlyin the S-CPR group (40.8% versus 23.1% withno bystander, p = 0.001), but survival was 7.6%versus 4.6%, p = 0.09. Presence or absence ofbystander CPR led to similar VF incidence andsurvival in the ACD+ITD group: 36.1% versus33.9% and 9.0% versus 8.9%, respectively, p >0.2. After propensity adjustment for witnessedarrest, age <67 years, gender, and public lo-cation, bystander CPR lost significance. In theabsence of bystander CPR, ACD+ITD signifi-cantly increased the incidence of first-recordedVF compared with S-CPR from 106 of 459 pa-tients (23.1%) to 164 of 484 patients (33.9%)(odds ratio [OR] 1.71, 95% confidence inter-val [CI] 1.27, 2.30, p < 0.001), and in patientswith VF, return of spontaneous circulation in-creased from 65 of 459 patients (14.2%) to 104of 484 patients (21.5%) (OR 1.66, 95% CI 1.16,2.37, p = 0.004), leading to an overall doublingof survival with MRS ≤3 from 21 of 455 pa-tients (4.6%) to 43 of 482 patients (8.9%) (OR2.02, 95% CI 1.15, 3.65, p = 0.009). After propen-sity adjustment, ACD+ITD remained a signifi-cant predictor of an MRS ≤3 (p = 0.02). Con-clusions. VF was the most important predic-tor of survival with MRS ≤3. In the absenceof bystander CPR, ACD+ITD increased VF in-cidence as the first-recorded rhythm and dou-bled survival to hospital discharge with MRS≤3 compared with S-CPR.

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


23. CAN PARAMEDICS DIAGNOSE SEPSIS IN THEPREHOSPITAL SETTING?: A FEASIBILITY STUDY

Andrew Travers, Robert Green, Ed Cain,Samuel Campbell, Jan Jensen, David Petrie,Dalhousie University

Background. Evidence demonstrates that sep-sis patients presenting to emergency depart-ments (EDs) benefit from early goal-directedtherapy (EGDT). Accurate paramedic diagno-sis is important to initiate EGDT promptly. Theobjective was to evaluate diagnostic perfor-mance of sepsis by paramedics, compared withemergency physician (EP) diagnosis. Methods.This prospective, observational study was ofa convenience sample of adult patients trans-ported to a tertiary ED by paramedics. Pa-tients were enrolled if dispatched as: abdomi-nal pain, breathing problems, sick person, un-known problem, unconscious/fainting, chestpain, or any case in which paramedics consid-ered sepsis a possible diagnosis. Patients werenot enrolled if they were interfacility trans-fers or had cardiac arrest, electrical shock, ortrauma. Paramedic diagnoses of sepsis wereentered into study data-collection forms, whichwere compared with blind, independent docu-mentation of admission diagnosis by attendingEPs (considered the “gold standard”). For miss-ing EP forms, ED chart review was conducted.Specificity, sensitivity, accuracy, positive andnegative predictive values, and ratios were cal-culated with 95% confidence intervals (CIs).Results. Nine hundred fifty-six patients wereenrolled by paramedics between January andSeptember 2008. Three hundred twenty-sevenof 956 (34.2%) were excluded for the follow-ing reasons: no paramedic diagnosis recorded(249/956, 26.0%), no EP diagnosis available(73/956, 7.6%), or patient left the ED withoutbeing seen/no EP diagnosis was made (5/956,0.5%). The paramedic and EP diagnoses wereavailable for 629 of 956 patients (65.8%), andwere included in the final analysis. Paramedicsidentified 170 of 629 (27.0%) patients as sep-tic, and EPs identified 71 of 629 (11.3%). Thesensitivity and specificity of paramedic sep-sis diagnosis were 73.24% (95% CI 61.40–83.05)and 78.85% (95% CI 75.23–82.17), respectively.The accuracy was 78% (492/629, 52 true posi-tive, 440 true negative). The positive and neg-ative predictive values were 30.59% (95% CI23.76–38.11) and 95.86% (95% CI 93.61–97.49),respectively. The positive likelihood ratio was3.46 (95% CI 2.80–4.29), and the negative like-lihood ratio was 0.34 (95% CI 0.23–0.50). Studylimitations included lack of prehospital use ofthermometers or point-of-source lactate test-ing and large proportion of incomplete EPforms, requiring retrospective chart review.Conclusion. Paramedic diagnosis of sepsis hasgreater specificity than sensitivity, with reason-able accuracy between paramedics and EPs.Further research should evaluate thermome-try and point-of-care lactate testing diagnostictools, derivation of a clinical prediction rule,and EGDT delivery in the prehospital setting.

24. EVALUATION OF A PREHOSPITAL SEPSISPROTOCOL

Jennifer Walker, Henderson McGinnis,Michael Halsey, Brian Hiestand,Wake Forest Baptist Medical Center

Background. Multiple studies show improvedoutcomes with early goal-directed therapy insepsis; however, most literature involves ini-tiation of therapy in hospital. The purposeof this study was to evaluate prehospitalprovider recognition of sepsis and the impacton treatment and patient outcomes. Methods.We performed a retrospective analysis of acritical care transport database involving all

adult interfacility transfers to a tertiary carecenter between September 2010 and Novem-ber 2011. We included patients who had sus-pected infection plus two or more systemic in-flammatory response syndrome (SIRS) criteria(temperature<36◦C or >38◦C, heart rate >90beats/min, respiratory rate >20 breaths/min orpartial pressure of carbon dioxide [PaCO2] <32mmHg, white blood cell [WBC] count <4,000 or>12,000); patients with clear noninfectious di-agnoses were excluded. Fisher’s exact test andthe Wilcoxon rank sum test were used to com-pare categorical and continuous data, respec-tively. Results. Of 717 patients transported dur-ing that time, 214 patients (29.8%) met the def-inition of sepsis. Of these, 106 (49.5%) weremale, 185 (86%) were white, and the mean agewas 57.6 years. Sepsis was documented by thetransport team in 59 cases (28%, 95% confidenceinterval [CI] 22–34%). Multivariate logistic re-gression models were prepared to determinecovariates that affected sepsis recognition. Inthe final model, the only significant factors af-fecting the ability to recognize sepsis were theinitial blood pressure (odds ratio [OR] 0.98, 95%CI 0.97–0.99) and the initiation of vasopressorsor inotropes at the outside hospital (OR 4.7,95% CI 2.2–10.3). Mean arterial pressure, heartrate, and systolic blood pressure improved dur-ing transport of both unrecognized and recog-nized septic patients; documented recognitionof sepsis did not make a significant differencein vital signs (p = 0.29, p = 0.19, and p = 0.14,respectively). If sepsis was recognized, how-ever, providers gave more fluid (p < 0.001),were more likely to start pressors (p < 0.001),and were less likely to turn off pressors startedby the outside hospital (p = 0.017). Discussion:Underrecognition of sepsis is present in this co-hort. More aggressive interventions occurredwhen sepsis was recognized; however, patientvital signs improved regardless of sepsis recog-nition. This may be confounded because of theretrospective nature of this analysis and theabsence of a true control group for compari-son. Conclusion. Patients with suspected sep-sis generally improve during interfacility trans-port in a system with a defined sepsis protocol,regardless of whether sepsis was explicitly rec-ognized by the providers.

25. BARRIERS TO IMPLEMENTATION OFRECOMMENDATIONS FOR TRANSPORT OFCHILDREN IN GROUND AMBULANCES

Rashida Woods, Cara Doughty, AnthonyGilchrest, Manish Shah, Rainbow Babies andChildren’s Hopsital

Background. Because of the risk of secondaryinjury in the event of an ambulance collision,the National Highway Traffic Safety Admin-istration (NHTSA) released draft recommen-dations in 2010 to define guidelines for safetransport of children in ground ambulances.Lack of awareness and other barriers may limitemergency medical services (EMS) agenciesfrom fully implementing these recommenda-tions. The purpose of this study was to assessawareness of the NHTSA guidelines amongEMS agencies in Texas and identify potentialbarriers in complying with them. Methods.This was a cross-sectional, online survey of asample of 9-1-1–responding ground transportEMS agencies in Texas. Using case-based sce-narios, the survey assessed each agency’s cur-rent transport methods with respect to the fivesituations defined in the NHTSA guidelines. Italso assessed each agency’s plans and potentialbarriers to implementation. Descriptive data re-porting was utilized. Results. Of the 160 EMSagencies contacted, 56 (35%) met the inclusioncriteria and completed the survey. Only 36%were aware that NHTSA issued the guidelines.

Agencies utilize ideal or acceptable transportmodalities 75% of the time when medical mon-itoring and/or interventions are required, and69% of the time when spinal immobilization isrequired. For children who are uninjured or notill, 93% of agencies use a mode of transport thatis not recommended by NHTSA. Also, 53% ofagencies transport ill and/or injured childrenwho do not require monitoring and/or inter-vention via a method that is not recommendedby NHTSA. Finally, 56% of agencies use in-appropriate methods when transporting chil-dren who are part of a multiple patient trans-port. All of the agencies reported necessarychanges to implement the recommendations,the most common of which were provision ofprovider education (75%) and the purchase ofnew equipment (52%). In addition, 61% wereunsure about the availability or did not havethe financial means to implement the recom-mendations. Conclusion. Few EMS agenciesare aware of the draft NHTSA guidelines onsafe transport of children in ground ambu-lances, and many agencies do not currently uti-lize acceptable transport methods. In additionto knowledge, cost of education and equipmentpurchase may prohibit implementation.

26. EMS STRETCHER “MISADVENTURES” IN ALARGE, URBAN EMS SYSTEM: A LONGITUDINALANALYSIS OF CONTRIBUTING FACTORS ANDRESULTANT INJURIES

Tyler Kallsen, Annette Arthur, StephenThomas, Jeffrey Goodloe, University ofOklahoma Health Science Center

Background. Earlier analysis in this large, ur-ban emergency medical services (EMS) systemdescribed EMS stretcher “misadventure” con-tributing factors and associated injuries. Therecontinues to be paucity of data regarding thisimportant aspect of safety for patient and EMSprofessional alike. This longitudinal study’spurpose was to continue to describe and an-alyze characteristics associated with undesir-able stretcher operations, with or without re-sultant injury to patients and/or EMS profes-sionals in a large, urban EMS agency. Methods.In the EMS agency studied, all stretcher-relatedmisadventures are required to be documentedby EMS personnel, regardless of whether in-jury results, using free-text incident-reportingsoftware. All such stretcher-related reports forincidents that occurred between July 1, 2009,and June 30, 2012, were queried from theagency’s risk-management database for retro-spective analysis, avoiding the Hawthorne ef-fect in stretcher operations. Results. Duringthe three years studied, the EMS agency trans-ported 404,178 patients. Fifty-nine stretcher in-cidents were reported (0.15 per 1,000 trans-ports). No substantive patient injury occurred.Eight EMS providers sustained minor injuries,including five back injuries, two knee injuries,and two arm contusions. There were three pri-mary times of stretcher operation problems: un-loading, loading, and surface movement. Forty-one of 59 incidents (69.5%) occurred duringunloading; five of 59 (8.5%) occurred duringloading; And 13 of 59 (22.0%) occurred duringsurface movement. There were five predomi-nant contributing aspects to stretcher operationproblems, with some incidents stemming frommultiple aspects: stretcher–ambulance safetylatch mechanism, ground surface conditions,equipment failure, bariatric patient size, andcombative patient behavior. Nineteen of 59(32.2%) related to the stretcher’s not engaginglocking mechanisms on the ambulance floor; 13of 59 (22.0%) related to poor ground surfaceconditions; five of 59 (8.5%) related to equip-ment malfunction; three of 59 (5.1%) related topatient weight exceeding 450 lb, compoundedby patient movement on the stretcher; And two

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


of 59 (3.4%) related to combative patient move-ment. There was no association between causeand crew injury, Fisher’s exact test = 0.087.Conclusions. In a large, urban EMS system, theincidence of injury related to stretcher opera-tions is markedly low, with few personnel in-juries and no substantive patient injuries in-curred during the three-year study period. EMSpersonnel should be particularly aware of in-creased risk for stretcher misadventure duringpatient unloading.

27. HEAD-TO-HEAD COMPARISON OF DISASTERTRIAGE METHODS IN PEDIATRIC, ADULT, ANDGERIATRIC PATIENTS

Keith Cross, Mark Cicero, University ofLouisville

Background. A variety of methods have beenproposed and used in disaster triage situations,but there is little more than expert opinionbehind most of them. Anecdotal disaster ex-periences have often reported mediocre real-world triage accuracy. Objective. To evalu-ate the comparative accuracy of several dis-aster triage methods for predicting mortalityin a large number of trauma patients. Meth-ods. Pediatric, adult, and geriatric trauma vic-tims from the National Trauma Data Bankwere assigned triage levels using each of sixdisaster triage methods: simple triage andrapid treatment (START); Fire Departmentof New York (FDNY); CareFlight; GlasgowComa Scale (GCS); Sacco Score; and Unad-justed Sacco Score. Triage assignments werecompared against patient mortality at hospi-tal discharge using area under the receiver-operating curve (AUC). Secondary outcomesincluded death on arrival, use of a venti-lator, and lengths of stay. Sensitivity anal-ysis assessed triage accuracy in patients byage, trauma type, and gender. Results. 530,583records were included in this study. The SaccoScore predicted mortality most accurately, withAUC of 0.883 (95% confidence interval [CI]:0.880–0.885), and performed well in most sub-groups. The FDNY criteria were more accu-rate than START for adults, but less accuratefor children. CareFlight did best in burn vic-tims, with an AUC of 0.87 (95% CI: 0.85–0.89),but mistriaged more salvageable trauma pa-tients to “Dead/Black” (41% survived) thandid other disaster triage methods (∼10% sur-vived). Conclusions. Among six disaster triagemethods compared against actual outcomes intrauma registry patients, the Sacco Score pre-dicted mortality most accurately. This analy-sis highlighted the comparative strengths andweakness of START, FDNY, CareFlight, andSacco, suggesting areas in which each might beimproved. The GCS predicted outcomes simi-larly to dedicated disaster triage strategies.

28. MASS-GATHERING MEDICINE: FACTORSPREDICTING PATIENT PRESENTATION

Samuel Locoh-Donou, Yan Guofen, MelanieWelcher, Thomas Berry, William Brady,University of Virginia

Background. This study was conducted toidentify the event characteristics predicting pa-tient presentation rates at all mass gatheringsheld at or near a Southeastern U.S. university.Methods. We conducted a retrospective reviewof all EMS records from mass-gathering pa-tient presentations between October 24, 2009,and August 27, 2011. All patrons seen by EMSwere included. Event characteristics includedcrowd size, venue percentage seating, venuelocation (inside vs. outside), venue bound-aries (bounded vs. unbounded), presence offree water, presence of alcohol, average heatindex, presence of climate control, and eventcategory (football, concerts, public exhibitions,nonfootball athletic events). We identified 79

mass-gathering events, for a total of 670 pa-tient presentations. The cumulative patron at-tendance was 917,307. The patient presentationrate (PPR) for each event was calculated as thenumber of patient presentations per 10,000 pa-trons in attendance. A Poisson logistic regres-sion model analysis was used to link this rateto the event characteristics while controlling forcrowd size. Results. Univariate logistic regres-sion analysis making use of rate ratios (RRs)revealed that increased rates were strongly as-sociated with outside venues (RR = 3.002, p =0.000), absence of free water (RR = 1.663, p =0.048), absence of climate control (RR = 0.330,p = 0.000), and a high heat index (RR = 1.211,p = 0.003). For every 10-unit increase in theheat index, the PPR increased by 21%, andthe presence of climate control resulted in a67% decrease in the PPR. The presence of al-cohol was not found to significantly affect thePPR. Football events were found to have thehighest PPR, followed sequentially by publicexhibitions, concerts, and nonfootball athleticevents. Models including event characteristicsas covariates were subsequently developed tojointly predict the rates of medical events. Con-clusion. Among the predictors studied, sev-eral were found to be strongly associated withthe rate of patient presentations. These find-ings should be considered during the processof EMS resource planning for future mass gath-erings.

29. PARAMEDIC WORK SAMPLING FOR THEASSESSMENT OF PARAMEDIC TRAINEES: ARELIABILITY, VALIDITY, AND FEASIBILITY STUDY

Walter Tavares, Justin Mausz, Victor Sun,Univerisity of Toronto

Background. Maintaining patient safety inprehospital settings involves ensuring thatparamedics are indeed competent. Workplace-based assessments (WBAs) are challenging;however, they have the advantage of authen-ticity and the potential for optimal constructvalidity if sampled appropriately. The pur-pose of this study was to establish the re-liability, validity, and feasibility of a WBAthat involved sampling trainee performancein an emergency medical services (EMS) con-text for entry-to-practice decisions. Methods.We used a prospective observational study de-sign. Paramedic clinical sampling (PCS) fol-lowed an objective structured clinical examina-tion (OSCE)-like structure by having traineesmove through five different “stations” (i.e.,raters and clinical cases) but in an EMS set-ting. Trainees responded to emergency callsand were required to demonstrate the techni-cal and nontechnical skills expected of an entry-level paramedic from point of contact to trans-fer of care, for any patient interaction withwhich they happened to be presented. Relia-bility was determined using scores assignedby raters on a seven-dimension global ratingscale. Content validity (evaluated by compar-isons with EMS data) and convergent validity(evaluated by comparing PCS scores to OSCEscores) were used to inform construct valid-ity. Feasibility was defined as rater satisfaction,compliance with the planned assessment pro-cess, and the absence of rater intervention. Re-sults. Forty-nine trainees were assessed overthree weeks using a PCS approach. The over-all mean score for trainees was 5.39 (standarddeviation ±0.48). Reliability was calculated us-ing generalizability theory and reached a g-coefficient of 0.49. Context specificity (i.e., vari-ance attributable to cases) was the most signif-icant threat to reliability. A D-study revealedthat 17 cases would be needed to achieve a reli-ability (g-coefficient) of 0.76 or 13 cases for aninterval of less than 1 point. Case variabilitywas similar to the most common case types ex-perienced by a large EMS. The disattenuated

correlation between this WBA and the OSCEwas r = 0.73 (p = 0.01). Rater satisfaction withthe process reached 72% with a 94% compliancerate. Raters needed to intervene in 12% of thetrainee-led patient interactions. Conclusions.PCS has the potential to be psychometricallydefensible and feasible, and therefore shouldbe considered for the assessment of paramedictrainees at the entry-to-practice level.

30. USE OF ONLINE MODULES IN TEACHINGTOXICOLOGIC EMERGENCIES TO EMS PROVIDERS

Jordan Guffin, Jennifer Werner, Adam Tobias,Michele Dorfsman, University of Pittsburgh

Background. Emergency medical services(EMS) providers are often the first contactwith patients presenting with toxicologicemergencies. While life-threatening toxicologicemergencies are rare, their recognition andtimely intervention are critical. The purpose ofthis intervention was to provide paramedicswith Web-based continuing education on themanagement of toxicologic emergencies in theprehospital setting. Methods. Between June1 and August 1, 2012, paramedics in WesternPennsylvania participated in an online modulecontaining multiple toxicology case scenariosand knowledge acquisition and retentionquestions. Participants completed pre- andpostmodule surveys to obtain feedback on theuse of the online module as an educational tooland to assess knowledge acquisition regardingthe toxicology topics. Results. Thirty-sevenparamedics participated in this module, and27 of them completed it. Of the paramedicswho completed the module, survey, andtest, the median age range and years of EMSexperience were 31–40 years old and greaterthan 10 years, respectively. Twenty-sevenparticipants (100%) had participated in onlinecontinuing education in the past. Following theprogram, 89% of the participants reported theywould participate in a similar training programagain, 74% reported enjoying learning ontheir own time via online education, and 67%reported that online education improved theirEMS field performance. Only 15% felt that theyhad better knowledge retention with onlinelearning. The knowledge acquisition questionsoverall were answered with 77% accuracy,compared with 59% prior to completing themodule. Conclusions. Paramedics participat-ing in this educational initiative enjoyed usingonline modules for continuing education oftoxicology cases and felt it would improvetheir EMS field performance. Paramedics wereable to demonstrate knowledge acquisitionfollowing participation in the module. Onlinemodules are feasible tools for use in continuingeducation for EMS providers. It will be impor-tant to determine what barriers led to someparticipants’ not completing the modules.Further study is needed to determine whetheronline modules offer the same level of knowl-edge retention over time when compared withtraditional methods of education, and whetherthis type of training can be used as a form ofremediation.

31. EXPERIENTIAL AND RATIONAL CLINICALDECISION MAKING: A SURVEY TO DETERMINEDECISION-MAKING STYLES OF PARAMEDICS

Jan Jensen, Lisa Calder, Mark Walker,Andrew Travers, Walter Tavares, AshaBienkowski, Pat Croskerry, DalhousieUniversity Division of EMS

Background. Two major processes underliehuman decision making: experiential (“intu-ition”) and rational (conscious and deliber-ate). The predominant process that paramedicsuse for making clinical decisions is unknown.The objective of this study was to determineparamedics’ preferences toward and perceived

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


ability to use experiential and rational thinking.Methods. Paramedics employed in a provincialground ambulance system voluntarily com-pleted a paper survey during 2012 professionaldevelopment sessions. The survey includedeight demographic questions and the RationalExperiential Inventory-40, a 40-question val-idated psychometric tool, scored on a five-point Likert scale. The tool has 20 questionson each thinking style, of which 10 assess fa-vorability and 10 assess ability to use thatstyle. Analysis included descriptive statistics,and t-tests to determine differences in over-all thinking style scores, favorability towardeach style, and ability to use each style. Differ-ences in thinking styles were evaluated withindemographic variables with t-tests and analy-sis of variance (ANOVA). Significance was setat 0.05 for all tests, with corrections for ANOVAcomparisons. Results. The response rate was96.1% (904/941). Most participants were male(n = 628, 69.5%), primary or advanced careparamedics (PCPs, n = 502, 55.5%; ACPs, n =294, 32.5%), the median age was 36 years (in-terquartile range [IQR] 29–42), and the mediannumber of years of experience was 10 (IQR4–16). The mean rational scores were higherthan experiential: 3.86/5 (95% confidence inter-val [CI] 3.83–3.87) vs. 3.41/5 (95% CI 3.38–3.44),p < 0.001. The participants scored their abil-ity to use rational thinking higher than expe-riential (3.93/5 [95% CI 3.90–3.96] vs. 3.60/5[95% CI 3.58–3.63], p < 0.001), and more fa-vored rational than experiential (3.79/5 [95%CI 3.76–3.82] vs. 3.22/5 [95% CI 3.18–3.26],p < 0.001). The only demographic subgroupin which a difference was found in experi-ential scores was gender, with males scoringhigher than females: 3.67 vs. 3.51, p < 0.001.The following demographic subgroups scoredrational questions higher: younger age (p <0.001), ACPs vs. PCPs (3.94/5 vs. 3.83/5, p =0.05), fewer years of working experience (p <0.001), and participants working in urban andmixed urban–rural areas versus rural areas(p = 0.001). Conclusion. Paramedics perceivethey have the ability to use and favor ratio-nal over experiential thinking. Future researchincludes determining differences in thinkingstyles of paramedics with work experience andstudents. This study adds to what is known onparamedic decision making, and is importantfor developing continuing education and clini-cal support tools.

32. EMERGENCY MEDICAL SERVICES EVALUATIONOF FALLS IN ASSISTED LIVING FACILITIES: ARETROSPECTIVE COHORT STUDY AND CLINICALPROTOCOL EVALUATION

Michael Bachman, J. Myers, JeffersonWilliams, Diane Miller, Benjamin Currie,Michael Lyons, Joseph Zalkin, Valerie DeMaio, Holly Tibbo-Valeriote, JohnaRegister-Mihalik, A. Jones, Alan Kronhaus,Wake County Department of EmergencyMedical Services

Background. Emergency medical services(EMS) often transports from assisted living fa-cilities (ALFs) patients who suffer simple falls.An EMS protocol could avoid unnecessarytransport for a subset of these patients, whileensuring that patients with time-sensitiveconditions are transported. Our objectivewas to begin to derive an EMS protocol todetermine which patients require transport.Methods. We conducted a retrospective cohortstudy of patients in our urban EMS system(population 900,000), in a subset of ALFsserved by a specific primary care group,who were transported to the emergencydepartment from July 2010 to June 2011 for achief complaint of “fall.” The primary outcomewas defined as “time-sensitive intervention”

(TSI) and was met by patients who had woundrepair or fracture, admission to the intensivecare unit, operating room, or cardiac catheteri-zation laboratory, death during hospitalization,or readmission within 48 hours. A priori,an EMS protocol to require transport forpatients needing TSI was derived by consensusbetween EMS and primary care. The protocolutilizes screening criteria including history andexamination findings to recommend transportversus nontransport with close primary carefollow-up. The EMS protocol was retrospec-tively applied to determine which patientsrequired transport. Data were analyzed usingstandard descriptive statistics. Results. Of 653patients transported across 30 facilities, 644had sufficient data. Of these, 197 (31%) met theprimary outcome. Most patients who requiredTSI had fracture (73) or wound repair (92). TheEMS protocol identified 190, for a sensitivity of96% (95% confidence interval [CI]: 93% to 98%),specificity of 46% (95% CI: 44% to 47%), andnegative predictive value of 97% (95% CI: 93%to 99%). Of seven false negatives, three werereadmitted (and redischarged) for another fall,three had hip fractures that were repaired,and one had a lumbar compression fracturethat was discharged. Conclusions. In thiscohort, one-third of patients with falls in ALFsrequired TSI and therefore EMS transport. AnEMS protocol may have sufficient sensitivityto safely allow for nontransport of a subsetof patients with falls in ALFs. Prospectiveevaluation of this protocol is necessary tovalidate this hypothesis.

33. DEVELOPING A FUNCTIONAL “GOLDSTANDARD” FOR MASS-CASUALTY TRIAGE

E. Brooke Lerner, Courtney McKee, CharlesCady, David Cone, M. Riccardo Colella,Arthur Cooper, Phillip Coule, Julio Lairet,J. Marc Liu, Ronald Pirrallo, Scott Sasser,Richard Schwartz, Greene Shepherd,Raymond Swienton, Medical College ofWisconsin

Background. Research on mass-casualty triagesystems is inhibited because there is no func-tional “gold standard” available to calculatesensitivity and specificity. Until there is agree-ment on the types of patients who should beidentified by each triage category, it is notpossible to evaluate or compare the accuracyof the various triage systems being utilized.The objective of this study was to develop aconsensus-based, functional gold-standard def-inition for each mass-casualty triage category.Methods. Experts were recruited through thelead investigators’ contacts and their suggestedcontacts. Key informant interviews were con-ducted to develop a list of potential criteriafor each triage category. Participants were in-terviewed in order of their availability until re-dundancy of themes. Participants were blindedto each other’s responses during the interview.Based on the results of the interviews, a mod-ified Delphi survey was developed and deliv-ered to all recruited experts. In the first tworounds, the participants could add, remove,or modify criteria. In the final rounds, ed-its were made to the criteria until there wasat least 80% agreement. Results. Thirteen na-tional and local experts were recruited to par-ticipate in the project. A total of six interviewswere conducted. Three rounds of voting wereperformed, with 12 respondents participatingin the first round, 12 in the second round,and 13 in the third round. After the first tworounds, the criteria were modified accordingto respondent commentary. In the final round,over 90% agreement was achieved for all butone criterion. A single e-mail vote was con-ducted on edits to that criterion and consen-sus was achieved. An example of a final crite-

rion is “minimal” patients are those dischargedfrom the emergency department with no x-rays or an extremity x-ray that was negative orshowed an uncomplicated fracture; had no lab-oratory testing; received only simple wound re-pair (single-layer suturing only); and receivedno intravenous medications from emergencymedical services or in the hospital. Conclusion.A consensus-based, functional gold-standarddefinition for each mass-casualty triage cate-gory has been developed. These gold-standarddefinitions can be used to evaluate the accuracyof mass-casualty triage categories after an ac-tual incident or during training.

34. COMPARING TWO PREDICTION MODELS FORMASS-GATHERING EVENTS

Jose Nable, Asa Margolis, Benjamin Lawner,Alexander Perricone, Michael Millin, SamuelGalvagno, Debra Lee, Richard Alcorta,University of Maryland School of Medicine

Background. Predicting medical resource us-age by spectators at mass gatherings has histor-ically been difficult. The literature describes atleast two models to forecast the numbers of pa-tients requesting medical assistance and thosewho require transport to an emergency depart-ment (ED). The objective of this study was tocompare the ability of two models to accuratelypredict the number of patients evaluated andtransported at a mass-gathering event. Meth-ods. This retrospective analysis of the 2011 Bal-timore Grand Prix (BGP) evaluated the abilityof the Arbon and Hartman methods to predictthe number of patient evaluations and patientsrequiring transport to an ED. In the Arbonmethod, several environmental features at theBGP, including crowd size, weather, and type ofevent, were input into an equation derived byregression modeling. The Hartman model uti-lized a scoring system based on weather, pres-ence of alcohol, crowd size, and age and in-tention of the spectators to stratify the incidentas a “minor,” “intermediate,” or “major” event,with predicted resource demand for each cat-egory. The actual number of patients evaluatedand transported at the BGP were tabulated withan electronic patient tracker system, and com-pared with the numbers predicted by the Ar-bon and Hartman models. Results. For the dayof the BGP with the largest crowd size (60,000),the Arbon method predicted a total of 64.4 pa-tient evaluations with 2.2 requiring transport.The Hartman scoring system categorized theevent as a “major” incident, with a mean pre-dicted 71 evaluations and 5.5 needing trans-port. Actual patient data demonstrated a totalof 52 evaluations with 4 transported. While theArbon prediction had a variance of 24% fromactual number of evaluations, the Hartman pre-diction was 37% greater than the true amount.The Arbon method for predicting transportswas 45% less than actual, whereas the Hartmanmethod predicted a value 38% greater than theactual number of transports. Conclusions. Boththe Arbon and Hartman methods are betterat predicting numbers of patient evaluationsthan transports. These observations call atten-tion to the need to develop a versatile and accu-rate model to predict resources needed at mass-gathering events.

35. A TARGETED APPROACH TO COMMUNITYCONSULTATION FOR AN EXCEPTION FROMINFORMED CONSENT STUDY: MORE SUPPORTTHAN WILLINGNESS TO PARTICIPATE

Eugene Vu, Kathy Arnold, Tony Carnevale,Mohamud Daya, Denise Griffiths, Dana Zive,Aarzoo Sidhu, Terri Schmidt,OHSU Emergency Medicine

Background. Community consultation (CC)is a key part of research conducted under

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


the Food and Drug Administration Excep-tion from Informed Consent (EFIC) frame-work. The best practices to achieve CC re-main uncertain. We report results from a tar-geted approach that surveyed at-risk membersof a health maintenance organization (HMO).Methods. The Resuscitation Outcomes Consor-tium (ROC) Amiodarone, Lidocaine, or PlaceboStudy (ALPS) is being conducted under EFIC.As part of the CC effort, we designed a surveywith the Kaiser Permanente Northwest Cen-ter for Health Research. Surveys were mailedto 2,000 Kaiser members between the ages of48 and 78 years residing in a three-county re-gion served by participating emergency medi-cal services agencies. The age group was con-sidered representative of the population atrisk for out-of-hospital cardiac arrest (OHCA).The survey described ROC-ALPS and the opt-out option. Respondents were asked whetherthe study was important to do, whether thebenefits justified doing it under EFIC, andwhether they would be willing to be enrolledin the study. Responses were recorded usinga five-point Likert scale. Demographic infor-mation collected included age, gender, andrace. Descriptive statistics were used to tab-ulate responses to survey questions. Pearsonchi-square test was used for subgroup analy-ses. Results. Of the 2,000 mailed surveys, 1,951were delivered and 337 were returned (17.2%response rate). The respondents were 48.8%male, with a mean age of 63.1 years. Most re-spondents were white (89.3%) and representa-tive of the study community. A majority (88.3%)of the respondents agreed that EFIC researchis important, and 73.9% felt that the benefitsjustify this method of research. Fewer respon-dents (53.1%) expressed a willingness to beenrolled in ROC-ALPS. There were no gen-der differences in question response. More re-spondents under the age of 65 years expresseda willingness to be enrolled in the study (p= 0.002). However, those ≥65 years old weremore likely to agree that the benefits justifythe risk in EFIC studies (p = 0.029). Con-clusions. A majority of at-risk HMO commu-nity member respondents expressed supportfor EFIC research, and more than half expresseda willingness to be enrolled in the ROC ran-domized placebo-controlled ALPS study. Theresponse rate was low, and differences wereidentified between respondents in relation toage.

36. THE INTERACTION OF CHEST COMPRESSIONRATES WITH THE IMPEDANCE THRESHOLDDEVICE AND ASSOCIATION WITH SURVIVALFOLLOWING OUT-OF-HOSPITAL CARDIAC ARREST

Ahamed Idris, Danielle Guffey, Paul Pepe,Siobhan Brown, Tom Aufderheide, StevenBrooks, Clifton Callaway, Jim Christenson,Daniel Davis, Mohamud Daya, Randal Gray,Peter Kudenchuk, Jonathan Larsen, StevenLin, James Menegazzi, UT SouthwesternMedical Center

Background. The Resuscitation OutcomesConsortium (ROC) PRIMED trial found nodifference in survival to hospital dischargewith a modified Rankin score (MRS) ≤3with standard cardiopulmonary resuscitation(CPR) plus a sham versus active impedancethreshold device (ITD) when used by emer-gency medical services (EMS) in adults without-of-hospital cardiac arrest. We evaluatedthe potential interaction of chest compressionrate with the sham versus active ITD andoutcome. Methods. Data were abstracted frommonitor–defibrillator recordings during thefirst 5 minutes of CPR in patients enrolled inthe ROC PRIMED ITD Trial. We retrospectivelyestimated the interaction between compressionrate and ITD on survival to hospital discharge

with MRS ≤3. We fit a natural cubic splinecurve to characterize the relationship betweencompression rate and outcomes. Results.Of 8,755 patients enrolled from June 2007 toNovember 2009, 6,188 had chest compressionrate and fraction data and 4,170 also had chestcompression depth data available, constitutingthe final study population. The mean age was68 years. The mean chest compression ratewas 106 compressions/min. The unadjustedcubic spline curves showed that peak survivalfor the sham and active ITD groups occurredat chest compression rates of 118 and 101compressions/min, respectively (p = 0.47).Peak survival with MRS ≤3 occurred at 118and 99 compressions/min for the sham andactive ITD groups, respectively (p = 0.58).After adjustment for gender, age, bystanderCPR, arrest location, ROC site, first EMSrhythm, witnessed status, and quality of CPR(chest compression fraction and depth), theinteraction between rate and ITD for survivalwith MRS ≤3 was statistically significant (p =0.036), but not for all survival irrespective ofMRS score (p = 0.093). Conclusion. In theadjusted model, we observed a significantchest compression rate–dependent interactionwith active ITD use for survival with MRS≤3. Optimal CPR rates were different withstandard CPR and a sham versus activeITD.

37. SLOW CHEST COMPRESSION RELEASEVELOCITY IMPAIRS HEMODYNAMIC POWER INTHE ABDOMINAL AORTA

Joshua Lampe, Josiah Garcia, Tai Yin, GeorgeBratinov, Christopher Kaufman, LanceBecker, University of Pennsylvania

Background. Research has suggested that chestcompression (CC) release velocity or wave-form impacts cardiopulmonary resuscitation(CPR) effectiveness. However, the impact ofchanges in CC waveform on blood flow andpressure in the abdominal aorta (AA) duringprolonged CPR has not been thoroughly in-vestigated. Methods. CPR hemodynamics ineight domestic swine (∼30 kg) were studied us-ing standard physiologic monitoring. A flowprobe was placed on the AA and a pressurecatheter tip was located in a corresponding re-gion. Ventricular fibrillation (VF) was electri-cally induced. Mechanical CCs were started af-ter 10 minutes of untreated VF. CC release wasadjusted so that sternal recoil lasted 100 ms,200 ms, or 300 ms. CCs were delivered over 54minutes at a rate of 100/min and at a depthof 1.25 in. Transitions between waveforms oc-curred every 2 minutes and were randomized.Hemodynamic power was calculated as flow× pressure. A peak decay phase was iden-tified during which the drop in power overtime became exponential. Results. Peak de-cay in AA power started after approximately12 minutes of CPR. Pairwise comparisons in-dicated a significant effect of CC waveform.Transitioning from 300-ms to 100-ms releasetime increased average AA power by 8.5% ±13.6%, while the reverse transition caused a de-crease of –14.1% ± 14.1% (p < 0.001). Tran-sitioning from 200-ms to 300-ms release timehad a similar negative effect on AA power(–8.7% ± 8.4%), whereas comparisons between100-ms and 200-ms release times showed nosignificant difference. Conclusions. CC releasevelocity significantly alters abdominal aortichemodynamic power during prolonged CPR.Faster release velocities were associated withpreserved or improved power, whereas slowerrelease velocities were associated with signif-icant reductions in power. Further researchis required to determine how best to takeadvantage of the power difference betweenwaveforms.

38. CHEST COMPRESSION RATES BETWEEN 100AND 120 PER MINUTE ARE INDEPENDENTLYASSOCIATED WITH INCREASED ADULT SURVIVALFOLLOWING OUT-OF-HOSPITAL CARDIAC ARREST

Ahamed Idris, Daniel Davis, CliftonCallaway, Jim Christenson, Peter Kudenchuk,Jonathan Larsen, Steven Lin, JamesMenegazzi, Kellie Sheehan, George Sopko,Mohamud Daya, Randal Gray, Ian Stiell,Graham Nichol, Danielle Guffey,UT Southwestern Medical Center

Background. American Heart Association car-diopulmonary resuscitation (CPR) guidelinesrecommend a chest compression rate of atleast 100 compressions/min, but provide noguidance for an upper rate limit. The ob-jective of this study was to assess the re-lationship between chest compression ratesused by emergency medical services (EMS)providers and survival to hospital dischargeafter out-of-hospital cardiac arrest (OHCA).Methods. Adults (≥20 years) with OHCAtreated by EMS providers were enrolled inthe Resuscitation Outcomes Consortium (ROC)PRIMED study. Data were abstracted frommonitor–defibrillator recordings for the first 5minutes of CPR. Multiple logistic regression as-sessed the odds ratio (OR) for survival by com-pression rate categories (<80, 80–99, 100–119,120–139, ≥140/min), unadjusted and adjustedOR for gender, age, witnessed status, attemptedbystander CPR, location of arrest, chest com-pression fraction and depth, rhythm, and ROCsite. We used a global test of association to as-sess significance. Results. 10,371 patients withOHCA had CPR from June 2007 to November2009 with compression rate available; 6,399 alsohad chest compression fraction and depth data.The mean (± standard deviation [SD]) age was67 ± 16 years. The mean (±SD) compressionrate was 111 ± 19/min, compression fractionwas 0.66 ± 0.16, and compression depth was42 ± 12 mm. Return of spontaneous circulationoccurred in 34%, and 9% survived to hospitaldischarge. The adjusted model (without chestcompression depth and fraction) did not showa significant relationship between chest com-pression rate categories and survival (p = 0.19).However, in the subgroup of subjects with com-pression depth data, the global test found asignificant relationship between chest compres-sion rate categories and survival, both with anadjustment for compression depth and fraction(p = 0.02) and without (p = 0.02). Conclusion.Survival to hospital discharge after OHCA isgreater when chest compression rates are main-tained in a range of 100 to 120/min.

39. CHANGES IN TISSUE OXIMETRY DURING THESPECTRUM OF CARDIAC ARREST ANDRESUSCITATION IN A PORCINE MODEL

Joshua Reynolds, David Salcido, AdamFrisch, Brian Suffoletto, James Menegazzi,University of Pittsburgh

Background. Intensive monitoring during re-suscitation remains relatively crude and is dif-ficult in the prehospital environment. Near-infrared spectroscopy (NIRS) noninvasivelymeasures real-time tissue oxygen content byapproximating the hemoglobin oxygen satu-ration fraction in terminal vasculature of tis-sue. Using our established porcine model ofout-of-hospital cardiac arrest (OHCA), we as-sessed the utility of continuous tissue oxy-gen saturation (StO2) measurement through-out the spectrum of resuscitation. Our hypoth-esis was that StO2 will decrease with loss ofpulses, increase with resuscitation measures(cardiopulmonary resuscitation [CPR] and ad-ministration of epinephrine), and return tobaseline with return of spontaneous circulation

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


(ROSC). Methods. We anesthetized and instru-mented seven female swine, placing a nonin-vasive NIRS probe on the left forelimb thatrecorded continuous StO2. After 8 minutes ofuntreated ventricular fibrillation and 2 min-utes of mechanical CPR, we randomized theanimals to 0.015 mg/kg (SDE) or 0.1 mg/kg(HDE) epinephrine. After 3 additional minutesof CPR, the animals were defibrillated (150 Jbiphasic). We attempted subsequent defibril-lations using quantitative waveform measuresto guide shock delivery. Data were analyzedwith descriptive statistics and a generalized lin-ear model with alpha = 0.05 to determine theoverall slope of the pooled StO2 across animalsfor each segment of resuscitation. Results. Themean weight was 29.7 ± 1.9 kg and the meananesthesia duration was 67.4 ± 13.0 minutes.Four animals received HDE and three SDE. Allanimals achieved ROSC. Significant coefficients(StO2/min) were noted for each segment of re-suscitation. The StO2 rapidly decreased imme-diately after loss of pulses (–29.1; 95% confi-dence interval [CI] –33.4, –24.7; p < 0.01), tak-ing 31.2 ± 7.8 seconds to decrease by 25%. Itdid not change during during CPR (–0.2; 95%CI –1.2, 0.8; p = 0.71). There was a graded de-cline in StO2 between SDE (–1.3; 95% CI –1.5,–1.2; p < 0.01) and HDE (–3.1, 95% CI –5.8, –0.4;p = 0.03). The slowest change occurred withROSC (0.4; 95% CI 0.3, 0.5; p < 0.01), taking10.2 ± 0.8 minutes to increase by 25%. Conclu-sions. In our porcine model of OHCA, periph-eral StO2 rapidly decreased after loss of pulses,but did not improve with CPR or epinephrine.It increased extremely slowly after ROSC. Pe-ripheral StO2 may be a useful indicator of“pulselessness” in critically ill prehospital pa-tients (especially pulseless electrical activity,which is not detected by electrocardiography),but does not appear to be a rapid indicator ofROSC.

40. CORONARY PERFUSION PRESSURE RESPONSETO INTRAOSSEOUS EPINEPHRINEADMINISTRATION AFTER PROLONGED CARDIACARREST

Joshua Harris, Ryan Coute, Adam Kellogg,Scot Millay, Timothy Mader,Baystate Medical Center

Background. While provision of vasopres-sors during attempted resuscitation of out-of-hospital cardiac arrest (OHCA) victims hasnever been shown to improve neurologicallyintact survival, its short-term benefits are clear.Despite evidence that the intraosseous (IO)dose should be substantially higher than thatgiven intravenously (IV) to achieve phar-macokinetic and hemodynamic equivalency,current guidelines still recommend a doseof 0.01 mg/kg regardless of delivery route(IO or IV). Objective. To compare the ef-fect of epinephrine 0.1 mg/kg (HDE) IO with0.01 mg/kg (SDE) IV on coronary perfu-sion pressure (CPP) during resuscitation in aswine model of prolonged ventricular fibrilla-tion (VF). Methods. This was a secondary anal-ysis of prospectively collected data from twoInstitutional Animal Care and Use Commit-tee (IACUC)-approved protocols. Seventy-nineYorkshire swine (25–30 kg) were surgically in-strumented under anesthesia and VF was elec-trically induced. After 10 minutes of untreatedVF in the IO study (n = 26) and 12 minutes ofuntreated VF in the IV study (n = 53), resusci-tation commenced with precordial chest com-pressions. After 30 seconds of initial chest com-pressions, a single dose of epinephrine (HDEIO or SDE IV) was given followed by a large-volume saline flush. An additional 2.5 minutesof compressions was provided after injection tocirculate the medication before the first rescue

shock (RS) was delivered. Cardiopulmonary re-suscitation (CPR) and RS attempts were stan-dardized for all animals. The CPP was definedas aortic diastolic pressure minus right atrial di-astolic pressure and the values were extractedimmediately following the last compression be-fore defibrillation for the first RS in each animal.Descriptive statistics were used to analyze thedata. Results. After 10 minutes of untreated VF,HDE IO epinephrine resulted in a mean CPP of33.2 mmHg (95% confidence interval [CI]: 26.6,39.9) just prior to the first RS. After 12 min-utes of untreated VF, SDE IV epinephrine re-sulted in a mean CPP of 25.0 mmHg (95% CI:20.5, 29.4) just prior to the first RS. Conclusion.This observation study reaffirms the assertionthat HDE may be required to generate CPP val-ues similar to SDE delivered IV during resus-citation of prolonged VF. A randomized com-parison of HDE and SDE IO in the metabolicphase of VF is needed to test this hypothesisand determine the impact on ROSC and short-term survival.

41. REGIONAL VARIATION IN POST–CARDIACARREST CARE—A STATEWIDE EVALUATION

Robert Swor, Kelly Sawyer, Carol Clark,Carman Turkleson, David Haines,William Beaumont Hospital

Background. There is an evolving body of liter-ature advocating for regional post–cardiac ar-rest systems of care. We sought to describe thevariation in post–cardiac arrest care providedin a single large state. Methods. We used theMichigan Inpatient Data Base (MIDB) to eval-uate care provided to post–cardiac arrest pa-tients from July 2008 to June 2011. The MIDBis a comprehensive source of inpatient activ-ity at Michigan acute care hospitals. We in-cluded all patients admitted with InternationalClassification of Diseases, Ninth Revision, Clin-ical Modification (ICD-9-CM) diagnoses of car-diac arrest (ICD 427.5) or ventricular fibril-lation (427.41). Data were collated regardingpatient demographics, final diagnoses, andprocedures performed: therapeutic hypother-mia (TH), percutaneous coronary intervention(PCI), implantable cardiac defibrillator (ICD)placement, and coronary artery bypass graft-ing (CABG). We evaluated variation in care byregion, using the eight-state designated med-ical emergency care regions (MECR), whichare designed to facilitate disaster, trauma, andcardiac care. Descriptive statistics, correlationcoefficients, and chi-squares were calculated.Results. During the study period, there were21,856 patients with cardiac arrest admitted,with 9,014 (41.2%) discharged. Rates of TH in-creased by year (0.7%, 1.7%, 2.2%; p < 0.001).Significant regional variation in postarrest carewas observed for TH (range 0.3% to 3.9%, p <0.001). There was also significant variation byregion in the rates of PCI, CABG, and ICD. Pa-tients with the diagnosis of VF were more likelyto received TH (2.3% vs. 1.2% [odds ratio 1.9;95% confidence interval 1.5, 2.3]), and variedmore than eightfold by region (0.6% to 5.1%,p < 0.001). There was no correlation betweenthe rate of TH provision and the rate of other in-terventional cardiac care procedures or regionalcardiac arrest case volume. Conclusion. Use ofTH is increasing in Michigan, and greater than10-fold variation in the rate of TH provision be-tween state regions was observed. Significantregional variation exists for post–cardiac arrestinterventional care. Regional rates of provisionof TH did not correlate with interventional car-diac care, or cardiac arrest case volume. Thesedata suggest that identification of comprehen-sive regional cardiac arrest centers based onvolume or ability to provide services may notresult in TH care for postarrest patients.

42. COMPARISON OF PREHOSPITAL CONTINUOUSPOSITIVE AIRWAY PRESSURE (CPAP)NONINVASIVE VENTILATION DEVICES’ ABILITYTO MAINTAIN CPAP OVER TIME IN A SIMULATEDLUNG MODEL

Ryan Hodnick, Daniel Barela, David Slattery,Darren Braude, Bryan Bledsoe,University of Nevada

Background. Prehospital continuous positiveairway pressure (CPAP) devices are highlyvariable in terms of cost, design, and ability tomaintain continuous airway pressures. Oxygenconsumption and CPAP maintenance becomecrucial factors as transport times increase. Ob-jective. To compare various CPAP devices withregard to useful operating time (UOT), definedas the duration the device maintained any con-tinuous airway pressure (CDOT0) on standard-ized settings. Methods. Prospective, observa-tional study. Inclusion criteria: CPAP devicesapproved for prehospital use as of November2011 (CAREvent, PortO2Vent, Oxy-PEEP, O2-ResQ, MACS CPAP, Boussignac, Flow Safe, andWhisperFlow). Exclusion criteria: Prehospitalventilators capable of delivering CPAP. CPAPdevices were attached to a standard D cylin-der via a calibrated regulator. Each device’smask was tested using a Laerdal Airway Man-agement Trainer attached to an IngMar Med-ical Demonstration Lung Model. The follow-ing settings were used: respiratory rate of 12breaths/min, tidal volume of 500 mL, compli-ance of 34 mL/cmH2O, resistance set to normalphysiologic value, and peak flow of 30 L/min.Each device was tested at a CPAP of 10 cmH2O.Devices that had a variable fractional concen-tration of oxygen in inspired gas (FiO2) weretested at the lowest and highest FiO2 settings.Positive end-expiratory pressure (PEEP), a sur-rogate for CPAP, was directly measured us-ing Pulmonary Mechanics Graphics 3000 Mod-ule software (IngMar). All times were recordedusing a digital stopwatch. The primary per-formance measure was PEEP drop-off time 0(CDOT0), defined as the time until completeloss of PEEP. We rank-ordered results by de-creasing overall CDOT0 and stratified them bymaximal and minimal FiO2 values. Results. Wemeasured eight devices, three of which werestudied twice (maximal/minimal FiO2 settings)and five of which were studied once, for a to-tal of 11 trials. For the eight devices tested, therange of CDOT0 was from 4 minutes (Whis-perFlow) to 78 minutes (Oxy-PEEP). CDOT0(minutes:seconds) in decreasing performancewas as follows: 1) Oxy-PEEP–32% (78:41); 2)MACS CPAP–65% (76:30); 3) PortO2Vent–100%(68:13); 4) WhisperFlow–28% (47:35); 5) MACSCPAP–100% (43:39); 6) O2-ResQ–30% (35:48);7) CAREvent–100% (34:19); 8) Flow Safe–100%(31:15); 9) Boussignac–100% (30:05); 10) Oxy-PEEP–95% (28:16); and 11) WhisperFlow–100%(04:14). Conclusion. There is high variabil-ity in UOT among common CPAP devices.These findings may have important resourceimplications for emergency medical servicesagencies.

43. BODY TEMPERATURE OF BURN PATIENTS ATHOSPITAL ARRIVAL IN BURN PATIENTSPRESENTING TO PENNSYLVANIA BURN CENTERSFROM 2000 TO 2011

Matthew Weaver, Jon Rittenberger, HeatherKowger, P. Patterson, Serina McEntire, AlainCorcos, Jenny Ziembicki, David Hostler,University of Pittsburgh

Background. Prehospital burns requireproviders to limit further injury, expose thepatient and provide initial treatment, andmaintain normothermia. Little is knownabout temperature at hospital presentation

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


for emergency medical services (EMS)-treatedburn patients. We characterized temperature athospital arrival for patients presenting to Penn-sylvania burn centers. Methods. We analyzed11 years of prehospital data from burn patientsincluded in the Pennsylvania Trauma Out-comes Study registry (traumatic diagnoses ofInternational Classification of Diseases, NinthRevision, Clinical Modification [ICD-9-CM]800–995 and documented burn injury). Ourdependent variable of interest was temperatureat hospital arrival—a continuous measure thatwe dichotomized into normothermia (>36.5◦C)and hypothermia (≤36.5◦C). We report thecharacteristics of these patients using descrip-tive statistics and evaluated factors associatedwith hypothermia using hierarchical logisticregression. Results. The data set included8,025 records of patients transported by EMS.We excluded data from patients who died inthe emergency department and those withmissing temperature information. Our finaldata set included 7,616 records submitted by22 institutions. Only 4,626 (61%) of the samplewas normothermic upon hospital arrival. Ofthe remaining 39% (n = 2,990), the mediantemperature was 36.1◦C, interquartile range(IQR) 35.8–36.3, with 5% of the hypothermicpatients presenting below 35◦C. Hypothermicburn patients were more likely to have beenparalyzed and intubated (odds ratio [OR] 1.99;95 confidence interval [CI] 1.79–2.20) and tohave arrived as a trauma alert (OR 1.52; 95 CI1.39–1.66) and demonstrated a lower GlasgowComa Scale score when compared withnormothermic burn patients (median 15 [IQR11–15] vs. median 15 [IQR 15–15]; p < 0.01).Temperature on arrival was most commonlymeasured via oral or tympanic methods inboth groups. The majority of patients expe-rienced thermal burns (73.8%), and 55% hadless than 10% total second- and third-degreetotal burn surface area. The normothermicpatients were younger than the hypothermicgroup (median 27 years [IQR 9–47] vs. median34 years [IQR 14–52]; p < 0.01). The twogroups also varied with respect to method oftransport (p < 0.01) and type of burn injury(p = 0.03). Conclusions. In this large sample,nearly 40% of burn patients transported byEMS arrived at the hospital mildly hypother-mic. Clinical characteristics suggest thesepatients may have been more severely injured.Maintenance of normothermia and good resus-citation practices should be stressed to EMSproviders.

44. PROMOTING EARLY INTERVENTION DURINGSIMULATED HEMORRHAGE WITH THE USE OF AREAL-TIME DECISION-ASSIST ALGORITHM

Victor Convertino, Gary Muniz, DavidWampler, Craig Manifold, Greg Grudic,Steven Moulton, Robert Gerhardt,United States Army Institute of SurgicalResearch

Background. Lifesaving interventions requiredto effectively treat trauma patients with sig-nificant hemorrhage can be delayed, leadingto poor outcomes, when critical care person-nel rely on standard vital signs obtained fromcurrent medical monitors. Using new moni-toring technologies, we developed a machine-modeling algorithm that accurately identifiesloss of central blood volume and provides anearly prediction of the point at which an in-dividual will experience hemodynamic insta-bility (onset of shock). A real-time compre-hensive physiologic monitor algorithm reducesthe time a paramedic will identify an un-stable computer simulated hemorrhaging pa-tient. Methods. Fifty paramedics reviewed amonitor screen that displayed standard vitalsigns during a simulated hemorrhage profile

on two different occasions, once with and oncewithout a decision-support algorithm designedto show real-time tracking of physiologic re-sponses directly associated with changes incentral blood volume. The paramedics wereasked to push a computer key only if therewas an indication from the monitor that ledthem to believe the patient’s condition hadbecome unstable. Results. The mean time re-quired by the paramedics to identify an un-stable patient was 18.3 ± 4.1 minutes with-out the algorithm and 10.7 ± 4.2 minutes withthe algorithm. Using the algorithm reduced theresponse time by >40% (p < 0.001). Con-clusion. These data demonstrate that algo-rithms with real-time tracking of physiologicresponses associated with loss of circulatingblood volume can lead to earlier interventionof a bleeding patient. A real-time comprehen-sive physiologic monitor algorithm reduces thetime a paramedic will identify an unstable,computer-simulated hemorrhaging patient.

45. SPINE INJURIES IN MOTORSPORTS, THEINDIANAPOLIS STYLE PROFESSIONAL RACINGEXPERIENCE

Andrew Stevens, Terry Trammell, GeoffreyBillows, Michael Olinger,Indiana University School of Medicine

Background. Emergency medical services(EMS) focuses on rapid assessment andtransportation of motor vehicle crash victims.Patients with serious “mechanism of injury”are extricated from the vehicle and placed inspinal immobilization. Upon transport to thehospital, it is common for immobilized patientsto receive whole-body computed tomography(CT) to rule out life-threatening injuries. Incontrast, drivers in the Indianapolis StyleRacing Series involved in high-speed (oftengreater than 200 mph) accidents commonlyself-extricate and are often transported tomedical facilities without cervical collars orbackboards. It is not known, however, whetherdriver-initiated self-extrication increases therisk of neurologic injury after spinal fracture.Our objective was to evaluate the safety ofself-extrication in Indianapolis Style Racingby comparing the use of CT scans, need forsurgery, and neurologic outcomes of spinalfractures in drivers who self-extricated com-pared with those who required EMS-facilitatedextrication. Methods. We conducted a retro-spective review of Indianapolis Style RacingDrivers’ medical records between the years of2006 and 2011. Cumulative radiation exposure,need for surgery, and reported neurologicoutcomes were obtained from medical records.This time period represents compulsory headand neck support (HANS) and unified chassisdesign, and all venues had installed SteelAnd Foam Energy Reducing (SAFER) barriers.Inclusion criteria were involvement in a crashincident and alert upon initial evaluation.Results. We examined 135 crash incidents. Self-extrication occurred in 121 (90%) and overallcumulative radiation exposure ranged from100 to 250 mSv or 0.82 to 2.06 mSv per driver.During this study period, 14 (10%) driverswere extricated and immobilized and theoverall cumulative radiation exposure rangedfrom 140 to 350 mSv or 10 to 25 mSv per driver.A total of 29 injuries were identified. Spinalinjury included nine (31%) of these injuries.Six EMS-extricated and three self-extricateddrivers had a spinal injury. Zero of thesehad surgical disease or neurologic deficit onfollow-up. One driver was excluded becauseof multiple trauma severity upon presentation.Conclusion. In our Indianapolis Style Racingexperience, a protocol-led self-extricationsystem did not miss any surgical spine injuriesregardless of mechanism of injury in this

cohort of patients and reduced group-averagedradiation exposure.

46. EFFECT OF EMS AIRWAY SELECTION ONNEUROLOGIC STATUS OF SURVIVORS AFTEROUT-OF-HOSPITAL CARDIAC ARREST

Jason McMullan, Ryan Gerecht, JordanBonomo, University of Cincinnati

Background. Early animal data suggest thatsupraglottic airways (SGAs) impair cerebralperfusion. A post hoc analysis of a clinical trialsuggests that there may be a beneficial im-pact on neurologically intact survival with theuse of endotracheal tube (ETT) use in out-of-hospital cardiac arrest (OHCA), but this hasnot been evaluated in an unselected OHCApopulation. This preliminary analysis was per-formed to support a larger effort investigat-ing the association of airway choice on func-tional status after OHCA. We hypothesizedthat there would be no difference in functionaloutcome in this small unselected population.Methods. OHCA patients who survived to hos-pital admission and had an ETT or SGA placedby emergency medical services (EMS) betweenJanuary 1, 2011, and April 30, 2012, were se-lected from the CARES registry of an urbanfire-based EMS system. Patients with no EMSairway, unknown airway, or unknown neuro-logic status at discharge were excluded. Groupswere compared with chi-square analysis. Goodfunctional outcome was defined as a cerebralperformance category (CPC) <3. Results. Over-all, 81 of 284 (28.5%) OHCA patients survivedto hospital admission during the study pe-riod, and 61 were included for analysis (ex-cluded: 1 unknown airway, 2 unknown neu-rologic status, 17 no airway). There was noobserved difference in subject (age, race, gen-der) or OHCA (witnessed, initial rhythm, by-stander cardiopulmonary resuscitation) charac-teristics. Five of 27 (18.5%) with ETT and sevenof 34 (20.5%; p = ns) with SGA were dischargedwith good neurologic status. Conclusion. Inthis small preliminary analysis, there is not alarge difference in functional outcome in sur-vivors of OHCA with SGA or ETT airway se-lection.

47. THE IMPACT OF PREHOSPITAL NONINVASIVEPOSITIVE PRESSURE SUPPORT VENTILATION INADULT PATIENTS WITH SEVERE RESPIRATORYDISTRESS: A SYSTEMATIC REVIEW ANDMETA-ANALYSIS

Sameer Mal, Shelley McLeod, AllaIansavichene, Adam Dukelow, MichaelLewell, University of Western Ontario

Background. Noninvasive positive pressureventilation (NIPPV), which includes contin-uous and bilevel pressure modalities, hasbeen shown to reduce mortality, intubationrates, and intensive care unit (ICU) length ofstay (LOS) for patients admitted to hospitalwith acute pulmonary edema and acuteexacerbation of chronic obstructive pulmonarydisease. NIPPV is increasingly being used byemergency medical services (EMS) agenciesfor the treatment of respiratory distress inthe prehospital setting. The primary objectiveof this systematic review was to determinewhether prehospital administered NIPPV forthe treatment of adults with severe respiratorydistress reduces 30-day mortality as comparedwith “standard” therapy. Secondary objectiveswere to examine the effect of prehospital-administered NIPPV on the need for invasiveventilation, ICU LOS, and hospital LOS. Meth-ods. Electronic searches of Medline, EMBASE,Cochrane Central Register of Controlled Trials,and CINAHL were conducted, and referencelists for relevant articles were hand-searched.Randomized controlled trials comparing

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


the use of prehospital NIPPV with standardtherapy in adults (age >16 years) with severerespiratory distress published in the Englishlanguage were included. Two reviewers in-dependently screened the titles and abstracts,assessed the quality of the studies, and inde-pendently extracted data. Where appropriate,data were pooled using random-effects modelsand reported as risk ratios (RRs) with 95%confidence intervals (CIs) and number neededto treat (NNT). Results. Six randomized con-trolled trials were included with a combinedtotal of 583 patients: 289 in the standardtherapy group and 294 in the NIPPV group. Inpatients treated with prehospital NIPPV, thepooled estimate showed a reduction in both30-day mortality (RR: 1.67; 95% CI: 1.01, 2.78;NNT = 18) and need for invasive ventilation(RR: 2.59; 95% CI: 1.66, 4.05; NNT = 8). Therewas no difference in ICU or hospital LOS.Conclusion. Out-of-hospital administration ofNIPPV appears to be an effective therapy foradult patients with severe respiratory distress.

48. AIRTRAQ VERSUS MACINTOSHLARYNGOSCOPE FOR MANIKIN INTUBATIONDURING ACTIVE CARDIOPULMONARYRESUSCITATION

Gregory Reimer, Erin Weldon, R. Grierson,Travis Hildebrand, University of Manitoba

Background. Advanced airway placement dur-ing active cardiopulmonary resuscitation (CPR)has been associated with prolonged pausesin chest compressions and worse patient out-comes. The Airtraq indirect laryngoscope hasbeen proven to be a useful difficult airway ad-junct and may have an advantage over tra-ditional direct laryngoscopy with respect toease and timeliness of endotracheal tube place-ment during continuous cardiopulmonary re-suscitation (CPR) in the prehospital setting.The purpose of our study was to determinewhether the Airtraq indirect laryngoscopewould positively affect intubation outcomes inmanikins with and without active chest com-pressions. Methods. We carried out a random-ized crossover study comparing traditional di-rect laryngoscopy with the Airtraq indirectlaryngoscope in a simulated setting. Eighty-five experienced paramedics participated infour intubation scenarios: 1) Macintosh laryn-goscope without chest compressions, 2) Mac-intosh laryngoscope during continuous chestcompressions, 3) Airtraq laryngoscope with-out chest compressions, and 4) Airtraq laryngo-scope during continuous chest compressions.Following all intubation scenarios, the partici-pants voluntarily completed a six-question sur-vey. Primary outcomes were success rate, timeto intubation, and number of intubation at-tempts. Secondary outcomes included view ofthe larynx and ease of intubation. Results.The Airtraq laryngoscope did not improve suc-cess rate, decrease time to intubation, or de-crease the number of intubation attempts (p> 0.05) with or without active CPR. Based onvisual analog scores, the participants subjec-tively found that the Airtraq provided a supe-rior view of the larynx compared with the Mac-intosh laryngoscope (p < 0.0001). Overall, theparticipants found it easier to intubate with theAirtraq both in the manikin at rest and duringactive CPR (p < 0.0001). Conclusion. The Air-traq indirect laryngoscope performs similarlyto the Macintosh direct laryngoscope in termsof paramedic success rate, time to intubation,and number of intubation attempts in manikinswith or without active chest compressions. Sub-jectively, the Airtraq provides a better view ofthe larynx and allows for easier intubation withand without chest compression. Further studiesare warranted to determine whether the Air-traq laryngoscope is superior in a human pa-tient population.

49. PEAK INSPIRATORY PRESSURES DURINGVENTILATION OF THE KING LARYNGEAL TUBEAIRWAY WITH A TRANSPORT VENTILATOR IN THEPREHOSPITAL SETTING

Christian Martin-Gill, Heather Prunty, SethRitter, Jestin Carlson, Francis Guyette,University of Pittsburgh

Background. Supraglottic airways, includingthe King laryngeal tube (LT), are widely usedin the prehospital setting after failed endotra-cheal intubation or as primary airways in casesof suspected difficult intubation. While peakinspiratory pressures (PIPs) of greater than 30cmH2O have been associated with air leak andgastric insufflation, the incidence of high PIPand the frequency of these adverse events dur-ing LT use with a transport ventilator are un-known. Methods. We conducted a retrospec-tive review of LT use (King LT or King LTS-D) with a transport ventilator by a large he-licopter emergency medical service from Jan-uary 1, 2006, to August 31, 2011. We identi-fied all cases transported after LT placementand collected demographics, transport time onthe ventilator, initial PIP, end-tidal carbon diox-ide (ETCO2), oxygen saturation (SpO2), andincidence of complications. Standard ventila-tor settings were assist control, tidal volume8 mL/kg, rate 12 respirations/min, positiveend-expiratory pressure (PEEP) 5 cmH2O, andfractional concentration of oxygen in inspiredgas (FiO2) 100%. We used descriptive statis-tics with 95% confidence intervals (CIs). Re-sults. A transport ventilator was used in 49 of146 (34%) cases of LT placement. Nine caseswere excluded because of inadequate docu-mentation. Of the 40 patients analyzed, 57%were male, the average age was 53.2 years, 26cases (65%) involved a traumatic mechanism,36 patients (90%) received rapid-sequence intu-bation, and no patients were in cardiac arrest atthe time of airway placement. Mechanical ven-tilation occurred for a mean of 22.4 minutes (95CI 15.4–29.5). Mean initial PIP was 31.8 cmH2O(CI 27.8–35.8), and 18 patients (45.0%) had PIP>30 cmH2O. The initial and final ETCO2 read-ings were 47.1 mmHg (95 CI 42.1–52.1) and 37.8mmHg (95 CI 34.0–41.5). The initial and finalSpO2 readings were 93.0% (95 CI 89.8–96.2) and96.6% (95 CI 94.5–98.6). Mechanical ventilationwas discontinued in five patients (13%, 95 CI3–23%) because of elevated PIP (n = 3, 8%, 95CI 0–16%) or hypoxia (n = 2, 5%, 95 CI 0–12%).One patient with a PIP <30 cmH2O vomitedand required suctioning through the LT’s gas-tric port. Conclusion. Peak inspiratory pres-sures above 30 cmH2O are common during me-chanical ventilation through an LT airway, yetthe majority of patients are adequately venti-lated and oxygenated. Caution should be takento monitor for elevated PIP, and gastric decom-pression should be performed when possible.

50. FACILITATION OF UNINTERRUPTED CHESTCOMPRESSIONS BY PARAMEDICS: THE ROLE OFTHE VIDEO LARYNGOSCOPE

Samantha Kealey, Aaron Burnett, KentGriffith, Sandi Wewerka, Joshua Salzman,Zabrina Evens, Ralph Frascone, RegionsHospital

Background. New American Heart Associationguidelines emphasize the importance of unin-terrupted chest compressions (CCs) in cardiacarrest. Recent studies suggest that direct laryn-goscopy (DL) in the field contributes to signifi-cant pauses in CCs. The aim of this study was toexamine the role of video laryngoscopes (VLs)in facilitating uninterrupted CCs for prehospi-tal cardiac arrest patients. Methods. This wasa post hoc analysis from a multisite, prospec-tive, nonrandomized, crossover trial compar-ing the placement success rates of two VLs

(CMAC, Karl Storz; King VISION [KV], KingSystems). Inclusion criteria for this analysisconsisted of need for advanced airway manage-ment and cardiac arrest as the primary impres-sion. Patient, provider, and clinical demograph-ics were compared between treatment arms.The associations between ongoing CCs and de-vice type, as well as CCs and attempt success,were examined using a chi-square test. An at-tempt was defined as the tip of the VL bladepassing the patient’s lips. Results. There wereno demographic differences (provider or pa-tient) between the VL treatment groups. A to-tal of 106 VL attempts (62 CMAC and 44 KV)were made by providers for 97 patients. Ofthe total attempts, 41.5% were made withoutstopping CCs, and there was no difference inplacement success between attempts with in-terrupted compressions and continuous com-pressions (45% vs. 50%; p = 0.62). There wasno statistically significant difference betweenthe frequency of KV and CMAC placement at-tempts with CCs (47.7% vs. 37.1%; p = 0.27).Though the percentage of successful attemptswith continued CCs appeared higher in theCMAC group (60.9% vs. 38.1%), the differencedid not meet statistical significance (p = 0.13).Conclusion. Overall, CCs were performed on41% of attempts and more frequently, thoughnot statistically significantly, with KV com-pared with CMAC. CMAC appeared to havea higher success rate while CCs are ongoingwhen compared with KV. Though this studydid not specifically determine which VL devicemight better facilitate endotracheal intubation(ETI) with uninterrupted CCs, it did show alarge percentage of successful VL intubationswithout pauses in CC. Future research directlycomparing ETI using VL versus DL may estab-lish a permanent place for VL in out-of-hospitalcardiac arrest.

51. SUCCESS OF OUT-OF-HOSPITAL PEDIATRICENDOTRACHEAL INTUBATIONS PERFORMED BYFLIGHT NURSES AND PARAMEDICS

Sean Button, Christian Martin-Gill, MioaraManole, Francis Guyette, Children’s Hospitalof Pittsburgh

Background. Bag–valve–mask ventilation(BVM) is the preferred method of pediatricventilatory support in the out-of-hospitalsetting. However, there is no consensusregarding the optimal method for pediatricairway management during long transportsperformed by critical care providers withrigorous training and ongoing experience.The purpose of this study was to describethe incidence of intubation success, first-passsuccess, immediate complications, adequacyof ventilation, and correct choice of tube sizefor pediatric intubations performed by flightnurses and paramedics. Methods. We retro-spectively reviewed all intubation attemptsperformed by a large regional helicopteremergency medical service (HEMS) in patientstransported to a single pediatric tertiary carecenter. We included patients less than 18 yearsold intubated by flight nurses and paramedicsfrom January 2003 to December 2011. Werecorded the age, gender, reason for intubation,number of attempts, initial and final end-tidalcarbon dioxide (EtCO2) and oxygen (O2)saturations, and endotracheal tube (ETT)size. We also reviewed the hospital record toidentify immediate complications, includingmissed esophageal intubation, pneumothorax,mainstem intubation, and persistent hypoxia.Results. Intubation was attempted in 196pediatric patients during the study period. Themedian patient age was 7 years (interquartilerange [IQR] 3–13) with a weight of 35 kg(IQR 20–50). The majority of children, 121(63%), were intubated at the scene, while

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


the remainder were intubated at outlyingfacilities. The first-pass success rate was 85.2%(95% confidence interval [CI] 80.2–90.2%).Successful intubations were performed withinthree attempts in 97% of patients (95% CI94.6–99.3%). There was one (0.5%, 95% CI0–1.5%) unrecognized esophageal intubation,zero pneumothoraxes, and 25 (13%, 95% CI8–18%) mainstem intubations identified at thereceiving hospital. Only two patients (1.0%,95% CI 0.4–2.4%) had hypoxia not corrected bythe flight crew (saturation of peripheral oxygen[SpO2] <95%). Hypoventilation (EtCO2 >45mmHg) and hyperventilation (EtCO2 <30mmHg) occurred in 18 (9.2%, 95% CI 5–13%)and 29 (15%, 95% CI 10–20%) patients, respec-tively. ETT size was correctly chosen in 141patients (74%, 95% CI 68–80%). Conclusions.In our series, HEMS nurses and paramedicsperformed successful intubations withoutcomplications in a vast majority of pediatricpatients. Correct calculation of pediatric ETTsize and depth may represent educationalopportunities in our cohort of providers.

52. BARRIERS TO PEDIATRIC VITAL SIGNASSESSMENT BY PREHOSPITAL PROVIDERS

Kathleen Adelgais, Lara Rappaport, JasonKotas, Kevin Waters, Maria Mandt,University of Colorado

Background. Studies demonstrate that incom-plete monitoring of vital signs (VS) by prehos-pital providers (PHPs) is common in children,while barriers to obtaining VS in children byPHPs have not been explored. Our objectivewas to test the content validity, face validity,and internal consistency of a survey designedto identify perceived barriers to obtaining VSin children. Methods. A draft survey was cre-ated after a review of the literature and non-structured interviews with PHP subject matterexperts (SMEs) to develop content and ascer-tain certain beliefs surrounding barriers to VSassessment in children. The survey was thenpilot-tested among additional SMEs for contentvalidity and a convenience sample of PHPs onface validity and internal consistency. We as-sessed content validity with a five-point Lik-ert scale. Cronbach’s alpha tested internal con-sistency for beliefs, including familiarity withpediatric VS norms, utility of VS in manage-ment, and importance of VS in children. Weevaluated the face validity of three case scenar-ios in different age groups depicting an unco-operative child created to evaluate PHP self-efficacy for pediatric VS assessment. Results.The survey was completed by seven SMEs and22 PHPs. The average age was 38.4 years, 78%were male, the median number of years as aPHP was 15, and 90.1% reported practice lo-cation as urban/suburban. Of the SMEs, >90%gave a rating of >3 (“agree”) for barriers andenablers to PHP VS assessment, with no recom-mended additional content. The PHPs agreedon the degree of cooperation for the infantand child age group case scenarios 72.7% and90.0% of the time, respectively, with disagree-ment noted for the toddler case. Cronbach’s al-pha for familiarity with VS norms was 0.72, theutility of VS was 0.76, and the importance ofVS was 0.86. Lack of equipment, patient com-pliance, and low pediatric call volume were re-ported by >80% of the PHPs as the greatest bar-riers to pediatric VS assessment. Conclusion.We have developed a survey tool with fair toexcellent consensus on content and face valid-ity and internal consistency to evaluate self-efficacy and perceived barriers to obtaining VSin children. Additional study into the impact oflocation, education, pediatric call volume, andequipment availability on PHP beliefs aroundVS assessment in children can be furtherexplored.

53. A RETROSPECTIVE ANALYSIS OFCOMPLICATIONS ARISING FROM INTRAOSSEOUSINSERTION DURING OUT-OF-HOSPITALCARDIOPULMONARY RESUSCITATION IN ADULTS

Andrew Wallace, Jose Cabanas, Valerie DeMaio, Brent Myers, University of NorthCarolina

Background. Intraosseous (IO) access is rou-tinely used during resuscitation from out-of-hospital cardiac arrest (OHCA), yet there is alack of data regarding complications in adults.The objective of this study was to identify com-plications from IO insertion in adult patientsresuscitated from OHCA. Methods. This wasa retrospective chart review of patients aged≥16 years admitted to hospital after success-ful resuscitation from OHCA with IO accessin an urban/suburban emergency medical ser-vices (EMS) system (population 900,000) fromJune 1, 2005, to December 31, 2009. Traumacases were excluded. A literature review re-vealed a list of potential IO complications: os-teomyelitis, periostitis, fracture, extravasation,compartment syndrome, cellulitis, skin abscess,ischemia, deep vein thrombosis (DVT), andpulmonary embolus (PE). Data abstracted byone observer to the study database from EMSand hospital records were demographics, med-ical history, IO insertion information, hospi-tal course, and complications. Follow-up phonecalls were made to survivors to screen forcomplications arising after hospital discharge.Analysis involved descriptive statistics with95% confidence intervals (CIs). Results. Thisstudy included 217 adults aged 16 to 98 years,comprising 273 IO insertions. For insertions, 14(5.1%, 95% CI 2.5–7.8%) had a defined com-plication, including compartment syndrome, 1(0.4%); ischemic injury, 1 (0.4%); and extrava-sation, 5 (1.8%). One extravasation was asso-ciated with local necrosis. DVT occurred infive insertions (1.8%) and PE occurred in twoinsertions (0.7%). Conclusion. This is one ofthe first reports on IO complications in adults,and the complication rates reported here arecomparable to the reports from pediatric pop-ulations and show that IO catheters are fa-vorable to traditional intravenous catheters.Additionally, the DVT and PE rates in thisstudy are similar to rates reported for hos-pitalized patients. IO appears to be a saferoute for rapid vascular access in emergencysituations, but further prospective evaluationis necessary to establish true complicationsrates.

54. THE IMPACT OF CPR DURATION ONSURVIVAL TO HOSPITAL DISCHARGE BETWEENINTEGRATED AUTOPULSE CPR AND MANUALCPR DURING OUT-OF-HOSPITAL CARDIACARREST OF PRESUMED CARDIAC ORIGIN

Lars Wik, Jan Olsen, David Persse, FritzSterz, Michael Lozano, Marc Brouwer, MarkWestfall, Chris Souders, Reinhard Malzer,Pierre van Grunsven, David Travis, UlrichHerken, James Brewer, E. Lerner,National Competence Center for EmergencyMedicine

Background. The Circulation Improving Re-suscitation Care (CIRC) trial found equiv-alent survival in out-of-hospital cardiac ar-rest (OHCA) patients who received integratedAutoPulse cardiopulmonary resuscitation (iA-CPR) compared with high-quality manual CPR(M-CPR). We hypothesized that as prehospi-tal CPR time increased, iA-CPR would providea survival benefit when compared with high-quality M-CPR. Methods. A subgroup analysisof the CIRC randomized clinical trial was con-ducted. Patients were included in the CIRC trialif they had an OHCA treated by a participat-

ing emergency medical services (EMS) systemin one of five study communities. Randomiza-tion occurred after manual compressions wereinitiated. Only those patients whose OHCAwas EMS- or bystander-witnessed and had ashockable initial rhythm were included in thisanalysis. Duration of CPR was obtained fromdata recorded by the EMS defibrillator, anddefined as the interval between the time thedefibrillator was turned on and the time re-suscitation was terminated or the time of thefirst documented return of spontaneous circu-lation. Logistic regression was used to modelthe interaction between treatment and lengthof resuscitation and was covariate-adjusted fortrial site and patient age. The primary out-come was survival to hospital discharge. Re-sults. 4,231 subjects were enrolled in the CIRCtrial; 674 patients had witnessed shockable ar-rests. Of those, 621 had complete outcome andduration of CPR data (294 iA-CPR, 327 M-CPR). The logistic model had an overall p-value<0.0001 and a Hosmer-Lemeshow goodness-of-fit p-value of 0.20. The covariate-adjustedodds ratio for survival to hospital discharge inthe iA-CPR arm was 1.49 compared with M-CPR with a p-value = 0.037 and a 95% con-fidence interval of 1.02 to 2.16. The odds ra-tio for survival to hospital discharge in favorof iA-CPR compared with M-CPR increased asthe duration of resuscitation increased. iA-CPRhad a survival benefit compared with M-CPRwhen the resuscitation duration was greaterthan 10 minutes. Conclusion. Compared withhigh-quality M-CPR, iA-CPR resulted in a sta-tistically significant improvement in survival tohospital discharge for adult witnessed shock-able OHCA patients with a longer duration ofCPR.

55. AMPLITUDE SPECTRUM AREA–BASEDDEFIBRILLATION DECISION GREATLY IMPROVESSHOCK SUCCESS RATE AND ACCURACY DURINGPREHOSPITAL CARDIOPULMONARYRESUSCITATION

Giuseppe Ristagno, Weilun Quan, GaryFreeman, Mario Negri Institute forPharmacological Research

Background. We retrospectively evaluated thecapability of amplitude spectrum area (AMSA)to predict the likelihood that a defibrillation(DF) would restore a perfusing rhythm duringcardiopulmonary resuscitation (CPR) in humanvictims of out-of-hospital cardiac arrest. Wehypothesized that threshold values of AMSAcould be identified to be used as a decisiontool for CPR intervention, i.e., chest compres-sion or DF, and such AMSA threshold wouldincrease DF success rate and accuracy. Meth-ods. Electrocardiographic (ECG) data, includ-ing 1,410 DF attempts, were obtained from 748cardiac arrest patients from multiple areas inthe United States. A 4-sec ECG window endingat 0.5 sec before DF was analyzed and AMSAwas calculated prior to DF attempts. Success-ful DF was defined as return of an organizedrhythm within 60 sec. For first and subsequentDFs, AMSA threshold values were analyzedwith regard to their ability to discriminateamong successful and unsuccessful DFs. TheDF performance was then compared betweentwo DF decision algorithms with and withoutAMSA threshold, respectively. Results. A to-tal of 1,221 qualified DF events from 607 pa-tients, with 578 first DF attempts and 543 subse-quent ones for DF-resistant VF, were includedin the analyses. For the DF decision algorithmwithout AMSA, the DF success rate and ac-curacy were 27% for the first DFs and 9% forthe subsequent DFs, respectively. An optimizedAMSA threshold was found to be 14 mV-Hz forfirst DFs and 12 mV-Hz for subsequent ones.

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


Incorporation of these AMSA thresholds into aDF decision algorithm increased both DF suc-cess rate and accuracy. The first DF achieveda DF success rate of 54% and an accuracy of75%, while subsequent DF attempts achieved aDF success rate of 42% and an accuracy of 89%.These results translated into an increase of DFsuccess rate and accuracy by 100% and 180%for the first DFs, and 360% and 880% for thesubsequent DFs. Conclusions. In this popula-tion, a defibrillation decision algorithm incor-porating an AMSA threshold was confirmed tobe capable to predict DF success with high ac-curacy. An AMSA-based DF decision algorithmtherefore should be a useful tool to guide CPRinterventions.

56. A COMPARISON OF COMPRESSION-RELATEDINJURIES IN SURVIVORS OF OUT-OF-HOSPITALCARDIAC ARREST TREATED WITH MANUAL VS.MECHANICAL CHEST COMPRESSION

Lori Boland, Paul Satterlee, JonathanHokanson, Craig Strauss, Dana Yost,Allina Health

Background. Compressing the chest dur-ing cardiopulmonary resuscitation can resultin injury. Previous reports on compression-related injuries in out-of-hospital cardiac ar-rest (OHCA) patients treated with manual vs.mechanical chest compression have relied ex-clusively on postmortem data. The purpose ofthis work was to examine the extent of injuryconferred by manual vs. mechanical prehos-pital chest compressions among survivors ofOHCA. Methods. A retrospective cohort studywas conducted among survivors of nontrau-matic OHCA who were discharged from hos-pitals belonging to a single health system be-tween January 2009 and May 2012. Cases wereeligible if the patient had received prehospi-tal compressions from an emergency medicalservices (EMS) provider. One EMS provider inthe area was using the LUCASTM device asstandard equipment for compression, while theremaining providers primarily used manualcompression. Hospital records were reviewedfor injuries documented during the postarresthospitalization that likely resulted from com-pressions. Information about prehospital carewas abstracted from EMS run sheets indepen-dently. Results. Among 117 eligible patients,78 (67%) received manual compressions onlyand 39 (33%) received compressions predom-inantly with LUCAS. Thirteen injuries wereidentified in 10 unique patients (8.5%; 8 male,2 female). The most common injuries were ribfractures (4/13; 31%) and hemorrhage (3/13;23%). Imaging in the manual vs. mechanicalgroups was comparable, as was the preva-lence of injury (8% vs. 10%, respectively, p= 0.64). When compressions were performedfor more than 10 minutes, injuries were morecommon (15% vs. 2% when compressions <10minutes; p = 0.012), but the prevalence ofinjury remained similar with manual (17%)vs. mechanical (14%) compression (p = 0.76).The mean length of stay for those with andwithout injuries was 12.6 and 10.9 days, re-spectively (p = 0.42). Conclusion. In this co-hort of OHCA survivors, longer duration ofcompressions was crudely associated with ahigher frequency of injury, while the use of me-chanical compression was not. Compression-induced injuries are likely underestimated inthis study as compared with autopsy stud-ies, but these data suggest that injuries in-curred among OHCA survivors are largelyunattributable to mechanical compression andmay be insignificant in terms of length ofrecovery.

57. A NOVEL VIDEO CONTENT ANALYSIS SYSTEMFOR INTERACTIVE VIDEO LARYNGOSCOPY

Jestin Carlson, Samarjit Das, Fernando De laTorre, Adam Frisch, Francis Guyette, JessicaHodgins, Donald Yealy,Saint Vincent Health Center

Background. A key step in successful endotra-cheal intubation (ETI) is obtaining a satisfactoryview of the glottic opening. To provide objec-tive feedback about the quality of glottic visu-alization, we developed a computer vision al-gorithm that can automatically detect the glot-tic opening using machine learning techniqueson videos collected during ETI. We sought totest the performance of this algorithm usingvideo laryngoscopy. Methods. Seven partici-pants trained in ETI performed 10 attemptseach on a manikin using a video laryngoscope(C-MAC Karl Storz Corp., El Segundo, CA).Using 3-second epochs, we recorded either thepresence or absence of the glottic opening onthe screen. Data from the first five trials for eachsubject were used to train a computer visionalgorithm, using image features known as thescale invariant feature transformation (SIFT),to represent either the presence or absence ofthe glottic opening (derivation cohort). Train-ing used five different types of common clas-sifiers: k-nearest neighbor (k-NN), support vec-tor machine (SVM), decision trees, kernel logis-tic regression, and neural networks. We treatedthe presence of the glottic opening lasting lessthan 1 second as spurious, and these were re-moved from the analysis to further refine theclassifiers. The five remaining trials per sub-ject were used for testing the accuracy of thealgorithm (validation cohort). Results. Becauseof difference in the length of time required forthe 70 ETI attempts, 1,145 time periods werein the derivation cohort and 1,320 time peri-ods were in the validation cohort. All classifiershad robust accuracy for detecting the glotticopening: k-NN 77%, SVM 77%, decision trees74%, kernel logistic regression 76%, and neu-ral networks 75%. The accuracy for automati-cally detecting the presence of glottic openingafter postprocessing was 81% for both k-NNand SVM. Conclusion. We found that a com-puter algorithm can be trained to identify air-way anatomy with good accuracy. This may al-low creation of an interactive laryngoscopy toolto provide procedural guidance and objectiveskill assessments.

58. ALTERNATIVE AIRWAY USE BY PARAMEDICSAFTER VIDEO LARYNGOSCOPE FAILURE

Zabrina Evens, Sandi Wewerka, KentGriffith, Joshua Salzman, Samantha Kealey,Aaron Burnett, Ralph Frascone,Regions Hospital

Background. Prehospital use of video laryn-goscopes (VLs) remains rare. We analyzedthe type, frequency, and overall success ratesfor alternative airway devices used when VLfailed during a prospective prehospital re-search study. Methods. This was a post hocanalysis from a multisite, prospective, nonran-domized, crossover trial comparing the place-ment success rates of two VLs (Storz CMAC,Karl Storz; King VISION [KV], King Sys-tems). Following failed airway managementwith a VL device, providers were allowed tochoose their alternative device (ETI via di-rect laryngscopy [DL], King, Combitube, orbag–valve–mask). Descriptive analyses werecompleted for patient, provider, and alterna-tive airway device variables. The frequencies ofachieving a good view (Cormack-Lehane score[CLS] ≤2) during failed VL attempts were com-pared between VL devices using a chi-square

test. Chi-square was also used to compare theCLS achieved during direct laryngoscopy (DL)with the CLS achieved during failed VL man-agement. Results. Between October 2011 andAugust 2012, there were 31 failed VL place-ments in the 94 patients treated. There wereno significant demographic differences in pa-tient or provider characteristics. Providers ex-perienced more failures with the KV devicethan with the CMAC (47.5% vs. 21.1%; p <0.006). The vast majority of the 31 VL fail-ures were managed with DL (84%). The com-bined DL success rate was high (88.4%), anddid not differ statistically between treatmentgroup (CMAC = 77.7%, KV = 94.1%; p = 0.24).A good view was achieved in 53% of the totalVL placement attempts (CMAC = 64%, KV =45.5%; p = 0.15). When the providers movedto DL, 64% of the total attempts had a goodview (CMAC = 46%, KV = 68%, p = 0.20).There was no significant difference in achiev-ing a good view between VL and DL (p = 0.40).Conclusion. Patients who failed VL were mostfrequently managed by DL, and managed suc-cessfully. CMAC offered a better but not sta-tistically better view than KV. When providersmoved from CMAC to DL, they reported aworse view, but it did not impact success rate.When providers moved from KV to DL, theyreported a better view, raising major concernabout the device as a prehospital VL.

59. DERIVATION OF A PREHOSPITAL DIFFICULTAIRWAY IDENTIFICATION TOOL: THE PREDAITSTUDY

Jestin Carlson, Francis Guyette, DavidHostler, Mark Pinchalk, ChristianMartin-Gill,Saint Vincent Health Center

Background. Endotracheal intubation (ETI) inthe prehospital setting is a challenging skillwith variable success rates. Several modelsare available to predict a difficult airway inthe hospital, but there are few options forpredicting a difficult airway in the field. Wesought to derive a model for predicting a dif-ficult airway in the prehospital setting. Meth-ods. We prospectively collected data on air-way management using an electronic data-collection form that was automatically gener-ated for all cases of airway intervention as partof the patient care record (emsCharts, Pitts-burgh, PA). Eighteen data elements were de-rived from previous studies of ETI and in-cluded patient demographics, circumstancessurrounding the ETI attempt, methods used,and difficulties encountered (emesis, cervicalcollar placement, trauma to the neck, etc.). Ourprimary outcome was first-pass success, de-fined as successfully passing the endotrachealtube through the vocal cords during the firstplacement of the laryngoscope into the oralcavity. We used logistic regression to derive amultivariate model to predict a difficult pre-hospital airway, defined as >1 attempt requiredto secure the airway. Results. We collecteddata from 16 EMS agencies over 18 months,which included 477 cases where airway inter-vention was performed. First-pass success oc-curred in 208 of 367 cases (57%) of ETI. Vari-ables with p-values <0.1 were included in themultivariate model: inability to manipulate theneck (odds ratio [OR] 0.42; confidence inter-val [CI] 0.18–0.94; p = 0.036), emesis (OR 0.48;CI 0.29–0.77; p = 0.002), and inability to pal-pate landmarks of the neck (OR 0.37; CI 0.14–1;p = 0.057). This three-variable model hadgood fit (Hosmer-Lemeshow = 0.55) with ad-equate discrimination (area under the receiver-operating curve = 0.78). Conclusions. We havederived a simple model to predict a diffi-cult airway (those requiring >1 ETI attempt)

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


prior to attempting ETI in the prehospital set-ting, including inability to manipulate the neck,emesis, and inability to palpate landmarks ofthe neck. While further study will be neededto validate this model, providers encounteringthese patients should consider using alternatetechniques or devices to improve first-pass air-way placement.

60. COMPARATIVE STUDY OF THE EFFECTIVENESSOF THE INTRAORAL MASK VERSUS TRADITIONALBAG–VALVE–MASK VENTILATION IN A LIGHTLYEMBALMED CADAVER MODEL

Richard Walker, Darcy Thompson, M.Hallbeck, Carol Lomneth, BernadetteMcCrory, Bethany Lowndes, Lina Lander,Michael Luethge, Michael Miller, WilliamRaynovich, Michael Wadman,UNMC

Background. Bag–valve–mask (BVM) venti-lation is more effective using a two-persontechnique. A novel intraoral mask (IOM)(NuMask R©) may allow effective BVM ventila-tion with a one-person technique.The objectiveof this study was to compare standard BVMventilation with the IOM measuring deliveredtidal volumes, leakages, and minute ventila-tions provided. Methods. This was a prospec-tive, randomized, nonblinded trial using alightly embalmed cadaver. The prewarmed ca-daver was intubated and attached to a ven-tilator. Lungs were expanded using increas-ing peak end-expiratory pressure (PEEP) un-til breath sounds could be heard bilaterally inthe bases along the anterior axillary line, af-ter which the cadaver was extubated. Subjectswere randomly assigned to either the IOM orthe BVM. Subjects watched a brief instructionalvideo and were permitted to ask questions onthe use and technique of each mask. The maskswere connected to the ventilator and the par-ticipants were asked to maintain a seal withthe nondominant hand for 2 minutes. The ven-tilator delivered 20 breaths at three differentset tidal volumes of 500, 750, and 1,000 mL,with a 1-minute rest between each set tidal vol-ume. The minute ventilation, tidal volume de-livered (TVi), and tidal volume received (TVo)were recorded for every breath. Leakage (inpercentage) was calculated for each breath us-ing the formula leak = TVo / TVi. Results. The27 subjects conducted over 3,204 ventilations.The TVo at the respective set tidal volumes of500, 750, and 1,000 mL for the BVM comparedwith the IOM was 61 versus 301 mL (p < 0.001),140 versus 498 mL (p < 0.001), and 304 versus647 mL (p < 0.001). The leakage at the respec-tive volumes for the BVM compared with theIOM was 88% versus 38% (p < 0.001), 81% ver-sus 32% (p < 0.001), and 66% versus 31% (p <0.001). The minute ventilation at the respectivevolumes for the BVM compared with the IOMwas 0.68 versus 2.89 L/min (p < 0.001), 1.37versus 4.51 L/min (p < 0.001), and 2.75 versus5.99 L/min (p < 0.001). Conclusions. The in-traoral mask (NuMask) was superior to the tra-ditional BVM in received tidal volume, leakage,and minute ventilation at all set tidal volumesin a lightly embalmed cadaver model.

61. INCIDENCE AND OUTCOMES OF UNRESOLVEDPREHOSPITAL REARREST IN EMS-TREATED CASESOF OUT-OF-HOSPITAL CARDIAC ARREST

David Salcido, Allison Koller, JamesMenegazzi, Matthew Sundermann, CliftonCallaway,University of Pittsburgh

Background. The phenomenon of rearrest(RA)—loss of pulses after successful return ofspontaneous circulation (ROSC)—is of inter-est in resuscitation research because even tran-sient loss of pulses prior to hospital arrivalmay have a detrimental impact on patient out-

comes. Patients who experience RA may alsodemonstrate specific pathology, the identifica-tion and treatment of which may lead to bet-ter outcomes. Our objective was to estimate theincidence and outcomes of one manifestationof RA, unresolved prehospital RA (UP-RA),wherein pulses are not reestablished followingat least one RA prior to arrival at the emer-gency department (ED). Methods. Case dataspanning 2006–2008 were obtained from theResuscitation Outcomes Consortium, a mul-tisite clinical research consortium with car-diac arrest surveillance programs in 10 sitesin North America. Nontraumatic emergencymedical services (EMS)-treated cases of out-of-hospital cardiac arrest (OHCA) with any in-stance of prehospital ROSC were included. Pre-hospital ROSC events, patient vital status atED admission, survival to hospital discharge,patient demographics, and ancillary resusci-tation variables were ascertained through re-view of paramedic-generated patient care re-ports (PCRs), defibrillator data downloads, andhospital records. Prehospital ROSC was de-fined as a detectible return of pulses resulting inan obvious suspension of cardiopulmonary re-suscitation. UP-RA status was assigned to anypatient with prehospital ROSC who did nothave pulses upon ED arrival. Results. Out of18,937 cases of OHCA across all sites, therewere 11,456 (60.5%) EMS-treated cases. Prehos-pital ROSC was found in 4,609 (40.2%) cases.Of these cases, mean (± standard deviation)patient age was 63.7 (±17) years, 37.1% pa-tients were female, 21.5% arrests occurred inpublic, 13.5% arrests were EMS-witnessed, andoverall survival to hospital discharge was 28%.Vital status at the ED was available for 3,116(67.6%) cases, of which UP-RA was present in473 (15.2%). Survival was 7.8% in cases withUP-RA, compared with 33.3% in cases without,and UP-RA was directly associated with deathprior to hospital discharge (odds ratio: 6.14,confidence interval: 4.31–8.75, p < 0.001). Con-clusion. When characterized as an irreversibleevent prior to ED arrival, the incidence of RA isrelatively uncommon but strongly predictive ofnonsurvival at hospital discharge.

62. LACK OF CORRELATION BETWEEN CHESTCOMPRESSION FRACTIONS OVER 0.4 ANDSURVIVAL

Robert Walker, Tom Brouwer, Fred Chapman,Rudolph Koster, Physio-Control

Background. Chest compression fraction (CCF)is a commonly measured cardiopulmonary re-suscitation (CPR) quality metric, but it remainsunknown whether there is an optimum CCF. Inventricular fibrillation (VF) patients, one studymeasured CCF over the first 1–3 minutes andfound worse survival for CCFs less than orequal to 0.2 than for higher CCFs. We analyzeda large resuscitation cohort and hypothesizedthat higher CCF would be associated with in-creased survival to hospital discharge and goodneurologic outcome throughout the range ofobserved CCFs. Methods. In an observationalstudy of prospectively collected data, we en-rolled all patients with ventricular tachycardia(VT)/VF out-of-hospital cardiac arrest duringcalendar year 2009 in five emergency medicalservices systems. We excluded resuscitationslasting <5 minutes, and analyzed electrocardio-graphic and transthoracic impedance record-ings to measure CCF over the first 5 minutes(CCF-5) and the entire resuscitation (CCF-all).Results. Of 263 patients, 72 patients (27%) sur-vived to discharge, 65 patients (25%) with goodneurologic outcome (cerebral performance cat-egory 1 or 2). CCF-5 was median (interquartilerange) 0.74 (0.63, 0.83), and CCF-all was 0.80(0.73, 0.85). No cases had a CCF-5 or CCF-allless than or equal to 0.2. For CCF-5 categories

of 0.21–0.40, 0.41–0.60, 0.61–0.80, and 0.81–1.0,survival rates (95% confidence intervals) were57% (20–88%, n = 7), 27% (15–42%, n = 45),24% (18–33%, n = 131), and 30% (21–41%, n= 80), respectively. Results for CCF-all weresimilar to results for CCF-5. In a multivariateUtstein-adjusted regression, there were no sig-nificant associations between each 0.1 increasein CCF and survival to discharge (CCF-5: oddsratio [OR] = 1.04 [0.83, 1.29]; CCF-all: OR =0.96 [0.72, 1.28]) or good neurologic outcome(CCF-5: OR = 0.99 [0.79, 1.24]; CCF-all: OR= 0.91 [0.68, 1.22]). Conclusions. In this co-hort, almost all patients received a chest com-pression fraction between 0.4 and 1.0, and wefound no significant association between in-creasing CCF and improved outcomes over thisrange.

63. MECHANICAL CHEST COMPRESSION AS ABRIDGE TO EXTRACORPOREAL MEMBRANOUSOXYGENATION (ECMO) IN A CLINICALLYREALISTIC PORCINE MODEL OFOUT-OF-HOSPITAL CARDIAC ARREST

Joshua Reynolds, David Salcido, MatthewSunderman, Allison Koller, JamesMenegazzi,University of Pittsburgh

Background. Extracorporeal membranous oxy-genation (ECMO) is an emerging tool in the re-suscitation armamentarium. One barrier to im-plementation in out-of-hospital cardiac arrest(OHCA) is the delay to the initiation of ECMO.Prehospital providers could play key roles inidentifying appropriate candidates, maintain-ing blood flow, and expediting transport to anECMO-capable center. We evaluated the fea-sibility of prolonged mechanical chest com-pression resuscitation followed by ECMO in aporcine model of ventricular fibrillation (VF)utilizing clinically realistic durations of delay toECMO. Methods. We anesthetized and instru-mented eight female swine (mean mass 31.9 kg)with 14-French and 18-French catheters via cut-down in the right femoral artery and exter-nal jugular vein, respectively. Catheters wereflushed with heparinized saline and clampeduntil initiation of ECMO. After 8 (n = 4) or 15(n = 4) minutes of untreated VF, the animals re-ceived 30, 40, 50, and 60 minutes of mechanicalchest compressions. All animals received drugs(0.6 U/kg vasopressin, 0.1 mg/kg epinephrine,0.1 mg/kg propranolol) after 5 minutes ofcardiopulmonary resuscitation (CPR). ECMOreperfusion was achieved at 3 L/min witha preheparinized centrifugal circuit. activatedclotting time (ACT) was maintained above 250seconds. After 15 minutes of reperfusion withECMO, the animals were eligible for defibrilla-tion attempts with 150 J based on the electro-cardiogram. ECMO flow was 3 L/min for up totwo hours and then was reduced to 1.5 L/minfor up to two additional hours (four hourstotal) before weaning. The primary outcomewas ROSC, defined as an organized rhythmwith systolic blood pressure (SBP) >80 mmHg;secondary outcomes were one-hour survivalon ECMO (SURV) and one-hour survival afterweaning off ECMO (WEAN). Results. Given asCPR duration in minutes (outcome). For 8 min-utes VF: 30 (WEAN), 40 (WEAN), 50 (ROSC),60 (SURV). For 15 minutes VF: 30 (SURV), 40(no ROSC), 50 (no ROSC), 60 (no ROSC). An-imals without ROSC spontaneously convertedto an organized rhythm, but had SBPs <80mmHg. The most prolonged successful resus-citation occurred one hour 27 minutes after VFonset. Conclusions. Mechanical chest compres-sion may be a suitable bridge to ECMO, whichmay be a feasible resuscitation tool for selectOHCA patients. Emergency medical serviceswill play an important role as this technologyadvances.

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


64. SURVIVAL RATES FOR OUT-OF-HOSPITALCARDIAC ARREST PATIENTS TRANSPORTEDWITHOUT A PREHOSPITAL RETURN OFSPONTANEOUS CIRCULATION

Ian Drennan, Steve Lin, Daniel Sidalak,Laurie Morrison,University of Toronto, St. Michael’s Hospital

Background. The implementation of the pre-hospital universal termination-of-resuscitation(TOR) guideline (terminate without a prehos-pital return of spontaneous circulation [ROSC]AND not emergency medical services [EMS]-witnessed AND no shock provided) has re-duced the transport rate to hospital. Some ser-vices currently use the absence of prehospitalROSC as the single criterion to terminate resus-citation, which may deny transport to potentialsurvivors. Few studies have examined the pa-tient characteristics and prehospital factors as-sociated with survival in OHCA patients trans-ported without a prehospital ROSC. This studyaims to describe patient characteristics and pre-hospital factors associated with survival to hos-pital discharge in nontraumatic, adult OHCApatients transported to hospital without a pre-hospital ROSC. Methods. This was a retro-spective observational study of consecutivelytreated OHCA patients without a prehospitalROSC who met the universal TOR guideline fortransport to hospital with ongoing resuscita-tion. Bivariate analyses were used to determineassociations of patient characteristics and pre-hospital variables with survival. Results. Therewere 19,571 OHCA patients treated by EMS, ofwhom 3,208 (16.4%) did not have a prehospi-tal ROSC but met the universal TOR guidelinesfor transport to hospital with ongoing resusci-tation (68.8% shocked, 25.2% EMS-witnessed,6.0% both). Of these patients, 62 (1.9%) sur-vived to hospital discharge. Survivors wereyounger (55.6 yr ± 15.1 vs. 65.4 yr ± 15.6, p< 0.01), had initial shockable rhythms (75.4%vs. 52.6%, p < 0.01), had no advanced airway(53.6% vs. 79.2%, p < 0.01), and were trans-ported to academic hospitals (30.7% vs. 17.6%,p = 0.01). No differences were noted in malegender (71.0% vs. 74.0%, p = 0.56), bystander-witnessed arrests (38.7% vs. 42.7%, p = 0.17),public location (36.1% vs. 28.7%, p = 0.25), by-stander cardiopulmonary resuscitation (47.2%vs. 42.8%, p = 0.60), EMS response (6.5 min±3.3 vs. 6.8 min ±3.5, p = 0.47), or advancedlife support provider (71.0% vs. 79.7%, p =0.11). Conclusion. In OHCA patients without aprehospital ROSC who met the universal TORguideline for transport with ongoing resusci-tation, 2% survived to hospital discharge. Thisrate of survival is higher than that defined asmedically futile (<1%) and suggests that ter-mination of resuscitation should not be basedon the absence of prehospital ROSC alone. Sur-vival was associated with younger age, initialshockable rhythms, no advanced airway, andacademic hospitals.

65. TRANSFER OF PATIENTS WITH ST-SEGMENTELEVATION MYOCARDIAL INFARCTION FORPRIMARY PERCUTANEOUS CORONARYINTERVENTION (PRIMARY PCI): IDENTIFYINGSOURCES OF DELAY IN THE FIRST HOSPITALCENTER

Kevin Brown, Eli Segal, Laurie Lambert,Celine Carroll, Dave Ross, Sebastien Maire,Lucy Boothroyd, Peter Bogaty,Institut National d’Excellence en Sante et enServices Sociaux

Introduction. Patients with ST-segment el-evation myocardial infarction (STEMI) whopresent to a center without capability to per-form primary percutaneous coronary interven-

tion (PCI) as reperfusion treatment and are thentransferred to a PCI center rarely achieve thetimely benchmark standard of a first “door-in to door-out” (DIDO) interval <30 minutes.Our objective was to characterize the factorsassociated with DIDO delay using data froma provincewide STEMI field evaluation. Meth-ods. Medical chart data of STEMI patients whounderwent interhospital transfer for primaryPCI were linked to the provincial emergencycare database. Using these two data sources,we recorded seven relevant time points: ini-tial triage (door-in time), first electrocardio-gram (ECG), transfer activation, arrival of am-bulance, departure of ambulance from initialcenter (door-out time), arrival of ambulance atPCI center, and time of primary PCI. Mediantimes with interquartile range (IQR) were mea-sured. Two coordinating center cardiologists re-viewed all presenting ECGs. Results. We iden-tified 990 interhospital transfers, of which 745(75%) could be linked to the emergency caredatabase. Time from first center triage to pri-mary PCI was 111 minutes (IQR: 91–146); DIDOtime was 50 (IQR: 35–79). Only 17.2% of pa-tients had a DIDO delay <30 minutes. Timefrom triage to first ECG was 7 minutes (IQR:2–15); from first ECG to transfer activation, 21minutes (IQR: 13–42); and from ambulance ar-rival to departure, 15 minutes (IQR: 11–19).Transport time was 27 minutes (IQR: 16–44).Median DIDO was shorter in patients trans-ported to the first center by ambulance (45 ver-sus 56 min, p < 0.001), especially when thesame ambulance was used for patient trans-fer (38 versus 52 min, p < 0.001). Older age,female gender, left bundle branch block, un-clear STEMI (cardiologist disagreement on thepresence of STEMI), and presentation outsideworking hours and on weekends were asso-ciated with longer DIDO times. Conclusions.Benchmark DIDO delays for transferred STEMIpatients in Quebec are rarely achieved. Timefrom ECG to transfer activation and from am-bulance arrival to departure were major con-tributors to DIDO delay. For patients arrivingat the initial center by ambulance, use of thesame ambulance for transfer was associatedwith shorter DIDO delay. These findings iden-tify where there is the most potential to reduceDIDO delays.

66. AMBULANCE USE VERSUS NONUSE INPATIENTS WITH ST-SEGMENT ELEVATIONMYOCARDIAL INFARCTION: A POWERFULPROGNOSTICATOR OF MORTALITY RISK

Lucy Boothroyd, Yongling Xiao, LaurieLambert, Eli Segal, Dave Ross, SebastienMaire, Peter Bogaty, Institut Nationald’Excellence en Sante et en Services Sociaux

Background. In a Canadian provincewide sys-tematic evaluation of care of patients withST-segment elevation myocardial infarction(STEMI), we compared the characteristics andsurvival of ambulance users and nonusers.Methods. All 82 acute care Quebec hospitalsthat treated at least 30 acute myocardial in-farctions (AMIs) participated in a six-monthevaluation in 2008–2009. Medical record librar-ians abstracted hospital chart data for AMI pa-tients who presented to the emergency room(ER) with acute symptoms. STEMI was con-firmed by core laboratory interpretation of thefirst electrocardiogram (ECG). Linkage to vi-tal statistics and hospital discharge databasesprovided data on survival and on diagnoses inthe previous five years, respectively. Results.Of 1,957 STEMI patients, 1,222 (62%) arrivedby ambulance. Compared with nonusers, am-bulance users were significantly older, morelikely to be female, and more likely to haveprevious myocardial infarction, peripheral vas-

cular disease, heart failure, hypertension, ar-rhythmia, renal disease, and chronic obstruc-tive pulmonary disease. At first medical con-tact, ambulance users were more likely to havelow systolic pressure (<90 mmHg), abnormalheart rate (<60 or >100 beats/min), and higherThrombolysis In Myocardial Infarction (TIMI)risk index. Ambulance users had shorter de-lays between 1) symptom onset and ER ar-rival; 2) ER arrival and first ECG; and 3) ER ar-rival and reperfusion treatment (thrombolysisor primary coronary intervention [PCI]). Theproportion receiving thrombolysis or sent forPCI within four hours was lower in the am-bulance group (78% vs. 83%; p = 0.01). Mor-tality (in-hospital, 30 days, 1 year) was higherfor ambulance users (9.1%, 12.6%, 18.7%, re-spectively) than nonusers (2.9%, 3.8%, 7.1%, re-spectively; p < 0.001 for all). Higher mortalitypersisted (hazard ratio [HR] = 1.5 [95% confi-dence interval (CI): 1.3–1.8]) after adjusting forgender, anterior AMI, TIMI risk index (whichincludes age), comorbidities, whether the pa-tient received thrombolysis or was sent for PCIwithin four hours or not, and the symptoms-to-ER delay (within 3 hours vs. >3 hours). Re-sults were similar when only treated patients(N = 1,570) were analyzed and timeliness ofreperfusion was included (HR = 1.5 [95% CI:1.0–2.4]). Conclusions. In this provincewideevaluation, ambulance users with STEMI wereolder and sicker than nonusers. Despite fasterreperfusion treatment overall, their mortalitywas higher. Ambulance use is a simple andpowerful marker of mortality in patients withSTEMI.

67. PREHOSPITAL ADMINISTRATION OF MULTIPLESIMULTANEOUS NITROGLYCERINE SUBLINGUALTABS RARELY CAUSES HYPOTENSION

Brian Clemency, Gina Tundo, JeffreyThompson, Heather Lindstrom,University at Buffalo, State University of NewYork

Background. High-dose intravenous nitroglyc-erin is a common in-hospital treatment for res-piratory distress due to congestive heart fail-ure (CHF) with hypertension. Intravenous ni-troglycerin administration is impractical in theprehospital setting. In 2011, a new regionalemergency medical services (EMS) protocolwas introduced allowing advanced providersto treat CHF with oral nitroglycerin. Patientswere treated with two sublingual tabs (0.8 mg)when systolic blood pressure (SBP) was >160mmHg or three sublingual tabs (1.2 mg) whenSBP was >200 mmHg every 5 minutes asneeded. To assess the protocol’s safety, westudied the incidence of hypotension follow-ing prehospital administration of multiple si-multaneous nitroglycerin (MSN) tabs by EMSproviders. Methods. A retrospective case re-view of records from a single commercial EMSagency over a six-month period (January–June2012). Cases with at least one administration ofMSN were reviewed. For each administration,the first documented vital signs before and af-ter administration were compared. Administra-tions were excluded if they were missing pre-or postadministration vital signs. Blood pres-sure was measured in mmHg. Results. Onehundred cases had at least one MSN admin-istration by an advanced provider during thestudy period. Twenty-five cases were excludedbecause of incomplete vital signs. Seventy-fivecases with 95 individual MSN administrationswere included for analysis. There were 65 ad-ministrations of two tabs, 29 administrations ofthree tabs, and one administration of four tabs.The mean change in SBP following MSN was–14.7 (standard deviation 30.7; range +59 to–132). Three administrations had documented

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


systolic hypotension in the postadministrationvital signs (97/71, 78/50, and 66/47 mmHg).All three patients were over 65 years old, wereadministered two tabs, had documented im-proved respiratory status, and had repeat SBPof at least 100 mmHg. The incidence of hy-potension following MSN administration was3.2%. Conclusion. Hypotension was rare andself-limited in this sample of prehospital pa-tients receiving MSN.

68. THE INCIDENCE AND CHARACTER OF STOKEMIMICS IN PATIENTS FLOWN TO A TERTIARYSTROKE CENTER

Laura Thompson, Seth Ritter, ChristianMartin-Gill, Francis Guyette,University of Pittsburgh

Background. Rapid clinical diagnosis of strokeis challenging for physicians and prehospitalproviders alike. However, identification andtreatment of stroke followed by rapid trans-fer of patients to a stroke center are neces-sary to improve morbidity and mortality. Theaim was to describe the incidence and charac-ter of conditions that mimic stroke in a pop-ulation of patients transported by a helicopteremergency medical service. Methods. We con-ducted a retrospective chart review of patientswho were transferred using a regional criticalcare transport to an academic medical centerwith a stroke team. We included patients trans-ported with a presumed diagnosis of strokefrom 2007 to 2010. We performed probabilis-tic linkage to associate prehospital records withinpatient data to compare patients diagnosedwith stroke syndromes vs. those who werediagnosed with conditions that mimic stroke.Results. A total of 1,439 patients were deter-mined to have stroke-like symptoms by pre-hospital providers using the Cincinnati StrokeScale. The average age for our study group was63.9 years, and the study included 51% men. Of1,439 subjects, 1,082 (75%, confidence interval[CI] 73–77%) were diagnosed with cerebrovas-cular accident (888, 62%; CI 59–65%), hemor-rhagic stroke (57, 4%; CI 3–5%), or transient is-chemic attack (137, 10%; CI 8–12%). These pa-tients were excluded from the mimic analysis.We were unable to match or obtain dischargediagnoses on 26 patient charts. A total of 332patients with stroke mimics were analyzed inour final analysis. Among stroke mimics, themost common diagnoses overall were seizure(86, 26%; CI 21–30%), headache (43, 13%; CI9–17%), and brain tumor (21, 6%; CI 3–9%). Hy-poglycemia (10, 3%; CI 1–5%) and intoxication(7, 2%; CI 0.5–4%) were rare mimics of stroke inthis population. Conclusion. Stroke mimic di-agnoses were widely varied among this pop-ulation. Many of the patients with mimics re-quired advanced care, and the current systemof triage to air did not result in significant over-triage. The most common stroke mimics wereseizure, tumor, and headache.

69. CHARACTERISTICS OF NONTRAUMA SCENEFLIGHTS FOR AIR MEDICAL TRANSPORT

Margaret Gluntz, Erica Fletcher, HowardWerman, Lara McKenzie,Ohio State University

Background. While there is a substantial bodyof literature on the use of direct scene responseto trauma by air medical services, little isknown about the use of air medical transporta-tion for patients with nontraumatic medicalemergencies. This study describes the practicesof air medical transport programs with respectto nontrauma scene responses in several dis-tinct geographic areas throughout the UnitedStates and Canada. Methods. A descriptive,retrospective study was conducted utilizingdata from five air medical transport programs.

Flight information and patient demographicdata were collected for all nontrauma sceneflights from 2008 to 2010. Descriptive statis-tics were used to examine indications for trans-port, Glasgow Coma Scale score (GCS) duringtransport, and loaded miles traveled. Results.A total of 1,785 nontrauma scene flights wereevaluated. The percentage of nontrauma flightsvaried between programs from 0% to 43.6% ofscene transports. The most common indicationfor transport was cardiac, non–ST-segment el-evation myocardial infarction, which was seenin 22.9% of medical transports, followed bygeneral medical (15.8%) and neurologic/stroke(11.7%). Cardiac arrest was the primary indi-cation for transport in only 2.5% of flights. Asingle program reported a large percentage ofneurologic/stroke calls (49.4%). The majority ofcases where a GCS was recorded had a scoreof 15 (74.4%). Medical scene flights includedin this study ranged from 0 to 680 miles, witha mean (± standard deviation) flight distanceof 43.4 (± 0.96) miles and a median (interquar-tile range) of 35.0 (22.0) miles. Conclusion. Theuse of air medical transport for nontraumaticemergencies varies considerably between airtransport programs and regions. Differences inthe geographic and economic diversity of a re-gion may affect the number and type of non-trauma air medical transports that take place.This study was limited in that only five pro-grams were included; these were not necessar-ily representative of all the geographic differ-ences that exist throughout the United Statesand Canada. Future research is needed to de-termine which nontraumatic emergencies ben-efit from air transportation from the scene. Fur-thermore, national guidelines defining the ap-propriate use of air medical transportation fornontraumatic scene responses are needed.

70. ASSESSING THE VALIDITY OF THE CINCINNATIPREHOSPITAL STROKE SCALE AND THE MEDICPREHOSPITAL ASSESSMENT FOR CODE STROKE INAN URBAN EMERGENCY MEDICAL SERVICESAGENCY

Jonathan Studnek, Steve Vandeventer, DougSwanson, Andrew Asimos, Jodi Dodds,Mecklenburg EMS Agency

Background. This study assessed the effective-ness of two prehospital stroke screens in cor-rectly classifying patients suspected of havinga stroke. Secondarily, the study assessed differ-ences in the sensitivity and specificity of thetwo screening tools. Methods. This retrospec-tive assessment of the Cincinnati PrehospitalStroke Screen (CPSS) and the Medic Prehospi-tal Assessment for Code Stroke (Med PACS)occurred between March 1, 2011, and Septem-ber 30, 2011, in a single emergency medicalservices (EMS) agency with seven local hospi-tals, all classified as stroke-capable. Data wereobtained from EMS electronic patient care re-ports and Get With The Guidelines R©-Stroke(GWTG-S) registries maintained by the localhealth care systems. Med PACS was devel-oped specifically for the EMS agency understudy by a local team of neurologists, emer-gency physicians, and paramedics. All phys-ical assessment elements of the CPSS wereincluded in Med PACS; items related to pa-tient history were utilized from the Los An-geles Prehospital Stroke Screen. Additionally,gaze and leg motor functions were included inMed PACS. Patients were classified as CPSS-positive or -negative and Med PACS-positiveor -negative if any one of the physical assess-ment findings was present and patient out-come was determined from the GWTG-S reg-istry. Sensitivity and specificity with resultant95% confidence intervals were calculated, andMcNemar’s chi-square analysis was used to as-sess differences in performance. Results. There

were 416 patients enrolled in this study, with186 (44.7%) diagnosed with a stroke. Med PACSscreen classified 293 (70.4%) patients as hav-ing an acute stroke, while CPSS classified 322(77.4%) patients as having an acute stroke. MedPACS scale demonstrated a sensitivity of 0.742(95% confidence interval [CI] 0.672–0.802) ver-sus 0.790 (95% CI 0.723–0.845) for CPSS. Thesensitivity of CPSS was significantly higherthan that of Med PACS, with a difference of0.048 (95% CI 0.009–0.088; p = 0.011). Thespecificities of these two scales were low: MedPACS 0.326 (95% CI 0.267–0.391) versus CPSS0.239 (95% CI 0.187–0.300), with the specificityof Med PACS significantly higher comparedwith CPSS, with a difference of 0.086 (95% CI0.042–0.131; p < 0.001). Conclusions. The re-sults of this study demonstrated no superior-ity between the two scales, with each scale per-forming marginally better in one of the metricsassessed.

71. OFF-DUTY SLEEP DISRUPTION NEGATIVELYIMPACTS BEHAVIORAL HEALTH IN PARAMEDICS

Marc Kruse, Jose Cabanas, Teresa Gardner,Jeffrey Hayes, Paul Parrish, Louis Gonzales,Paul Hinchey, Austin Fire Department &Austin–Travis County EMS

Background. In spite of increased awarenessof the effects of sleep deprivation on job per-formance in emergency medical services (EMS)providers, the impact of off-duty sleep habitsis not well understood. The goal of the cur-rent study was to describe the off-duty sleeppatterns of paramedics and these patterns’association with behavioral health. Methods.An anonymous, voluntary, cross-sectional sur-vey was conducted during scheduled con-tinuing education sessions at a large third-service EMS agency. In addition to provid-ing demographic and off-duty sleep data, theparticipants completed the Beck DepressionInventory-II, PTSD Checklist, Brief ResilienceScale, Sources of Occupational Stress Scale, Pro-fessional Quality of Life Scale, Brief COPE,and measures of alcohol consumption andlife satisfaction. Results. Of the 320 eligibleparamedics, 256 completed the survey (80% re-sponse rate; 81% male; 82% non-Hispanic Cau-casian). Even though on-duty sleep disruptionwas identified as the greatest source of occupa-tional stress, the paramedics reported off-dutysleep for an average of only 6.3 hours (stan-dard deviation [SD] = 1.5; range 3–12). Anal-ysis of variance (ANOVA) revealed a signifi-cant association between depression symptomsand off-duty sleep. Those paramedics reportingno/minimal (53.8%), mild (20.7%), moderate(19.9%), and severe (5.6%) depression symp-toms reported average sleep of 6.6 (SD = 1.4),6.2 (SD = 1.2), 5.9 (SD = 1.4), and 5.5 (SD = 2.3)hours, respectively, F(3, 243) = 5.08, p = 0.002.A series of hierarchical linear regressions alsodemonstrated that after controlling for gender,race/ethnicity, years of EMS experience, andhours worked, reduced off-duty sleep was sig-nificantly associated with higher posttraumaticstress disorder (PTSD) symptoms (p < 0.001),lower life satisfaction (p < 0.001), lower psy-chological resilience (p = 0.004), higher oc-cupational stress (p = 0.004), higher burnout(p = 0.006), more frequent drinking to intox-ication (p = 0.022), and poorer coping strate-gies, including behavioral disengagement (p =0.002), using substances (p = 0.002), and us-ing food (p = 0.026) to deal with stress. Con-clusions. The findings from this cross-sectionalstudy demonstrate a strong association be-tween off-duty sleep disruption and behav-ioral health in paramedics and suggest thatEMS providers might benefit from interven-tions designed to improve sleep quantity andquality.

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


72. THE EPIDEMIOLOGY OF NEEDLEDECOMPRESSION IN A LARGE URBAN EMSSYSTEM

Michael Redlener, David Ben-Eli, DanielGoldberg, Robert Silverman, Pamela Lai,Bradley Kaufman, Douglas Isaacs, JamesBraun, John Freese,St. Luke’s–Roosvelt Medical Center

Background. To describe the utilization of nee-dle decompression (ND) for suspected ten-sion pneumothorax in a large urban emergencymedical services (EMS) system. Methods. Ret-rospective chart review and case series of allNDs performed by a municipal ambulance ser-vice from January 1, 2007, to December 31, 2011.Call type, category of clinical presentation,trauma versus nontrauma, cardiac arrest sta-tus, evidence of clinical improvement, and on-scene and transport times were collected. Re-sults. Over the study period, 140 NDs were per-formed. A total of 99 (71%) NDs were duringtrauma calls (78 [55.7%] penetrating trauma, 21[15%] blunt trauma) and 41 (29%) were dur-ing medical calls (22 [15.7%] in medical ar-rests, three [2.1%] had spontaneous pneumoth-orax, five [3.6%] had an iatrogenic cause, and11 [7.9%] had an unknown cause). Thirty (21%)of the overall patients were female. The over-all median age was 42 years. The median ageof patients in the trauma group was 26 yearsand in the medical group was 72.5 years (p< 0.01). A total of 69 (49.3%) experienced car-diopulmonary arrest prior to or during theEMS call (36/69 in trauma, 33/69 in medicalgroup), five achieving return of spontaneouscirculation (ROSC) after ND, with four main-taining ROSC until emergency department ad-mission. All ROSCs occurred in medical arrestpatients. Of the 71 patients who did not ex-preience cardiac arrest, 51 (71.8%) were foundto have clinical improvement by 1) return ofthe trachea to midline, 2) return of lung soundsto the affected side, 3) patient report of relief,and/or 4) improvement in systolic blood pres-sure and/or pulse rate. In the trauma group,63 (63.6%) of the patients did not have cardiacarrest; of those patients, 45 of 63 (71.4%) ex-perienced clinical improvement. In the medi-cal group, of the eight patients who did not ex-perience arrest, six (75%) of the patients expe-rienced clinical improvement. Conclusions. Inthis large urban EMS system, ND appeared toprimarily benefit trauma patients who were notin cardiac arrest. Early identification and use ofneedle decompression in patients with trauma-related pneumothorax should be encouraged.These retrospective data also suggest that nee-dle decompression may have a role in ap-propriate medical scenarios, including cardiacarrest.

73. BARRIERS, FACILITATORS,RECOMMENDATIONS, AND PRIORITIES FOR EMSRESEARCH: A SCOPING REVIEW OF RESEARCHAGENDAS

Ian Blanchard, Ryan Brown, Jan Jensen, BlairBigham, Alix Carter, Andrew Travers, LaurieMorrison, University of Calgary

Background. Several countries and organiza-tions have created a research agenda in an ef-fort to improve and focus the emergency medi-cal services (EMS) research enterprise. To date,results from these agendas have not been com-piled, which may hinder knowledge transla-tion. Scoping reviews are used to map broadtopics and summarize and disseminate re-search findings. The purpose of this study wasto map reported research agenda methods, thebarriers and facilitators to EMS research, therecommendations made, and the prioritization

of research topics and outcomes. Methods. Acombination of Medical Subject Headings andkey words was used to search Medline, EM-BASE, CINAHL, Google Scholar, and the greyliterature using Google with no date restric-tions. Search results were subject to two reviewrounds for inclusion: 1) title, Web-link, and ab-stract and 2) full article screening. Articles wereincluded if they were a “research agenda,” de-fined as a knowledge-generating project wherea group of national stakeholders in EMS re-search reached consensus on at least one ofthe following: barriers, facilitators, recommen-dations, or prioritization of research topics andoutcomes. Non–English-language articles wereexcluded. Results. Three thousand six hun-dred eighteen titles, Web-links, and abstractsand 52 full-text articles were reviewed. Tendistinct EMS research agendas reported in 17articles were included; 13 articles from peer-reviewed journals and four from non–peer-reviewed sources. Agendas from Australia (n =1), Canada (n = 1), Europe (n = 1), Ireland(n = 1), the United Kingdom (n = 1), and theUnited States (n = 5) used 13 unique method-ologies to report 22 barriers and five facilita-tors to EMS research. Agendas proffered 46 rec-ommendations for improving the research en-terprise with some setting-specific implemen-tation strategies, and 217 prioritized topics andoutcomes. Conclusions. Multiple EMS researchagendas were identified employing a variety ofmethods and revealing many barriers, recom-mendations, and priority research topics andoutcomes, but few reported facilitators. Whilesome of the results are setting-specific, therewere numerous similarities between the agen-das. Translation of the results of these agendasat the local level may be a starting point foractionable changes to the EMS research enter-prise. Future research should quantify the im-pact that these agendas have had on improvingthe quality, quantity, and usefulness of EMS re-search.

74. THE MASSACHUSETTS EMERGENCY MEDICALSERVICE STROKE QUALITY IMPROVEMENTCOLLABORATIVE

Denise Daudelin, Erin Kulick, JenniferDonovan, Mirian Barrientos, KatrinaD’Amore, Robin Ruthazer, Kathy Foell,Tufts Medical Center/Tufts Medical School

Background. Quality improvement collabora-tives (QICs) are a popular approach address-ing gaps between evidence-based practices andactual patient care. Little is known about theiruse in emergency medical services (EMS), par-ticularly to improve prehospital stroke care.Despite EMS’s important role in the strokesystem of care, no nationally recognized EMSstroke performance measures exist. We createda prehospital stroke care QIC involving Mas-sachusetts EMS agencies. Our objective wasto determine the feasibility of using a strokeQIC to improve EMS stroke care. Methods. AQIC was conducted with 16 EMS agencies. Fiveprehospital stroke performance measures weredeveloped to quantify the quality of prehos-pital care, guide QIC activities, and monitorchange in performance over time. During learn-ing sessions, participants trained in QI and per-formance measurement collected and analyzedperformance measure results and shared suc-cesses and challenges. Focus groups were con-ducted to understand participants’ experienceswith the collaborative. Mixed-model logistic re-gression was used to compare the changes inthe five measures over time. Results. Partic-ipating EMS agencies collected stroke perfor-mance measures on 2,272 patients. Adherenceto four of the five performance measures in-creased significantly overall as well as between

the first and the last three months of partici-pation. Patients with a clinical impression ofstroke had a high rate of formal stroke screen-ing performed; from the first three months tothe final three months the rate increased from94% to 96% (p = 0.32). Blood glucose testingincreased from 92% to 99% (p = 0.0002). Doc-umentation of the patient’s time last known tobe well increased from 79% to 95% (p = 0.0001).Time of stroke symptom discovery documenta-tion increased from 85% to 96% (p = 0.0004).Hospital prenotification of suspected stroke in-creased from 60% to 90% (p = 0.0001). Par-ticipants acknowledged that the QIC providedthem with an efficient and effective frameworkfor stroke QI and peer-learning opportunities.Conclusions. As evidenced through the Mas-sachusetts prehospital stroke QI collaborativeexperience, QICs can be an effective tool to im-prove EMS stroke care. The data collected, im-provements made, participation of EMS agen-cies, and positive experiences with the collab-orative support the continued use of this ap-proach.

75. IMPLEMENTATION OF PREHOSPITAL DISPATCHPROTOCOLS TO TRIAGE LOW-ACUITY PATIENTSTO ADVICE LINE NURSES

Allison Infinger, Jonathan Studnek, SteveVandeventer, Eric Hawkins, Doug Swanson,Mecklenburg EMS Agency

Background. Although emergency medical ser-vices (EMS) agencies have been designed to ef-ficiently provide medical assistance to individ-uals, the overuse of 9-1-1 as an alternate to pri-mary medical care has resulted in the need formethods to respond to this increasing demand.This study analyzes the efficacy of classifyingspecific low-acuity calls that can be transferredto an advice line nurse for further medical in-struction. The objectives of this study were toanalyze the impact of implementing this pro-tocol and resultant patient satisfaction with thetransfer to an advice line nurse. Methods. Wecollected data for retrospective review fromApril 2011 to April 2012 from a single municipalEMS agency with an average annual call vol-ume of approximately 90,000. Medical PriorityDispatch System response codes were assignedto calls based on patient acuity. Patients clas-sified under Omega response codes were as-sessed for eligibility of transfer to nurse advicelines. Exclusion criteria included the following:if the call was placed by a third-party caller; thepatient refused to be transferred to the adviceline nurse; anytime the MPDS system was notused; or the patient was referred from a skillednursing facility, school or university nursing of-fice, or physician’s office. Telephone surveyswere conducted for those patients who spoketo an advice line nurse and did not receive anambulance response 24 hours after calling 9-1-1 to determine patient satisfaction. Results.The database included 1,660 patients initiallyclassified as Omega and eligible for transferto an advice line nurse. After applying the ex-clusion criteria, 329 (19.8%) patients were ulti-mately transferred to an advice line nurse and204 (12.3%) received no ambulance response.Of those patients who were not transportedby ambulance, 118 (57.8%) patients completedtelephone follow-up with 104 (88.1%) reportingsatisfaction with the nontransport option and108 (91.5%) responding they would accept thetransfer again for a similar complaint. Conclu-sion. We identified an average of two patientsper day as eligible for transfer to the advice linenurse, with less than one patient successfullycompleting the Omega protocol per day. Whileimpact was limited, there was a decrease in am-bulance response, with maintenance of patientsatisfaction.

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


76. REGIONAL DISTRIBUTION OF PREHOSPITALPATIENTS AT RISK OF CRITICAL ILLNESS

Adam Frisch, Brian Suffoletto, CristopherSeymour, Chris Martin-Gill, University ofPittsburgh

Background. Admission to a hospital with in-creased volume of intensive care unit (ICU)admissions has been associated with reducedmortality of critically ill patients comparedwith low-volume hospitals. Calls for improve-ments in regionalized specialty care have in-cluded the preferential transfer of patients atrisk of critical illness to high-volume facili-ties. The current distribution of such patientsby emergency medical services (EMS) to high-and low-volume facilities is unknown. Meth-ods. We studied an existing cohort of prehos-pital medical records for patients transportedfrom a scene to an emergency department by37 local EMS agencies over 12 months. We usedAmerican Hospital Association data to strat-ify the 161 acute care facilities in Pennsylvaniabased on quartiles of total patient ICU days,with the highest and lowest quartiles identi-fied as high- and low-volume facilities. Dataobtained from prehospital medical records in-cluded age, vital signs, Glasgow Coma Scalescore, and receiving facility. We used a criticalillness score (Seymour et al., 2010) to stratify pa-tients based on risk of critical illness (score ≥)and identified the proportions of patients trans-ported to high- and low-volume facilities. Miss-ing data were replaced with worse-case scores.Results. A total of 50,364 medical records werereviewed. The distribution of critical illnessscores was 0 (4,885, 9.7%, 95% confidence inter-val [CI] 9.4–9.95), 1 (23,946, 47.6%, 47.1–47.9),2 (12,788, 25.39%, 25.0–25.7), 3 (5,327, 10.67%,10.4–10.9), 4 (2,265, 4.5%, 4.3–4.6), 5 (830, 1.65%,1.5–1.8), 6 (230, 0.46%, 0.39–0.51), 7 (44, 0.09%,0–0.1), and 8 (4, 0.01%). Of 21,533 patientswith critical illness score ≥2, 8,056 (37.4%,95% CI 36.8–38.1) were transported to high-volume facilities and 870 (4.0%, 3.7–4.3) to low-volume facilities (p < 0.05). 12,607 patients(58.5%, 57.9–59.2) were transported to otherfacilities. Conclusions. Many patients at highrisk for critical illness are transported to high-volume facilities, but the majority are trans-ported to other facilities and a small percent-age of these patients are transported to low-volume facilities. By identifying patients withhigh critical illness scores in the prehospital set-ting, there is potential that EMS could initiatetransport of these patients to a facility that ismore experienced in the care of the criticallyill.

77. PREVALENCE OF CERVICAL SPINE FRACTURESAMONG ELDERLY PATIENTS WHO SUFFER HIPFRACTURES DURING LOW-LEVEL FALLS: ANOPPORTUNITY TO REFINE PREHOSPITAL SPINALIMMOBILIZATION GUIDELINES

Paul Satterlee, Lori Boland, Paul Jansen,Allina Health EMS

Background. Conventional emergency medicalservices (EMS) spine-assessment approachesbased on low index of suspicion and mecha-nism of injury (MOI) result in the liberal ap-plication of prehospital spinal immobilizationin trauma patients. Since EMS protocols of-ten designate the presence of a distracting in-jury as a sufficient condition for immobiliza-tion, a painful hip fracture often obligates im-mobilization, even in an elderly patient whosuffers from a simple fall from standing andhas no apparent neurologic deficit or cervi-cal tenderness. To inform further refinement ofspine-assessment protocols, we examined theprevalence of cervical spine fractures (CSFs) in

elderly patients with hip fracture from a fallfrom standing height or less. Methods. Billingrecords of inpatient and outpatient dischargesfrom hospitals in Minnesota were used to iden-tify all cases of traumatic hip fracture that oc-curred in Minnesota in 2010. Concurrent di-agnosis codes were reviewed and the preva-lence of CSF by age and MOI was examined.Results. Among 2,747 patients with traumatichip fractures, only 1.5% (n = 40) had a CSFamong diagnosis codes for the same admission.The prevalence of CSF was only 0.3% (4/1,346)when the MOI was a same- or low-level fall,and 2.6% when the hip fracture was causedby other traumatic mechanisms (36/1,401; p< 0.001). Among the four same- or low-levelfall patients with CSF, three were over the ageof 65 years, and diagnosis codes indicative ofblunt head trauma or altered mental status(e.g., concussion, facial fractures) were foundin all four patients. Conclusions. Spinal immo-bilization is associated with discomfort, pres-sure sores, and respiratory compromise, partic-ularly in the elderly, and EMS providers con-tinue to refine spine-assessment tools in aneffort to limit unnecessary use of this proce-dure. These cross-sectional data suggest thatelderly patients with suspected hip fractureafter a fall from standing height or similarrarely suffer CSFs that would require spinalimmobilization by prehospital providers, andwhen CSF does occur in this patient popu-lation, it is frequently accompanied by headtrauma or altered mentation. Conservative useof spinal immobilization may be warrantedin elderly patients who suffer hip fractureduring low-level falls when the only crite-rion for immobilization is the distracting hipinjury.

78. DETECTING PREHOSPITAL HEMOPERITONEUMREMOTELY THROUGH FOCUSED ASSESSMENTWITH SONOGRAPHY IN TRAUMA (FAST) AND A3G NETWORK: A SIMULATION STUDY

Ki Jeong Hong, Kyoung Jun Song, SungWook Song, Sand Do Shin, NationalUniversity Boramae Medical Center, Seoul

Background. Prehospital focused assessmentwith sonography in trauma (FAST) is usedto diagnose hemoperitoneum and determinewhether transportation to a trauma center isnecessary. The goal of this study was to simu-late detecting prehospital hemoperitoneum re-motely through FAST and a 3G network. Meth-ods. We developed a real-time image trans-mission system for prehospital ultrasound. Inthe system, the ultrasound image is initiallyacquired using portable sonography (SonoSiteInc., Bothell, WA, USA) and transmitted tothe emergency department (ED) through a3G network. One emergency medical techni-cian (EMT) acquired the prehospital FAST im-age inside an ambulance. Image acquisitionand transmission was conducted at 3, 5, 10,and 15 km from the ED and during migra-tion between these points. At each point, theEMT performed FAST with a hepatorenal viewusing two phantom models randomly: nor-mal model and hemoperitoneum model. Emer-gency physicians interpreted the FAST images.We analyzed sensitivity, specificity, and areaunder the curve (AUC). We also conducted sub-group analysis by grade of emergency physi-cians (board-certified, senior resident, and ju-nior resident), moving status of the ambulancevehicle, and distance from the ED. Results. Atotal of 17 image acquisitions and transmis-sions were attempted, with a success rate of 15of 17 (88.2%). Two board-certified emergencymedicine (EM) physicians, four senior resi-dents, and two junior residents were recruitedto detected the presence of fluid collection

in the hepatorenal area. The sensitivity, speci-ficity, and AUC value of the emergency physi-cians overall were 67.9%, 78.1%, and 0.73 (95%confidence interval [CI]: 0.65–0.81), respec-tively. Results of subgroup analysis of the EMboard-certified physicians were 85.7%, 95.8%,and 0.90. Higher-grade emergency physiciansshowed a significantly higher value for theAUC (board-certified: 0.90, senior resident:0.69, junior resident: 0.63, p = 0.01). Therewas no significant difference of AUC regard-ing moving status of ambulance or distancefrom the ED (p = 0.34, 0.98). Conclusions.Simulation of detecting hemoperitoneum us-ing prehospital FAST through a 3G net-work showed acceptable performance. Highergrade of emergency physicians showed betterperformance.

79. PREHOSPITAL INTUBATION RESULTS ININCREASES IN PULMONARY INFECTION INTRAUMATICALLY INJURED PATIENTS

Dug Andrusiek, Danny Szydlo, MartinSchrieber, Karen Brasel, Russell MacDonald,Joseph Minei, Rardi van Heest, Susanne May,Emergency and Health Services Commission

Background. Infection is a major cause of mor-bidity and mortality in multisystem trauma.Sources of infection in trauma are not wellunderstood. The impact of invasive prehospi-tal procedures, including intubation and nee-dle thoracostomy, on the incidence of infec-tion is not known.We hypothesized that traumapatients who are exposed to prehospital in-tubation and needle thoracostomy will havehigher rates of pneumonia and empyema com-pared with those who have no exposure orexposure to the same procedures performedin the hospital. Methods. This was an obser-vational cohort study of data previously col-lected from the Resuscitation Outcomes Con-sortium (ROC) hypertonic saline (HS) trial. Pa-tients were included if they were found tohave an injury that resulted in shock, trau-matic brain injury, or both. Patients were ex-cluded if they had an infection detected or diedwithin the first 24 hours after injury, or if in-fection data were missing. Descriptive statis-tics were calculated, and unadjusted and ad-justed logistic regression was used to estimatethe odds ratio (OR) of having an infection ifthe patient was exposed in the prehospital set-ting compared with exposure in the hospitalor no exposure. Multivariable models were ad-justed for Abbreviated Injury Scale (AIS) score,type of injury, age, and HS treatment group.Results. Of 2,222 patients enrolled in the HSgroup, 1,676 patients met the enrollment cri-teria. Patients suffered from either pneumo-nia or empyema 4.5% of the time. Comparedwith no intubation, intubation in the prehos-pital setting was associated with a 7.7-fold in-crease (95% confidence interval [CI] 2.0, 23.0; p= 0.003) in the adjusted odds of having pneu-monia, while in-hospital intubation was associ-ated with a 4.8-fold increase (95% CI 1.4, 16.6;p = 0.01). Compared with no or in-hospitalneedle thoracostomy, prehospital needle thora-costomy was not associated with a statisticallysignificant increase in empyema or pneumonia(OR 0.34; 95% CI 0.05, 2.05; p = 0.29). Conclu-sions. In this study, exposure to intubation inthe prehospital setting was associated with anincrease in pneumonia, while prehospital expo-sure to needle thoracostomy was not associatedwith an increase in empyema. Additional re-search is needed to determine whether the in-creased risk of pulmonary infections associatedwith prehospital intubation is due to prehospi-tal airway intervention or confounded by otherfactors.

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


80. PROVISION OF PREHOSPITAL ANALGESIA TOOLDER FALLERS WITH SUSPECTED FRACTURES:ABOVE PAR, BUT OPPORTUNITIES FORIMPROVEMENT EXIST

Paul Simpson, Jason Bendall, AnneTiedemann, Stephen Lord, Jacqueline Close,Ambulance Service of New South Wales

Background. Paramedics frequently attendolder patients who have fallen and sustainedsuspected fractures, a population of patientswho may be at risk of inadequate analgesiccare. This prospective study aimed to describethe rate and effectiveness of analgesia adminis-tered by paramedics to older patients with sus-pected fractures secondary to falls, and to iden-tify predictive factors associated with the provi-sion of analgesia. Methods. A prospective, ob-servational study of patients aged 65 years andolder who had fallen was conducted from Octo-ber 1, 2010, through June 30, 2011. Fall-specificdata, collected on scene by paramedics usinga specially designed data form, were linked topatient clinical records and dispatch informa-tion. Cases in which a patient was diagnosedby paramedics as having a suspected fracturewere extracted to form the study population.Descriptive analyses were performed to de-scribe rates and effectiveness of analgesic ad-ministration, and multivariate logistic regres-sion was conducted to identify factors associ-ated with the provision of analgesia. Results.There were 333 eligible patients identified. Themean age was 82 years (standard deviation[SD] 8), and 75% were female. Suspected frac-tures of the hip were most common (42%).An initial pain score was recorded in 67% ofcases, and the median initial pain severity was8 (interquartile range [IQR] 5–9). Overall, 60%received analgesia and 80% of those receiveda parenteral opiate. Intravenous morphinewas most commonly administered (63%), fol-lowed by methoxyflurane (39%) and intranasalfentanyl (17%). Administration of an oral anal-gesic was uncommon. Analgesia was effec-tive (≥30% reduction in initial pain severity)in 62% of cases. Patients with suspected hipfractures were more likely to receive analge-sia compared with other sites (odds ratio [OR]2.7, 95% confidence interval [CI] 1.17–6.32; p= 0.02). Compared with those with mild pain,the adjusted odds of receiving analgesia in-creased for patients with moderate pain (OR6.5, 95% CI 2.3–18.8; p < 0.0001) and severepain (OR 31.1, 95% CI 9.9–97.6; p < 0.0001).Conclusions. While two-thirds of older pa-tients with suspected fractures received anal-gesia, more clinical initiatives are required inorder to optimize out-of-hospital pain man-agement. Implementation of pain measure-ment and assessment strategies specific toolder patients, and further exploration of anal-gesic alternatives such as simple analgesicsand regional anesthesia, could optimize out-of-hospital management of traumatic pain in olderpeople.

81. ANALYSIS OF PREHOSPITAL TREATMENT OFPAIN IN THE MULTISYSTEM TRAUMA PATIENT ATA COMMUNITY LEVEL II TRAUMA CENTER

Ranko Bulatovic, Paul Bradford, PeterMorassutti, Michael Lewell, Elsie Galbraith,Suzanne McKenzie, Diane Bradford, ShelleyMcLeod, Michael Peddle, Adam Dukelow,Don Eby, University of Western Ontario

Background. The Ontario advanced life sup-port patient care standards (ALS-PCS) limitthe delivery of analgesia by advanced careparamedics (ACPs) to trauma patients whohave suffered from isolated extremity trauma.However, ACPs are able to establish online

medical control to request analgesia for traumapatients who do not meet the ALS-PCS. Theprimary objective of this study was to deter-mine how often analgesia is provided to traumapatients as defined by an Injury Severity Score(ISS) >12 either by the ALS-PCS medical di-rective or through online medical control. Sec-ondary outcomes included the proportion ofpatients who were transported by ACPs versusprimary care paramedics (PCPs) and the timesaved if analgesia was delivered in the field ver-sus in the emergency department (ED). Meth-ods. A retrospective chart review of trauma pa-tients transported to a level II trauma centerfrom April 1, 2010, to March 31, 2011, was per-formed. Interfacility transports, walk-ins, andpatients who fell in hospital were excluded.Cases were reviewed by a trained ACP auditor,medical student, and emergency medical ser-vices (EMS) trauma team leader physician. Re-sults. Two hundred twenty-eight patients withISS >12 were reviewed. Seventy-eight were ex-cluded (53 interfacility transports, 23 walk-ins,and two in-hospital falls). Of the remaining 150patients, 62 (41%) had an ACP response wherethe potential to provide analgesia existed. Ofthese 62 patients, only five (8.1%) received pre-hospital analgesia. The median (interquartilerange [IQR]) time to prehospital analgesia was29 (19, 29) minutes. No patients were coveredby the ALS-PCS medical directive and all fivepatients who received analgesia required on-line medical control. Of the 57 ACP patientswho did not receive prehospital analgesia, 37(64.9%) were given analgesia in the ED (me-dian [IQR] of 68 [48, 165] minutes). Conclu-sions. Despite demonstrated rapid delivery, thefrequency of prehospital analgesia use for mul-tisystem trauma patients is extremely low. Themajority are attended by PCPs who cannot ad-minister analgesia. Promoting more frequentuse of online medical control by ACPs may al-low patients to receive analgesia much sooner.Consideration should be given to expandingprehospital directives for all paramedics toinclude pain control for multisystem traumapatients.

82. IMPACT OF A COUNTYWIDE PREHOSPITALDESTINATION PROTOCOL ON THROMBOLYTICRATES FOR ACUTE ISCHEMIC STROKE (AIS)

Prasanthi Govindarajan, David Ghilarducci,Larry Cook, Barbara Grimes, StephenShiboski, S. Claiborne Johnston,University of California San Francisco

Background. Population-based studiessupport regionalization of stroke care. Ob-jectives. We describe the intravenous (IV)tissue-plasminogen activator (t-PA) rates inambulance-transported acute ischemic stroke(AIS) patients before and during implementa-tion of a countywide prehospital stroke centerdestination protocol (SDP) and examine theassociation between SDP and IV t-PA rates inambulance-transported AIS patients. Methods.This was a cross-sectional observational studyof patients with a hospital-based diagnosis ofAIS identified using validated InternationalClassification of Diseases, Ninth Revision(ICD-9) codes. Patient records from 2005–2007were obtained from the discharge abstractfile of the statewide administrative databaseand were linked to the prehospital electronicrecords using patient-level identifiers andprobabilistic linkage methodology. Throm-bolytic use for AIS was identified using theprocedure codes in the discharge database. Weexcluded direct admissions and interfacilitytransfers. The IV t-PA rate by year wascalculated. Logistic regression was used todetermine association between SDP and the

IV t-PA rate in AIS patients after controllingfor patient and hospital demographics, strokecenter designation and teaching status of thehospital, patient residence, and day of theweek using logistic regression. Data analysiswas performed using SAS 9.2. Results. Duringa three-year period, 6,181 patients with aprimary or secondary diagnosis of strokewere transported by ambulance. The meanage at the time of admission was 74 (±15)years; 54% (n = 3,312) were female; and63% (n = 3,870) were white. The majority ofpatients, 4,132 (70%), were treated at strokecenters, and 6,005 (97%) were treated atcommunity hospitals. Among ambulancetransports, the IV t-PA rate did not increaseduring the implementation of the strokecenter designation protocol (preprotocol phase2.82%, postprotocol phase 2.85%, p-value 0.95).After controlling for patient demographics,stroke center status, teaching status of thehospital, and the weekend effect, SDP imple-mentation was not independently associatedwith increased rate of IV t-PA in AIS (oddsratio 0.96, 95 confidence interval 0.63–1.47).Conclusions. Among ambulance-transportedpatients, our preliminary findings do notshow an increase in thrombolytic rates duringimplementation of a stroke center destinationprogram.

83. ARE EMERGENCY MEDICAL SERVICES (EMS)USED FOR EMERGENCY MEDICAL TRANSPORTS?FINDINGS FROM THE NHAMCS EMERGENCYDEPARTMENT DATA FILE ON NATIONALPATTERNS OF AMBULANCE UTILIZATION(2003–2009)

Katie Tataris, John Stein, Kristin Kuzma,Renee Hsia, Judith Maselli, Ralph Gonzales,Prasanthi Govindarajan,University of California San Francisco–SanFrancisco General Hospital

Background. Emergency medical services(EMS) transports for medically unnecessaryconditions have increased over the years. Ourobjective was to identify patient and regionalcharacteristics associated with transport byEMS for these noncritical conditions. Methods.This was a cross-sectional observational studyconducted using the 2003–2009 NationalHospital Ambulatory Medical Care Survey(NHAMCS) database. The definition ofnoncritical condition was based on the Neelyconsensus conference criteria in which physi-cians coded the transport decision for eachInternational Classification of Diseases, NinthRevision (ICD-9) code as medically necessary,unnecessary, or uncertain. Majority agreementamong physicians determined the transportdecision code. Our analysis included medicallyunnecessary transports in patients ≥18 years ofage who were discharged from the emergencydepartment (ED). The characteristics of EMStransports for noncritical conditions werecompared with other modes of transport usingdescriptive statistics. Logistic regression wasused to identify independent predictors ofarrival by ambulance to the ED. Results. Theaverage proportion of noncritical transportsfor the study period was 9.4% (actual N =5,160, weighted N = 16,735,915). Factorsindependently associated with transports byEMS for noncritical conditions included age,insurance status, geographic location of the ED,urban or rural location of the ED, and patientresidence. Older patients were more likelyto use ambulances for noncritical conditions(odds ratio [OR] 1.32, 95% confidence interval[CI] 1.28–1.35). Patients with Medicaid hadhigher odds of ambulance utilization (OR 1.24,95% CI 1.07–1.43), while those with private

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


insurance had lower odds of using ambulances(OR 0.74, 95% CI 0.65–0.85). Nursing homeresidents had a higher odds compared withthose transported from other locations (OR9.43, 95% CI 7.86–11.31). Patients transportedto EDs in urban locations had a higher useof ambulances compared with those in rurallocations (OR 1.86, 95% CI 1.55–2.22). Patientsresiding in the South, Midwest, and West hadlower odds of ambulance utilization comparedwith those in the Northeast (OR 0.64, 95%CI 0.52–0.78) (OR 0.64, 95% CI 0.52–0.78)(OR 0.69, 95% CI 0.56–0.84). Conclusions.Reducing unnecessary noncritical medicaltransports by EMS could decrease the burdenon overcrowded EDs. Future efforts shouldattempt to discern patient and organizationalfactors within nursing homes associated withincreased utilization of EMS for noncriticalconditions.

84. MODIFICATION AND VALIDATION OF THEBASIC LIFE SUPPORTTERMINATION-OF-RESUSCITATION RULE INADULT OUT-OF-HOSPITAL CARDIAC ARREST INKOREA

TaeHan Kim, Sang Do Shin, Kyoung JunSong, Chang Bae Park, Eui Jung Lee, YoungSun Ro, ChungJu Medical Center, Korea

Background. The basic life supporttermination-ofresuscitation (BLS TOR)rule has been shown to reduce unnecessaryuse of emergency medical services (EMS) re-sources. The BLS TOR rule had been validatedin several countries and has proven to havehigh predictive value. East Asian EMS systemsare different from those in North Americaand Europe. All out-of-hospital cardiac arrest(OHCA) patients are transported with receiptof cardiopulmonary resuscitation (CPR) duringambulance transport in these systems. Wevalidated the BLS TOR rule and modified it tofit Korean EMS. Methods. We used the OHCAdatabase of the Seoul metropolitan area,composed of hospital and ambulance data,which contained the Utstein risks and hospitaloutcomes. We included EMS-treated victimswho were aged 18 years or older. Cases withpresumed noncardiac etiology as well as thosewithout available hospital outcome data wereexcluded. The primary outcomes were survivalto hospital discharge and good neurologicoutcome (cerebral performance category 1 to2). We tested the predictive performance of theTOR rule, calculating sensitivity, specificity,and positive and negative predictive values.We supplemented and tested the decision rulewith each response time interval (limit of 10minutes). Results. Of 3,812 OHCA patients,we excluded 1,127 (noncardiac etiology), 72(younger than 18 years), and 859 (unknowninformation on witness). A total of 2,127 cases(55.8%) were eligible for the final analysis.After applying the BLS TOR rule, 1,368 (64.3%)patients met all three criteria of the BLS TORrule. Of these, 109 (8.0%) survived to discharge(sensitivity 69.2%, specificity 64.6%, positivepredictive value 92.0%, and negative predictivevalue 26.2%) and good neurologic outcome(sensitivity 66.7%, specificity 91.8%, positivepredictive value 99.6%, and negative predictivevalue 5.1%). The addition of an additionalresponse time criterion (over 10 minutes) tothe original rule appeared to improve thespecificity (93.5%). Conclusions. The BLS TORrule showed relatively lower specificities inretrospective validation of Korean OHCA data.However, with the addition of response timeintervals, the modified TOR rules showedbetter predictive performance for differentcharacteristics of Korean EMS and prehospitalcare.

85. PATIENT-CENTERED PREHOSPITALCARE—ASSESSMENT OF PATIENT PERSPECTIVES

Kevin Munjal, Daniel Reid, Corita Grudzen,Brandon First, Kelly Silverman, KevinChason, Lynne Richardson, Mount SinaiMedical Center

Background. There is increasing interestamong the emergency medical services(EMS) community to break from traditionalemergency response paradigms that bringall patients to the emergency department(ED), in favor of a patient-centered modelfeaturing alternative destination options suchas a primary care office, clinic, or urgent carecenter, as well as “treat-and-release” protocols.However, little data is currently available onthe attitudes of patients toward assessmentby EMS professionals in the field and theappropriateness of alternative transportationdestinations. Successful redesign of the systemrequires the accurate assessment of patientperspectives. Methods. We conducted across-sectional survey among a conveniencesample of patients or their family caregiverswho presented to the ED. A survey wasdeveloped through an iterative process ofwriting, piloting, and evaluating questionsfor comprehension and content. In additionto basic demographic information and priorexperience with EMS, the participants wereasked for their level of agreement with 14 state-ments regarding current practice and proposedchanges to traditional EMS systems using aLikert scale. Data analysis and descriptivestatistics were performed using SPSS v. 18.Results. Among 54 subjects who completedthe survey, 78% agreed or strongly agreed withthe statement “I want EMS to do an evaluationand then advise me whether I need to go to thehospital.” Among the subjects, 67% indicatedthat they were comfortable being treated andreleased without seeing a physician and 70%felt comfortable being brought to an alternativedestination. Also, 91% and 94% of the sampledpatients agreed that EMS should have accessto their medical records and be able to sendcare information electronically, respectively;and 85% indicated comfort with EMS’s coordi-nating treatment and transportation decisionswith their doctor. Conclusion. Our resultssuggest that a significant proportion of patientsexpect EMS to provide an evaluation thatinforms the subsequent transportation decisionand that patient-centered prehospital caresystems would allow for a variety of treatmentand transport options. Currently, financial andregulatory barriers exist that impede redesignof the system in accordance with the patientperspectives described in this study. Perhapsovercoming these barriers would allow EMS tobetter meet public expectations.

86. DEMOGRAPHIC TRENDS IN AMBULANCE USE

Stephen Pitts, Graydon Lord, HHSEmergency Care Coordination Center

Hypothesis. We hypothesized that as the babyboom cohort ages, it will place an increasingburden on the providers of and payers for am-bulance transport. Methods. We describe thecharacteristics of ambulance use by age andpayer status in an annual sample of emergencydepartment (ED) visits, the National HospitalAmbulatory Medical Care Survey (NHAMCS).Confounding and interaction were addressedwith a regression model using complex surveymethods. We defined an urgent visit as a visitclassified by nurse triage as levels 1–3 of thefive-level triage assessment used by NHAMCS.Results. Between 2003 and 2009, there were18.3 million annual emergency medical services(EMS) transports, 16% of all ED visits to U.S.

EDs. Of all ambulance visits, 7.4 million (36%,95% confidence interval [CI] 34 to 37) were aged65 years or older. From 2003 to 2009, the av-erage age of all patients visiting EDs increasedsignificantly, by 0.18 years for each survey year.This trend disappeared when ambulance sta-tus was accounted for. Of Medicare visits, 35%(95% CI 34 to 37) were by ambulance, com-pared with 12% of other visits. Compared withMedicare-only patients, dual-eligible patientswith both Medicare and Medicaid—a proxy fordisability—were even more likely to arrive byambulance (p = 0.02 for the difference), de-spite being on average 7 years younger. The ur-gency of a visit was higher among ambulancearrivals, but Medicare patients arriving by am-bulance had a lower-than-expected level of ur-gency. Conclusions. The cause of high and in-creasing ambulance use by the elderly and dis-abled deserves further study. It may be a conse-quence of either an unmet need or reimburse-ment policy, and poses challenges for the futureof EMS.

87. PNEUMOTHORAX VOLUME EXPANSION INHELICOPTER EMS TRANSPORT

Derek Knotts, Annette Arthur, MatthewThomas, Tim Herrington, Stephen Thomas,University of Oklahoma

Background. During air medical transport, thevolume of a pneumothorax will increase asbarometric pressure decreases in accordancewith Boyle’s law. While not all medical litera-ture agrees, preflight thoracostomy is often rec-ommended even for patients with small pneu-mothoraxes. We sought to characterize altitude-related volume changes in a pneumothoraxmodel, aiming to improve clinical decisionsfor preflight thoracostomy in helicopter emer-gency medical services (HEMS) patients. Meth-ods. This prospective study used three de-vices to measure air expansion at HEMS alti-tudes. The main device was an artificial pneu-mothorax model that mimicked a human pul-monary system with a 40-mL pneumothorax.Volume measurements were made by directobservation of the meniscus at the device sy-ringe’s air–water interface. In addition, volumechanges were calculated in two spherical bal-loons (6 L and 25 L) by measuring equatorialcircumferences. All three models were flownin a Bell 206 aircraft, and measurements wererecorded at 500-foot altitude increments from1,000 to 5,000 feet above ground level. Re-sults. The three models exhibited volume in-creases of 12.7%–16.2% at 5,000 feet comparedwith ground level. Univariate linear regres-sion yielded similar increases, 1.27%–1.52%, involume per 500-foot altitude increase for allthree models. Bivariate-indexed linear regres-sion identified no association between volumeincrease and assessment model (p-values 0.19and 0.29). Locally weighted scatterplot smooth-ing (LOWESS) plots indicated linearity of thealtitude–volume relationship. Conclusion. Thisstudy modeled predictable pneumothorax vol-ume changes at typical HEMS altitudes. In-creased understanding of altitude-related vol-ume changes will aid in peritransport decisionmaking.

88. PARAMEDIC ATTITUDES REGARDINGEXCEPTION FROM INFORMED CONSENTRESEARCH

Ralph Bledsoe, Glenn Bloom,York Hospital Emergency Medicine Residency

Background. Emergency medical services(EMS) research suffers from the paradox ofinformed consent and life-threatening situa-tions. Research under exception from informed

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


consent (EFIC) was designed to foster researchin situations where existing treatments wereunproven or unsatisfactory and traditionalinformed consent was impractical given thepatient’s immediate need. The patient or his orher legally authorized representative is underduress in these situations. We investigatedthe understanding and comfort RAMPART-participating paramedics had with EFIC in apilot survey. The authors know of no publishedresearch into paramedics’ attitudes and beliefsabout EFIC research. Methods. FollowingRAMPART’s closure, EMS site coordinatorswere e-mailed a link to a SurveyMonkey.comsurvey of attitudes and feelings about EFICresearch and asked to forward it to theirenrolling paramedics. Wellspan Health’s insti-tutional review board exempted the study fromreview. The survey instrument was validatedby cognitive testing of local RAMPART-trainedparamedics. Paramedics in the authors’ systemdid not participate. After completion, theparticipants could enter a drawing for $50 giftcards. Results. Fifty-five paramedics averaging17 years’ experience completed the survey; 50responded to all questions. Of the respondents,49% felt research was more important than theright to consent, 16% felt the opposite, and35% were neutral. Enrollers were significantlymore likely to believe benefits outweighedautonomy; nonenrollers believed in equalitybetween the two (p = 0.006). Eighty-eightpercent believed research on ambulances wasimportant to improve future patient care, andthree were negative; 64% believed a paramedicshould “fully participate in any research theirorganization is participating in,” while 34% feltobjectors should be allowed not to enroll. Inaddition, 72% felt EFIC was ethical and 74%felt it to be legal. Further, 46% had enrolledpatients, 48% had not but would enroll, and6% would not have enrolled a patient. Analysisof variance was used to seek relationshipsbetween data points. Conclusions. Assuringfuture buy-in requires further research intoparamedics’ attitudes, and EMS workforce ed-ucation is needed as prehospital EFIC researchincreases. The enrollers’ greater belief in theweight of societal benefit merits further study.The respondents’ experience demonstrates aneed to train established providers in additionto newly certifying providers. Our smallsample size and inability to track the responserate are limitations of this pilot.

89. COMPARING WORK-RELATED STRESS ANDSTRESS REACTIONS IN AMERICAN ANDCANADIAN EMS PERSONNEL

Elizabeth Donnelly, Paul Bradford, SeveroRodriguez, University of Windsor

Background. There are many similarities inhow emergency medical services (EMS) areprovided in the United States and Canada;however, significant structural and systemicdifferences in service provision exist. Whileextant literature has linked workplace stressto stress reactions in EMS personnel, virtu-ally no attention has been given to explor-ing how those stresses and stress reactionsmay be influenced by the differences in theAmerican and Canadian EMS systems. Thegoal of this study was to assess whether thereare differences in work-related stresses andstress reactions in American and Canadian re-sponders. Methods. Two online surveys wereconducted utilizing the same instruments. In2009, a probability sample of 12,000 nationallyregistered emergency medical technicians andparamedics were surveyed, with a 13.6% re-sponse rate (n = 1,633). In 2011, paramedics ina municipality-based service in southwest On-tario were surveyed, with a 54% response rate(n = 145). Respondents reported levels of oper-

ational and organizational chronic stress, criti-cal incident stress, posttraumatic stress symp-tomatology (PTSS), alcohol use, and demo-graphic characteristics. t-Tests and chi-squareanalyses were used to assess for significantdifferences. Results. American responders re-ported higher mean levels of operational stress(39.0 vs. 31.4; delta 7.6; 95% confidence in-terval [CI]: 5.5–9.7). Canadian responders re-ported higher levels of alcohol use (5.9 vs.4.3; delta 1.6; 95% CI: 0.9–2.4). No significantdifferences were identified in organizationalstress, critical incident stress, or PTSS. Signif-icant demographic differences were also iden-tified; American responders were significantlyyounger (35.1 vs. 38.3 years; delta 3.2; 95%CI: 1.3–5.1), had fewer years of experience inEMS (9.2 vs. 13.8; delta 4.6; 95% CI: 2.7–6.5),worked more hours weekly (4.1 vs. 3.9; delta0.3; 95% CI 0.01–0.4), and reported lower in-come (3.6 vs. 6.7; delta 3.1; 95% CI 2.8–3.4).Conclusion. Significant variations were iden-tified in self-reported operational stress, in al-cohol use, and in demographic factors. Thehigher levels of operational stress in Americansmay be the result of the increased number ofhours at work, fewer years of experience, andlower wages. More investigation is needed toexplicate how systemic differences in EMS sys-tems may influence the health and well-beingof EMS personnel.

90. ALCOHOL USE AND MISUSE IN EMERGENCYMEDICAL SERVICES

Elizabeth Donnelly, University of Windsor

Background. Alcohol misuse has been identi-fied as problematic for police officers and fire-fighters. Further, research has demonstratedthat in police subculture, alcohol use is mod-eled as a way to deal with workplace-relatedstress. Despite the recognition that alcohol mis-use occurs among other first responders, nosystematic inquiry has been made into alcoholuse among emergency medical services (EMS)personnel. The purpose of this study was to as-sess the prevalence of alcohol use among EMSpersonnel and ascertain what work-related fac-tors were influential in predicting increased useof alcohol. Methods. A probabilistic sample (n= 12,000) of nationally registered emergencymedical services and paramedics were askedto respond to an online survey, reporting lev-els of operational stress, organizational stress,critical incident stress, posttraumatic stress, al-cohol use (including whether they used alco-hol to cope with a bad call or shift or they sawtheir colleagues using alcohol to cope), and de-mographic characteristics. Pearson correlationcoefficients were used to estimate linear de-pendence between stress variables. Ordinaryleast-squares regression determined predictorvariables independently associated with alco-hol use. Results. A total of 1,633 responses werereceived (13.6% response rate). Among the re-spondents, 15.1% reported risky drinking; 1.3%reported possible alcohol dependence; 18.8%reported binge drinking at least monthly; 49.2%reported using alcohol to cope with a bad call orshift; and 83.7% reported seeing coworkers usealcohol to cope with a bad call or shift. Alcoholuse was significantly correlated with critical in-cident stress (r = 0.08, p < 0.01) and with post-traumatic stress (r = 0.187, p < 0.01). Multivari-ate analysis revealed multiple significant pre-dictors for alcohol use, including age (p < 0.01),length of service (p < 0.05), gender (p < 0.001),marital status (p < 0.001), operational stress (p< 0.001), posttraumatic stress (p < 0.001), usingalcohol to cope (p < 0.001), and seeing othersuse alcohol to cope (p < 0.001). The final modelhad an adjusted R2 of 0.424. Conclusion. Thesefindings indicate that alcohol use is prevalentamong EMS personnel. To address potential

alcohol misuse, greater investment should bemade to educate personnel about the risks as-sociated with alcohol use and healthy copingtechniques. Further, these findings reinforce theimportance of making resources available tosupport personnel that may be misusing alco-hol as a result of workplace stress.

91. INCREASING PREHOSPITAL RECOGNITION OFSEPSIS: A FEASIBILITY STUDY TO EVALUATE THEUSE OF LACTATE METERS AND TEMPORALARTERY THERMOMETERS BY PARAMEDICS

Lori Boland, Jonathan Hokanson, TonyOlson, Karl Fernstrom, Charles Lick,Allina Health

Background. To assess the feasibility of equip-ping prehospital providers with temporalartery thermometers (TATs) and handheld lac-tate meters to increase identification of patientsat risk of sepsis. The correlation between pre-hospital and emergency department (ED) lac-tate values and the time interval that sepa-rates the availability of the two measures toclinicians were also evaluated. Methods. Thispilot study used a convenience sample of pre-hospital patients meeting risk criteria for sep-sis. Paramedics received education on systemicinflammatory response syndrome (SIRS) crite-ria and instruction on the use of TATs andhandheld lactate meters. Patients were enrolledif they had a recent history of infection, met≥2 SIRS criteria, and were being transportedto a participating hospital. A lactate test wasperformed by paramedics in the prehospitalsetting and again upon arrival in the ED viausual care (i.e., venipuncture and laboratoryprocessing). Paramedics entered study data us-ing an online database accessible at the pointof care. Results. Among 64 patients enrolledover 12 months, 32 had a prehospital bodytemperature of <36◦C or >38◦C by TAT. Therange of prehospital lactate values was 0.8 to9.8 mmol/L and the unadjusted Pearson cor-relation between prehospital and ED-measuredlactate values was 0.70 (p < 0.001). Among 10patients whose prehospital lactate value wasabove the recommended threshold for earlygoal-directed therapy (EGDT; i.e., 4 mmol/L),eight had ED lactate values <4 mmol/L, al-though five had received fluids in the prehos-pital setting. The median time intervals be-tween prehospital lactate measurement and 1)ED venipuncture and 2) ED lactate result avail-able in the electronic hospital record were 76minutes and 116 minutes, respectively. Of 56study patients admitted to the hospital, 14 ulti-mately received a diagnosis of sepsis. Conclu-sion. Paramedics can identify patients meetingSIRS criteria and use TATs and handheld lactatedevices to evaluate patients more comprehen-sively for sepsis risk. The clinical significance ofprehospital lactate values >4 mmol/L requiresfurther study, but if proven a reliable indicator,paramedics may be able to relay a key clinicalalert value for EGDT approximately two hoursearlier than is currently possible.

92. A RELIABLE AND VALID TOOL FORDETECTING ADVERSE EVENTS IN HELICOPTEREMS

P. Patterson, Judy Lave, Matthew Weaver,Chris Martin-Gill, Francis Guyette, RonaldRoth, Jon Rittenberger, Richard Wadas,Vincent Mosesso, Robert Arnold, DonaldYealy, Department of Emergency Medicine,University of Pittsburgh

Background. We sought to develop a reli-able and valid tool to identify adverse events(AEs) in helicopter emergency medical services(HEMS). Methods. We adhered to a recom-mended multistep process for developing and

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


testing new measurements. First, we soughtto develop a content-valid tool by conven-ing an expert panel of three senior emergencymedicine (EM)-trained physicians, three mid-career EM physicians, and four quality officerscertified as flight paramedics and flight nurses.These experts met in eight face-to-face sessionsto edit a draft list of 20 AE triggers and es-tablish methods for rating proximal cause andseverity of AEs. An example of a trigger is “car-diac arrest during transport.” The experts ratedthe relevance of each AE trigger, each com-ponent of proximal cause, and each compo-nent of AE severity using an established four-point scale developed by Lynn et al., 1985. Thescale included Not Relevant, Somewhat Rel-evant, Quite Relevant, and Highly Relevant.We then calculated the Item-level Content Va-lidity Index (I-CVI) and removed items withless than 0.78 I-CVI in accordance with es-tablished benchmarks. Finally, we calculatedthe average Scale-level Content Validity Index(S-CVI) developed by Rubio et al. 2003 andWaltz, et al. 2005. Results. Experts reachedconsensus on 14 AE triggers, five categoriesfor assigning proximal cause, and three cat-egories for rating AE severity. We removedthree of the 14 triggers with an I-CVI lessthan 0.78 (e.g., “time from dispatch to initialpatient contact exceeds accepted standards”).The remaining triggers had a content valid-ity mean I-CVI score of 0.94. The mean I-CVIfor the five categories of proximal cause was0.92. The mean I-CVI for the three levels ofrating AE severity was 0.93. Overall, the con-tent validity (S-CVI) for the three componentsof our method (triggers, proximal cause, andseverity) was 0.93, with I-CVI ranging from0.80 to 1.0. Conclusions. Initial tests confirma content valid method for identifying AEs inHEMS.

93. THE IMPACT OF HYPOTHERMIA TREATMENTON SURVIVAL TO HOSPITAL DISCHARGE FOROUT-OF-HOSPITAL CARDIAC ARREST PATIENTS INTHE CIRCULATION IMPROVING RESUSCITATIONCARE (CIRC) TRIAL

Lars Wik, Jan Olsen, David Persse, FritzSterz, Michael Lozano, Marc Brouwer, MarkWestfall, Chris Souders, Reinhard Malzer,Pierre van Grunsven, David Travis, UlrichHerken, Brooke Lerner,National Competence Center for EmergencyMedicine

Background. Therapeutic hypothermia (TH)has been associated with increased survivalafter out-of-hospital cardiac arrest (OHCA).There is debate about when TH should be ap-plied after OHCA. In the Circulation Improv-ing Resuscitation Care (CIRC) trial, applicationof TH was captured for three distinct treatmentperiods: prehospital (PH), in the emergency de-partment (ED), and in hospital (IH). A post hocanalysis of the CIRC database evaluated the ef-fect of TH during these periods on survival todischarge. Methods. All study patients (botharms) admitted to hospital were included. Be-cause patients could have TH initiated in thePH, ED, or IH phase of their care and not allpatients had hypothermia maintained betweenone phase of care and the next, a TH score wascreated. The score awarded points for each lo-cation where TH was provided: 3 for PH, 2 forED, and 1 for IH. A maximum score of 6 wasreceived if they had TH in all three settings; theminimum score was 0 if no TH was provided.Logistic regression was used to determine theinteraction between the TH score and survivalto hospital discharge, adjusting for the same co-variates used in the CIRC study’s survival anal-ysis: shockable initial rhythm, witnessed arrest,age group, and study site. Results. Of the 4,231subjects enrolled, 1,068 were admitted to hospi-

tal. Survival information unknown for two sub-jects. Of the remaining 1,066, 36% had a score of0 (no TH), 8% score of 1 (IH-TH only), 3% scoreof 2 (ED-TH only), 26% score of 3, 6% scoreof 4, 3% score of 5, and 17% score of 6 (PH-,ED-, and IH-TH). The adjusted odds ratio (OR)for survival to discharge was 1.105 (95% con-fidence interval 1.030–1.186, p < 0.01) for eachone-point increase in the TH score. For exam-ple, a subject who received PH-, ED-, and IH-TH (score of 6) had an OR of 1.8 for survivalto hospital discharge. Conclusion. Our analysisindicates that TH treatment in OHCA patientsshows the most benefit when started in the fieldand continued into the hospital without inter-ruptions in the ED.

94. EFFECTS OF COMPRESSION RATE ONCORONARY PERFUSION PRESSURE AND CAROTIDBLOOD FLOW IN A PORCINE MODEL OFSHOCK-INDUCED PULSELESS ELECTRICALACTIVITY

Xiaobo Wu, Weilun Quan, Yinlun Weng, WeiChen, Shijie Sun, Wanchun Tang,Weil Institute of Critical Care Medicine

Background. The 2010 American Heart Associ-ation guidelines recommend a chest compres-sion (CC) depth of at least 2 inches. However,most clinical studies found an inverse associa-tion between CC depth and rate. The effect ofCC rate on coronary perfusion pressure (CPP)and carotid blood flow (CBF) has not been de-termined. This study was conducted to inves-tigate hemodynamic responses to CC at dif-ferent compression rates in a porcine modelof postshock pulseless electrical activity (PEA).Methods. Ventricular fibrillation was electri-cally induced and untreated for 2–7 minutesin 12 domestic pigs weighing 22–24 kg. Post-shock PEA was induced with electric shock.Peak aortic pressure below 40 mmHg was usedto identify a qualified PEA. Once PEA wasinduced, the animals received 30 seconds ofPEA-triggered synchronized sternal CCs usinga modified Thumper device. The CC depth wasmaintained at 2 inches. If the animals were re-suscitated after the study sequence, PEA in-duction and the study sequence were repeatedafter 30 minutes of recovery. Results. A to-tal of 1,102 compression cycles from 29 quali-fied PEA events were included in the analyses.The rate of induced PEA varied from 30 to 125bpm. When the CC rate increased from 80 to120 bpm with fixed CC depth, CPP increasedby 5.6 mmHg (32.9%) and CBF increased by46.4 mL/min (50.2%). Both CPP and CBF werepositively correlated to CC rate, with correla-tion coefficients of 0.32 and 0.49, respectively.Linear regression found a positive trend forboth CPP and CBF, with slopes over the CCrate of 0.14 and 1.16, respectively. Conclusions.In this shock-induced PEA porcine model, bothCPP and CBF increased with CC rate. These re-sults indicate that faster sternal CC would gen-erate better coronary and cerebral perfusion inthe range from 80 to 120 bpm.

95. REEVALUATING EPINEPHRINE INOUT-OF-HOSPITAL CARDIAC ARREST: ASYSTEMATIC REVIEW AND META-ANALYSIS

Steve Lin, Clifton Callaway, PrakeshkumarShah, Justin Wagner, Joseph Beyene, LaurieMorrison,Rescu, Li Ka Shing Knowledge Institute, St.Michael’s Hospital

Background. International resuscitation guide-lines support the routine administration ofepinephrine during out-of-hospital cardiac ar-rest (OHCA). However, the evidence is notclear for the efficacy of epinephrine in long-and short-term survival. Objective. To evalu-

ate the efficacy of epinephrine (all doses andcombinations) in adult OHCA patients on long-term outcomes (survival to discharge) andshort-term outcomes (survival to admissionor return of spontaneous circulation [ROSC]).Methods. The search included Medline, EM-BASE, and the Cochrane Library up to March1, 2012, and hand-searches of bibliographiesand electronic resources to identify eligiblepublished and unpublished randomized con-trolled trials (RCTs). Eligible trials comparedepinephrine with vasopressin or placebo inadult OHCA patients. Two independent re-viewers conducted the hierarchical selection,abstracted data, and assessed quality. Disagree-ment was resolved by consensus. The Mantel-Haenszel random-effects method was usedto test for differences. A subgroup analysiswas performed using stratification by location(prehospital or emergency department). Re-sults. Fourteen RCTs (N = 12,246) met the in-clusion criteria: six compared standard-doseepinephrine (SDE) with ≥5-times higher-doseepinephrine (HDE) (n = 6,174), six comparedSDE with epinephrine–vasopressin combina-tion (n = 5,202), one compared SDE withvasopressin alone (n = 336), and one com-pared SDE with placebo (n = 534). Therewas no survival-to-discharge advantage. HDEshowed improved survival to admission (rel-ative risk [RR] 1.15 [95% confidence inter-val (CI): 1.00–1.32], p = 0.05) and ROSC (RR1.17 [95% CI: 1.03–1.34], p = 0.02) over SDE.SDE showed improved survival to admis-sion (RR 1.95 [95% CI: 1.34–2.84], p < 0.001)and ROSC (RR 2.80 [95% CI: 1.78–4.41], p <0.001) over placebo in the single RCT. HDEand the epinephrine–vasopressin combinationimproved survival to admission when strati-fied by prehospital setting (RR 1.15 [95% CI:1.00–1.32], p = 0.05] and emergency depart-ment (RR 1.20 [95% CI: 1.04–1.38], p = 0.01),respectively. Conclusions. There was no clearadvantage of any vasopressor in long-term sur-vival. Similar short-term benefit was seen withHDE over SDE and SDE over placebo, suggest-ing that HDE should be a treatment option incurrent resuscitation guidelines.

96. TEAM CPR: EVALUATION OF A METHOD TOENABLE YOUNGER CHILDREN TO PERFORMMORE EFFECTIVE CHEST COMPRESSIONS

Chad Panke, Angelo Salvucci, Lynn White,David Chase, Barbara Spraktes-Wilkins,American Medical Response

Background. Prompt continuous high-qualitycardiopulmonary resuscitation (CPR) increasessurvival in cardiac arrest. Although childrenaged 8 years can learn the theory and me-chanics of CPR, they typically do not havesufficient size or strength to perform effectivechest compressions. The purpose of this studywas to examine the effectiveness of a simulta-neous two-rescuer chest compression methodperformed by young children. Methods. Thiswas a pilot study to describe the quality ofchest compressions done by children as indi-viduals and as a two-person team. Volunteerchildren, without previous CPR training, weretrained and asked to perform 150 chest com-pressions at a 30:2 compression-to-ventilationratio on the SmartMan manikin system (Ambu,Inc.). A metronome was used to prompt therate. Within two-person teams, subjects firstperformed alone. They were then taught howto position themselves on opposite sides ofthe manikin chest and place their hands nextto each other on the lower sternum to per-form simultaneous chest compressions, andthen performed 150 compressions at 30:2 as ateam. The examiner and subjects were blindedto the results. Two-tailed t-tests were used to

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


compare chest compression quality parame-ters in individual and team tests. Results. Six-teen subjects completed both tests. Eight weremale; their ages were 6–9 years, heights were44–53 inches, and weights were 43–80 lb. Ofthe 2,268 compressions performed by individ-ual subjects, the median depth was 0.89 inches.Of the 1,245 compressions performed by two-person teams, the median depth was 1.17 in.The median increase in depth by team over in-dividual chest compressions was 0.29 in (p =0.015). No individual or team median achievedthe recommended minimum depth of 2 in. Themetronome-guided rate was equivalent amongall groups. Conclusions. In this study the qual-ity of chest compressions performed by youngchildren was significantly better when done us-ing the team method. Although teams did notreach the American Heart Association Guide-lines 2010–recommended depth, team CPR wassuperior to individual efforts. The team CPRmethod may allow younger children, or anygroup of rescuers unable to reach adequatechest compression depth as individuals, to im-prove the effectiveness of CPR.

97. A BLINDED EVALUATION OF COMBINATIONDRUG THERAPY FOR PROLONGED VENTRICULARFIBRILLATION CARDIAC ARREST

Timothy Mader, Ryan Coute, Scot Millay,Adam Kellogg,Baystate Medical Center/Tufts UniversitySchool of Medicine

Background. There are many examples inwhich medical disease treatments have beeninnovated from single-agent, failure-based, se-rial medication delivery to first-line combina-tion drug therapy. Could the same approachimprove outcomes in the metabolic phase ofventricular fibrillation (VF)? Objective. Usinga swine model of prolonged VF, we comparedreturn of spontaneous circulation (ROSC) and20-minute survival in animals treated with sin-gle agents given in series to animals providedfirst-line combination drug therapy. Methods.The study was approved by the InstitutionalAnimal Care and Use Committee. Eighty swine(25–30 kg) were surgically instrumented un-der anesthesia, and VF was electrically in-duced. After 12 minutes of untreated VF, theanimals were 2:1 block-randomized to one oftwo resuscitation schemes. The SERIES group (n= 53) received the following series: standard-dose epinephrine (0.01 mg/kg [SDE]); vaso-pressin (0.57 mg/kg); amiodarone (4.3 mg/kg)and SDE; sodium bicarbonate (1 mEq/kg) andSDE—each delivery punctuated by a rescueshock (RS) after 3 minutes of cardiopulmonaryresuscitation (CPR). If ROSC occurred after anyRS, supportive care was provided and subse-quent medications were not given. The COCK-TAIL group (n = 27) received all drugs together(with metoprolol [0.2 mg/kg]), followed byCPR and the first RS. An SDE followed eachfailed RS. All drugs were given intravenouslywith a flush. CPR and RS attempts were stan-dardized. Resuscitation continued until ROSCwas achieved or 20 minutes elapsed with-out ROSC. Group comparisons were assessedusing descriptive statistics. Proportions with95% confidence interval (CI) were calculatedfor VF termination, ROSC, and survival. Re-sults. At baseline, the two groups were simi-lar. Of the 27 animals in the COCKTAIL group,VF was terminated in 24 (proportion: 0.89 [95%CI 0.72–0.96]), six animals experienced ROSC(proportion: 0.22 [95% CI 0.11–0.41]), and fiveachieved 20-minute survival (proportion: 0.19[95% CI 0.08–0.37]). Of the 53 animals in theSERIES group, VF was terminated in 47 (pro-portion: 0.89 [95% CI 0.77–0.95]), 30 animalsexperienced ROSC (proportion: 0.57 [95% CI0.43–0.69]), and 28 animals achieved 20-minute

survival (proportion: 0.53 [95% CI 0.40–0.66]).The SERIES outcomes mirrored those of histor-ical controls (VF termination proportion: 0.82,ROSC proportion: 0.59, and 20-minute survivalproportion: 0.40). Conclusion. In this swinemodel of prolonged out-of-hospital cardiac ar-rest, short-term outcomes were adversely af-fected by this drug–dosage combination.

98. VIDEO-ASSISTED FEEDBACK DURING CPR:ANALYSIS OF SMARTPHONE VIDEO FOOTAGEACCURATELY CLASSIFIES CHEST COMPRESSIONRATE

Adam Frisch, Samarjit Das, Joshua Reynolds,Jestin Carlson, Fernando De la Torre, JessicaHodgins, UPMC

Background. Real-time cardiopulmonary re-suscitation (CPR) feedback improves chestcompression (CC) rate in the prehospital set-ting; however, few such tools exist for layrescuers. Smartphone applications utilizing anembedded gyroscope can provide CPR feed-back, but this requires purchase of additionalaccessories that may not be readily avail-able when needed. We evaluated whethervideo footage from a smartphone camera couldbe used to determine CC rate during simu-lated bystander CPR. Hypothesis. Analysis ofsmartphone video footage can discriminate be-tween CCs performed too slowly (<100 com-pressions/min), too quickly (>120 compres-sions/min), or within recommended ranges(100–120 compressions/min). Methods. Sixsubjects previously trained in CPR performedCCs on a CPR manikin. Each subject performedfive 30-second bouts at specified parameters:normal rate/normal depth, normal rate/toodeep, normal rate/too shallow, too fast/normaldepth, and too slow/normal depth. Partici-pants were recorded using a smartphone cam-era placed flat on the floor between them andthe manikin. Inertial measurement devices at-tached to the participants’ hands determinedactual compression rate. We divided each videorecording into 2-second epochs, calculating theoverall classification accuracy of video seg-ments via two different methods using a com-puter vision algorithm for determining repeti-tive movement patterns from video. Analysis I:Half the video segments from each subject wereused as a training set, and the other half wasa test set. Analysis II: All the video segmentsfrom half of the subjects were used as a train-ing set, and the other subjects were a test set.We determined overall classification accuracyby k-nearest neighbors. Results. Smartphonevideo recording yielded high-quality video foranalysis, generating a total of 153 video seg-ments. Recorded CC rates ranged from 60 to144 per minute. Analysis I yielded an overallclassification accuracy of 88% (95% confidenceinterval 82.2–92.4); analysis II yielded an over-all classification accuracy of 80% (72.6–85.3).Conclusion. Analysis of video obtained froma smartphone accurately classifies CC rate intothree categories: too fast, too slow, or withina desired range. Smartphone applications inte-grating this technology could provide real-timeCPR feedback regarding CC rate to lay rescuerswithout the need for additional accessories.

99. ASSESSING THE ACCURACY OF ED VS EMSCOMPUTER ECG INTERPRETATION FORIDENTIFYING ACUTE MYOCARDIAL INFARCTION

Kelly Sawyer, Stephanie Hang, Amy Kule,Alfred Burris, Justin Trivax, Robert Swor,William Beaumont Hospital

Background. Emergency medical services(EMS) systems and emergency departments(EDs) have used computer interpretationof electrocardiograms (CI-ECG) to shorten

triage time and decrease door-to-balloontime for patients with acute myocardialinfarction (AMI). However, device-specificalgorithms may potentiate inconsistencies ininterpretation. Our objectives were to evaluatethe performance characteristics of EMS andED CI-ECG and to assess their agreement forthe diagnosis of coronary artery occlusionusing angiography as the “gold standard.”Methods. We examined a retrospective cohortof adult, EMS-transported patients whoseECGs were obtained both by EMS (EMS-ECG)and ED (ED-ECG) ECG devices and were takenemergently to the cardiac catheterization lab-oratory for suspicion of AMI. This study wasconducted at a single academic communityED from January 2006 to February 2011. Caseswere dichotomized as either “CI-ECG+,”defined as “AMI suspected” by computerprintout, or as “CI-ECG–.” The primaryoutcome (“definite AMI”) was determined byevidence for acute coronary vessel occlusionor presence of thrombus in a culprit vessel onangiography. To assess the relative accuracyof EMS-ECG and ED-ECG, we calculated thesensitivity, specificity, and likelihood ratios(LRs) of each device type to predict coronaryartery occlusion; the kappa value assesseddevice agreement for AMI diagnosis. Results.A total of 173 patients were identified. Themean time from EMS-ECG to reperfusion was81.9 (standard deviation [SD] ±25.6) minutesand that from ED-ECG to reperfusion was56.0 (SD ±20.4) minutes. Overall, 116 (67.0%)cases had coronary occlusion on angiography.Performance characteristics for EMS-ECG andED-ECG for definite AMI revealed sensitivityof 62.9% (53.4, 71.6) and 50.0% (40.5, 59.4);specificity of 36.8% (24.8, 50.7) and 64.9%(51.0, 76.8); LR+ of 0.99 (0.78, 1.27) and 1.42(0.96, 2.12); and LR– of 1.01 (0.75, 1.35) and0.77 (0.62, 0.94), respectively. There was pooragreement on AMI diagnosis between devices(κ 0.23 [95% confidence interval 0.09, 0.37]).Conclusion. Using coronary angiographyas a gold standard, neither device’s CI-ECGwas accurate for identifying coronary arteryocclusion, nor was there significant agreementbetween EMS and ED CI-ECG diagnoses.Further work is needed to understand therelative value of CI-ECG in the diagnosis ofcoronary artery occlusion in the ED.

100. ASSESSMENT OF ADVERSE EVENTS IN APRIMARY CARE PARAMEDIC ST-SEGMENTELEVATION MYOCARDIAL INFARCTION BYPASSPROGRAM IN A LARGE RURAL AREA

Mark Froats, Andrew Reed, Richard Dionne,Justin Maloney, Rob Burns,Queen’s University

Background. This study was conducted to de-termine the frequency and nature of adverseevents in a basic life support (BLS) ST-segmentelevation myocardial infarction (STEMI) by-pass program in a large rural area. In our re-gion, BLS providers bypass closer emergencydepartments (EDs) to deliver patients to per-cutaneous coronary intervention (PCI) centersup to 60 minutes away, exceeding the 30-minute limit recommended by the AmericanHeart Association. The safety of this transporttime and non–advanced life support (non-ALS)providers attending STEMI patients during by-pass has been questioned. Methods. We con-ducted a health records review for patientstransported by a rural BLS emergency med-ical services (EMS) agency under bypass toour regional PCI center. Patients were eligi-ble if they had less than 12 hours of chestpain, a STEMI-positive electrocardiogram, anda drive time to the PCI center within 60 min-utes. We determined transport times and ad-verse events during transport, which were

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


bradycardia (<50 bpm), tachycardia (>140bpm), hypotension (systolic blood pressure<90 mmHg), cardiac arrest, and death. We con-ducted descriptive data analyses with 95% con-fidence intervals (CIs). Results. Forty-five con-secutive cases were identified between Febru-ary 2005 and February 2012. The mean agewas 61.2 years, with 32 (71.1%) male patients.The mean transport time was 30.0 minutes, and13 cases (28.9%) exceeded this (range 31–62minutes). Twenty cases (44.4%) had an ad-verse event. The event rate for cases withdrive time under 30 minutes was 40.0% (95%CI 22.5–57.5%), compared with 46.2% (95% CI19.1–73.3%) for cases with longer drive times.Of three events occurring after 30 minutes oftransport, there was one ventricular fibrilla-tion arrest that responded to one shock, andtwo transient episodes of hypotension. Fifteen(75.0%) of the adverse events were transientlyabnormal vital signs requiring no intervention.Four adverse events (20.0%) would have bene-fited from an ALS intervention. None of thesewould have been transported if the vital signadverse event parameters and abnormal Glas-gow Coma Scale score were contraindicationsto bypass. There were no deaths. Conclusion.BLS EMS providers in a rural county maybe able to safely bypass the closest ED andtransport STEMI patients to a PCI facility pro-vided there are no vital sign contraindicationsat initial presentation. There was no differ-ence in adverse event rate with extended drivetimes.

101. EFFECT OF CONTINUOUS POSITIVE AIRWAYPRESSURE (CPAP) ON MORTALITY IN THETREATMENT OF ACUTE CARDIOGENICPULMONARY EDEMA IN THE PREHOSPITALSETTING: RANDOMIZED CONTROLLED TRIAL

Michael Austin, Karen Wills, DavidKilpatrick, Ottawa Hospital

Background. The prehospital use of continuouspositive airway pressure (CPAP) ventilation isa relatively new management for acute cardio-genic pulmonary edema (ACPE), and there islittle high-quality evidence of the benefits orpotential dangers in this setting. The aim ofthis study was to determine whether patientsin severe respiratory distress treated with CPAPin the prehospital setting have a lower mortal-ity than those treated with usual care. Meth-ods. Randomized controlled trial comparingusual care versus CPAP (WhisperFlow R©) in aprehospital setting, for adults experiencing se-vere respiratory distress, with falling respira-tory efforts, due to a presumed ACPE. Patientswere randomized to receive either usual care,including conventional medications (nitrates,furosemide, and oxygen) plus bag–valve–maskventilation, versus conventional medicationsplus CPAP. The primary outcome was prehos-pital or in-hospital mortality. Secondary out-comes were the need for tracheal intubation,length of hospital stay, change in vital signs,and arterial blood gas results. We calculatedrelative risk (RR) with 95% confidence inter-vals (CIs). Results. Fifty patients were enrolled,with a mean age of 79.8 years (± standard de-viation [SD] 11.9), male 56.0%, and mortality20.0%. The risk of death was significantly re-duced in the CPAP arm, with mortality 34.6%(9 deaths) in the usual-care arm compared with4.2% (1 death) in the CPAP arm (RR, 0.12; 95%CI 0.02 to 0.88; p = 0.04). Patients who receivedCPAP were significantly less likely to have res-piratory acidosis (mean difference in pH 0.09;95% CI 0.01 to 0.16; p = 0.02; n = 24) than pa-tients receiving usual care. The length of hospi-tal stay was significantly shorter for the patientswho received CPAP (mean difference 2.3 days;

95% CI –0.01 to 4.6, p = 0. 05). Conclusion. Wefound that CPAP significantly reduced mortal-ity, respiratory acidosis, and length of hospitalstay for patients in severe respiratory distresscaused by ACPE. This study shows that the useof CPAP in the prehospital setting for ACPE im-proves patient outcomes. Trial reg.: ANZCTRACTRN12609000410257. Funding: Fisher andPaykal, suppliers of the WhisperFlow CPAPdevice.

102. INITIAL IMPACT OF PREHOSPITALELECTROCARDIOGRAMS ACQUIRED BY BASIC LIFESUPPORT PARAMEDICS ON PROCESSES OF CARE INPATIENTS WITH ST-SEGMENT ELEVATIONMYOCARDIAL INFARCTION: A PROVINCEWIDEEVALUATION

Lucy Boothroyd, Yongling Xiao, LaurieLambert, Eli Segal, Dave Ross, SebastienMaire, Peter Bogaty,Institut National d’Excellence en Sante et enServices Sociaux

Background. In a systematic, Canadianprovincewide evaluation of ST-segmentelevation myocardial infarction (STEMI), weexamined the association of processes of careand mortality with prehospital electrocardio-grams (phECGs), early in their implementationby basic life support (BLS) paramedics.Methods. All 82 acute care Quebec hospitalsthat treated at least 30 patients with acutemyocardial infarction (AMI) participatedin a six-month evaluation in 2008–2009. Atthat time, seven of 16 provincial regions hadimplemented phECG acquisition by BLSparamedics. Hospital charts, phECGs, and firstemergency room (ER) ECGs of AMI patientspresenting to an ER with acute symptomswere systematically reviewed by medicalrecord librarians and two cardiologists toidentify STEMI and ambulance use. Survivaldata and diagnoses in the previous five yearswere identified from linkage to vital statisticsand hospital discharge databases. Timesare expressed as medians with interquartilerange (IQR). Results. Of 1,222 STEMI patientstransported by ambulance, 139 (11%) had aphECG. For 79% of these, paramedics alertedthe receiving ER or transmitted the phECG.Two-thirds of phECG patients were trans-ported directly to a percutaneous coronaryintervention (PCI) center (vs. 33% withoutphECG), and 18% were transferred to a secondhospital for PCI (vs. 50% without phECG; bothp < 0.001). Presenting patient characteristics(including anterior AMI and ThrombolysisIn Myocardial Infarction [TIMI] risk index),cardiac and noncardiac comorbidities, andsymptoms-to-ER delay did not differ forpatients with and without phECGs. In-hospitaldelays were shorter for phECG patients, withminimal increase (3 min) in prehospital times:door-to-ECG time was 5 minutes (IQR = 2–10)vs. 8 minutes (2–15) without phECG (p <0.001); door-to-needle time was 19 minutes(17–29) vs. 26 minutes (20–45) (p = 0.09); door-to-balloon time was 50 minutes (37–76) vs. 95minutes (71–123) (p < 0.001). There was a trendfor lower risk-adjusted one-year mortality forphECG patients. Despite similarity in patientfactors, phECG transmission versus nontrans-mission was associated with more timelyPCI (p = 0.001). Conclusions. In this initialimplementation of phECG capability in theBLS setting, critical treatment delays have beenmarkedly reduced. Given the well-establishedrelation between faster treatment and mortalityreduction, widespread introduction of phECGscan be expected to improve survival of STEMIpatients.

103. DOES PREHOSPITAL ADMINISTRATION OFNITROGLYCERIN FOR CHEST PAIN CAUSEHYPOTENSION IN ACUTE INFERIOR WALLSTEMI? A RETROSPECTIVE COHORT STUDY

Dave Ross, Laurie Robichaud, Marie-HeleneProulx, Sebastien Legare, Charlene Vacon,Xiaoqing Xue, Eli Segal,Urgences-sante; Hopital du Sacre-Coeur deMontreal

Background. Patients with inferior ST-segmentelevation myocardial infarction (STEMI), asso-ciated with right ventricular infarction, are po-tentially at higher risk of developing hypoten-sion when administered nitroglycerin (NTG).However, current basic life support primarycare paramedic (PCP) protocols do not differ-entiate location of STEMI prior to NTG admin-istration. Objective. We sought to determinewhether NTG administration is more likelyto cause hypotension (systolic blood pressure<90 mmHg) in inferior STEMI compared withnoninferior STEMI. Methods. We conducted aretrospective chart review of prehospital pa-tients with chest pain of suspected cardiac ori-gin and computer-interpreted prehospital elec-trocardiograms (ECGs) indicating “acute MI.”Computerized interpretation was performedby the GE Marquette 12SL R©-Zoll E Series. Pa-tients were treated by PCPs. We included all lo-cal STEMI cases identified as part of a provin-cial STEMI registry project. Charts were re-viewed by trained data extractors using a pre-defined instruction list. Univariate analysis wasused to compare differences in proportions ofhypotension after NTG administration, drop insystolic blood pressure greater than or equal to30 mmHg, and hypotension on initial prehos-pital blood pressure between patients with in-ferior wall STEMI and those with STEMI in an-other region (noninferior). Results. Over a 29-month period, we identified 1,466 STEMI pa-tients. Of those, 798 (54.4%) had complete dataand received NTG. Hypotension occurred afterNTG in 36 of 461 inferior STEMIs and 29 of 337noninferior STEMIs, 7.8% vs. 8.6%, p = 0.69. Adrop in systolic blood pressure greater than orequal to 30 mmHg occurred in 23.5% of infe-rior STEMIs and 23.8% of noninferior STEMIs,p = 0.91. Initial hypotension was noted in sig-nificantly more inferior STEMIs compared withnoninferior STEMIs, 9.9% vs. 4.9%, p = 0.005.Interrater agreement for chart review of the pri-mary outcome was excellent (kappa = 0.94).Conclusion. Patients with chest pain and infe-rior wall STEMI on their computer-interpretedprehospital ECG who receive nitroglycerin donot seem to develop hypotension more fre-quently than patients with STEMI in other ter-ritories, although they are more commonly hy-potensive on presentation. Current PCP pro-tocols for NTG administration in computer-interpreted prehospital ECG STEMI appear tobe safe.

104. COMPUTER 12-LEAD INTERPRETATIONINFLUENCES PARAMEDICS TO OVERTRIAGEACUTE MYOCARDIAL INFARCTION

Todd Burgbacher, Spencer Brady, CraigManifold, Christopher Velasquez, DavidWampler,University of Texas Health Science Center atSan Antonio

Background. Emergency medical services(EMS) personnel are often tasked withmaking the decision to call “heart alert,”which prepares hospitals to receive the mosttime-sensitive form of heart attacks. There issignificant hospital cost associated with heartalert activation. Historically, the computer’sinterpretation of the 12-lead electrocardiogram

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


(ECG) is hypersensitive, and overcalls acutemyocardial infarction (AMI). It is imperativethat paramedics be able to accurately interpretthe 12-lead ECG in order to reduce delay indefinitive care, thereby reducing morbidityand mortality. The objective of this study wasto determine the impact of the computer-generated diagnostic interpretation on thecorrect paramedic activation of heart alertbased on difficult ECG. Methods. Twenty-fiveECGs were selected from an in-house libraryto include a mix of false AMI determinations,appropriate AMI determinations, and missedAMI interpretations. Each ECG was duplicatedwith the computer interpretation deleted fromthe duplicate. All ECGs were formally random-ized and formatted as a test. All 50 questionscontained identical patient information (i.e.,patient was diaphoretic and complaining ofchest pain) and administered to a randomselection of on-duty paramedics. Responseswere dichotomous for patient eligibility forheart alert. Results. The test was given to42 paramedics resulting in 1,050 ECG pairsassessed. Each ECG pair had at least 1 incon-sistency with a total of 183 inconsistencies.Of the inconsistencies, 94 (51%) followed thepattern of calling heart alert with the computerinterpretation of AMI and not calling theheart alert on the same ECG if the computerinterpretation was blinded. Twenty-fivetrue AMIs were not identified without thecomputer-generated interpretation. Two ECGsrevealed AMI where the computer algorithmfailed to recognize it; the paramedics weretwice as likely to appropriately identify thoseST-segment elevation myocardial infarctions(STEMIs) when the computer interpretationwas blinded. Conclusion. Paramedics weremore likely to trigger heart alert on difficultECGs if the computer interpretation displayedAMI, rather than if the computer interpretationwas blinded. False activations of heart alertmay be reduced if computer interpretationis disabled, at the risk of missing moreAMIs.

105. TIME INTERVALS AND REARREST AFTEROUT-OF-HOSPITAL CARDIAC ARREST

Allison Koller, David Salcido, CliftonCallaway, James Menegazzi,University of Pittsburgh

Background. Rearrest (RA) occurs when a pa-tient loses pulses following return of sponta-neous circulation (ROSC) after out-of-hospitalcardiac arrest (OHCA), but the exact causes ofRA are not fully understood. Time-to-treatmentintervals may affect patient outcome and couldbe a plausible contributor to RA. Objectives.To compare emergency medical services (EMS)time intervals between cases with and withoutRA. We hypothesized that RA cases will havesignificantly longer time intervals than no-RA cases. Methods. The Institutional ReviewBoard of the University of Pittsburgh approvedthis study. Cases of EMS-treated, nontraumaticOHCA from 2006 to 2011 with at least oneinstance of prehospital ROSC were retrospec-tively gathered from the Pittsburgh site of theResuscitation Outcomes Consortium. Prehospi-tal event times were derived from computer-assisted dispatch records. We calculated timeintervals to the nearest minute, including 9-1-1 call to EMS arrival, arrival to first EMS car-diopulmonary resuscitation (CPR), EMS CPR toROSC, and 9-1-1 call to arrival at the emergencydepartment (ED). RA status was determinedfrom electronic defibrillator downloads and pa-tient care reports. We used logistic regressionto examine the association between RA andeach time interval while controlling for patient

demographic and clinical variables with alpha= 0.05. Results. Two-hundred thirty-five caseswere analyzed, and 79 (34%) cases had an in-stance of RA. The RA group had a significantlyhigher proportion of males and shocked casesthan the non-RA group. Additionally, the RAgroup had a significantly higher amount ofshocks delivered than the non-RA group. Oddsratios for the outcome of RA by time intervalwere as follows: 9-1-1 to arrival: 1.09 (confi-dence interval [CI]: 0.98–1.21, p = 0.10); arrivalto CPR: 1.02 (CI: 0.89–1.19, p = 0.74); CPR toROSC: 0.99 (CI: 0.96–1.02, p = 0.86); and 9-1-1 toED: 0.99 (CI: 0.96–1.01, p = 0.35). All other vari-ables were not significant. Conclusion. EMStime intervals for OHCA were not predictive ofRA.

106. TREATMENT AT A STEMI CENTER ISASSOCIATED WITH FAVORABLE OUTCOMESFOLLOWING OUT-OF-HOSPITAL CARDIAC ARREST

Bryn Mumma, Deborah Diercks, BeateDanielsen, James Holmes,UC Davis

Background. The American Heart Associationrecommends regionalized post–out-of-hospitalcardiac arrest (OOHCA) care at cardiac re-suscitation centers that are closely alignedwith ST-segment elevation myocardial infarc-tion (STEMI) centers. However, the effect ofSTEMI centers on outcomes following OOHCAremains unknown. We hypothesize that treat-ment at a STEMI center is associated withincreased survival and neurologic recoveryamong OOHCA patients. Methods. We in-cluded patients aged 18 years or older withOOHCA in the 2010 California EmergencyMedical Services Information Systems (CEM-SIS) database whose CEMSIS record was linkedto an inpatient record from the California Of-fice of Statewide Health Planning and De-velopment (OSHPD) database. The CEMSISdatabase is a unified emergency medical ser-vices (EMS) data-collection system, and theOSHPD database contains patient-level datafor all inpatient and emergency departmentencounters. We linked CEMSIS and OSHPDrecords using probabilistic linkage. Multiple lo-gistic regression models including age, gen-der, race, ethnicity, EMS response time, car-diac arrest rhythm, payer category, hospitalsize, teaching status, and trauma center sta-tus were used to evaluate the association be-tween STEMI center treatment and outcomes.A STEMI center was defined as one with 24-hours-a-day/seven-days-a-week percutaneouscoronary intervention capability. Good neuro-logic recovery was defined as discharge tohome, a residential care facility, a psychiatricfacility, or prison/jail, or leaving against medi-cal advice. Results. We identified 8,481 patientswith OOHCA in the CEMSIS database; 3,620were linked to OSHPD inpatient records. Themajority (74.0%) were treated at STEMI cen-ters. The median age was 71 years (interquar-tile range 60–83). Survival to hospital dischargewas similar between the two groups (26.3% vs.25.8%; p = 0.76), but good neurologic recov-ery was higher among those treated at STEMIcenters (18.2% vs. 10.4%; p < 0.0001). In theadjusted analysis, treatment at a STEMI cen-ter was associated with good neurologic recov-ery (odds ratio [OR] 2.3, 95% confidence in-terval [CI] 1.6–3.4; p < 0.0001) but not sur-vival to discharge (OR 1.3, 95% CI 0.98–1.7; p =0.07). Conclusions. Treatment at a STEMI cen-ter following OOHCA was associated with fa-vorable recovery outcomes. Our data suggestthat regionalized post-OOHCA care in accor-dance with American Heart Association guide-lines may improve neurologic recovery follow-ing OOHCA.

107. CPR QUALITY AND OUTCOMES INRESUSCITATION OF SUSPECTED DRUG-RELATEDCARDIAC ARRESTS

Allison Koller, David Salcido, CliftonCallaway, James Menegazzi,University of Pittsburgh

Background. Cardiac arrest resulting fromdrug overdose (OD) is a significant publichealth issue and is the cause of many pre-ventable deaths each year. Little is knownabout actual resuscitation process parametersof ODs, and such arrests are often excludedfrom out-of-hospital cardiac arrest (OHCA)studies. We sought to investigate the charac-teristics of emergency medical services (EMS)-treated OHCA cases resulting from suspectedOD. Methods. The University of Pittsburgh In-stitutional Review Board approved this study.Data from EMS-treated nontraumatic OHCAswere obtained from the Pittsburgh site of theResuscitation Outcomes Consortium, a multi-center clinical research consortium with 10 sitesacross North America. Cases from 2006–2011were analyzed. Case definition for OD wasnaloxone administration or drug overdose in-dicated on the patient care report. Resuscitationparameters including chest compression frac-tion, compression rate, and compression depth,shock delivery, and the administration of resus-citation drugs were collected and compared be-tween OD and non-OD groups. Demographicand outcome variables including age, gender,EMS- or bystander-witnessed status, return ofspontaneous circulation, and survival to hospi-tal discharge were also compared by OD status.Resuscitation parameters were compared be-tween groups using logistic regression or t-testswith alpha = 0.05. Results. We identified 180OD cases and compared them with 2,162 non-OD OHCAs. Both groups were predominantlymale (OD: 66%, non-OD: 59%; p = 0.0498). ODOHCAs were on average 20 years younger thannon-OD OHCAs (45 vs. 65 years, p < 0.001).OD cases were associated with higher overallcardiopulmonary resuscitation (CPR) fractionthan non-OD (66% vs. 63%, p = 0.0158) andsurvival (OD: 19%, non-OD: 11%; odds ratio[OR]: 1.82, CI: 1.12–2.94, p = 0.015). OD caseswere also associated with higher probability ofepinephrine (OR: 2.05, CI: 1.33–3.15, p < 0.001)and sodium bicarbonate administration (OR:2.70, CI: 1.98–3.68, p < 0.001). Other variablesdid not differ. Conclusion. Patients with OD-related OHCA were more likely to receive re-suscitation drugs, receive higher CPR fraction,and survive than non-OD OHCAs.

108. THE TIMING OF PREHOSPITAL INITIATION OFTHERAPEUTIC HYPOTHERMIA INTRA-ARRESTVERSUS POSTARREST: A RANDOMIZEDCONTROLLED TRIAL

Jonathan Studnek, John Garrett, SteveVandeventer, David Pearson,Mecklenburg EMS Agency

Background. This study assessed the frequencyof return of spontaneous circulation (ROSC)and survival to hospital discharge among pa-tients experiencing out-of-hospital cardiac ar-rest (OOHCA) randomized to receive eitherintra-arrest therapeutic hypothermia (IATH)or postarrest therapeutic hypothermia (TH).Methods. This was a single-center randomizedcontrolled trial performed in a municipal emer-gency medical services agency from Septem-ber 1, 2011, to May 31, 2012. Patients were in-cluded in this study if they had a nontrau-matic OOHCA and were >18 years of age.All arrest rhythms were randomized to a treat-ment arm. Patients were randomized to induc-tion of therapeutic hypothermia via 4◦C normalsaline (NS) bolus in two arms: intra-arrest ver-sus postarrest following ROSC. Those patients

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


randomized to the postarrest arm receiveda room-temperature saline bolus immediatelyupon intravenous/intraosseous access. Duringthe arrest period, no patient received greaterthan 1,000 mL of NS. If ROSC occurred, pa-tients received up to 2,000 mL of 4◦C NS bo-lus. The primary outcome of this study wassurvival to hospital discharge with a secondaryoutcome of ROSC. Data analysis consisted ofdescriptive statistics with chi-square analysis.Results. There were 356 patients enrolled inthis study with 13 (4.6%) excluded becauseof protocol violations or unknown outcome.Overall, 150 (43.0%) patients achieved prehos-pital ROSC and 34 (9.7%) patients survived tohospital discharge. Patients were distributedequally between study arms, with no differencein demographic or arrest characteristics. Therewas no significant difference in ROSC or sur-vival to hospital discharge based on the tim-ing of prehospital initiation of TH. Of thosepatients who achieved ROSC, 75 (50%) re-ceived IATH and 75 (50%) received postarrestTH; p = 0.89. This lack of difference persistedwhen assessing survival to discharge, with 17(50%) patients surviving in the IATH groupand 17 (50%) in the postarrest group; p = 0.97.There were 59 patients who were witnessedby a bystander and presented in ventricularfibrillation, with 16 (27.1%) surviving to dis-charge and 10 (62.5%) of those in the postar-rest group; p = 0.11. Conclusion. In this ran-domized controlled trial, the timing of pre-hospital initiation of TH was not associatedwith improved outcome. A multicenter non-inferiority study is needed to validate theseresults.

109. REGIONAL VARIATION OF OUT-OF-HOSPITALCARDIAC ARREST (OHCA) SURVIVAL: DEFININGMEDICALLY UNDERSERVED AREA FOR OHCA

Won chul Cha, Jung Eui Rhee, Yu Jin Kim,Suk Ran Yeom, Sun Hyu Kim, Hyun Ho Ryu,Samsung Medical Center

Background. Out-of-hospital cardiac arrest(OHCA) outcome varies greatly even amongadjacent communities. In this study, we aimedto evaluate regional characteristics as indepen-dent factors influencing outcome of regionalOHCA survival. Methods. We used the na-tional OHCA cohort database from 2006 to2008. We included emergency medical services(EMS)-assessed cardiac arrest with presum-ably cardiac origin in GyeonGi-Province. Ut-stein factors were extracted from the OHCAdatabase. The study region has 10 million res-idents in 31 counties. One hundred nineteenambulances provide sole, single-tiered prehos-pital service, which is based on the fire de-partment. Ambulances are staffed with emer-gency medical technician (EMT)-Intermediatesand EMT-Basics. We performed a multivari-ate analysis with two regional factors (popula-tion density and presence of high-volume cen-ter) along with Utstein factors predicting sur-vival of OHCA victims. Results. Overall, 8,347cardiac arrests were included for the study.The median age of the patients was 68 years(interquartile range [IQR]: 53–78), and 3,113(37.3%) were female. Of the cardiac arrests,7.50% occurred in public places; 43.7% werewitnessed, but only 2.40% received bystanderCPR; and 4.79% showed an initial shockablerhythm. The median response interval was 7minutes (IQR: 5–9). The rate of ROSC was19.6%, and the rate of survival to discharge was3.04%. There was significant variation in therate of survival to discharge among countiesfrom 0.0% to 7.2% (IQR: 0.7–2.7). The multi-variate logistic regression predicting survival todischarge revealed an odds ratio (OR) of 1.031(95% confidence interval [CI]: 1.007–1.055) for

every 1,000 increment of population density.The adjusted OR for the presence of emergencycenters was 1.69 (95% CI: 1.24–2.33), even afteradjustment for destination hospital level andtransport time. However, when put together inthe regression model, only presence of emer-gency centers showed significant association(1.56; 95% CI: 1.08–2.25) due to strong inter-action (p < 0.001) between emergency centersand population density. Conclusion. Survivalof OHCA patients varied significantly evenamong adjacent counties. Low population den-sity along with absence of a low-volume centercan be used to define medically underservedareas for OHCA.

110. EFFECT OF THE 2010 GUIDELINES ON TIMETO DRUG ADMINISTRATION DURINGRESUSCITATION AND OUTCOMES FOROUT-OF-HOSPITAL CARDIAC ARREST PATIENTS

Leticia Huynh, David Salcido, Allison Koller,James Menegazzi,University of Pittsburgh School of MedicineDepartment of Emergency Medicine

Background. The American Heart Association(AHA) 2010 Advanced Cardiac Life Support(ACLS) Guidelines placed new emphasis onthe importance of rapid time to drug adminis-tration, moving vascular access in front of ad-vanced airway management in the suggestedorder of interventions. We hypothesized thatthere would be a decrease in time to drug ad-ministration following publication of the 2010AHA Guidelines and an increased proportionof intraosseous (IO) access utilization. We alsohypothesized that improved patient outcomemarkers would be associated with decreasedtime to drug administration. Methods. We per-formed a retrospective analysis of data for car-diac arrest patients who were registered in asingle Regional Clinical Center of the Resusci-tation Outcomes Consortium (ROC) databasebetween November 2006 and April 2011. Us-ing Microsoft Excel and STATA12, we trackedmean time to drug administration on a monthlybasis and proportion of IO vs. intravenous (IV)drug administration. We used two-tailed t-testsand chi-square tests to compare pre- and post-guideline data. We also tracked patient survivaland Modified Rankin Scale (MRS) score andperformed multiple logistic regression analysesin order to assess the association between timeto drug administration and outcome variables.Results. We analyzed 1,685 cases (1,321 casesbefore and 364 cases after publication of theGuidelines). The mean time to drug adminis-tration before Guidelines publication was 10.17(6.19) minutes and after publication was 10.16(5.84) minutes, with p = 0.93. The proportionof cases involving IO access only before Guide-lines publication was 0.09 and after publicationwas 0.31, with p < 0.001. The proportion ofcases involving both IO and IV did not changesignificantly after Guidelines publication. Lo-gistic regression results indicated that each ad-ditional minute between arrival and drug ad-ministration resulted in an 8% decrease in sur-vival probability with p = 0.002. There wasno association between time to drug admin-istration and MRS score. Conclusions. In thisROC Regional Center, there was no decreasein time to drug administration after publica-tion of the Guidelines. However, time to drugadministration was already fairly rapid priorto publication of the Guidelines. There was asignificant increase in proportion of cases us-ing IO access after Guidelines publication, withIO use more than doubling. Additionally, wefound time to drug administration to be a sig-nificant predictor of survival, but not of MRSscore.

111. DOES PREHOSPITAL CONTINUOUS POSITIVEAIRWAY PRESSURE IMPACT THE RATE OFINTUBATION AND MORTALITY OF ACUTERESPIRATORY EMERGENCIES?

Sheldon Cheskes, Linda Turner, SueThomson, Nawfal Aljerian,Sunnybrook Centre for Prehospital Medicine

Background. Previous small studies havedemonstrated decreased rates of intubationand mortality with prehospital use of con-tinuous positive airway pressure (CPAP). Wesought to validate these findings in a largerobservational study. Methods. We conductedan observational study of patients transportedby emergency medical services (EMS) duringthe 12 months before and the 12 months fol-lowing implementation of a prehospital CPAPprotocol for acute respiratory distress. This24-month consecutive period ended June 24,2010. Included were all patients transportedby EMS meeting preestablished criteria indica-tive of acute respiratory distress and CPAP use(patient’s problem specified as cardiac, respi-ratory distress, respiratory disease, or conges-tive heart failure (CHF); age = 12 years; chestsounds documented as wheezes or rales; Glas-gow Coma Scale score = 11; respiratory rate =24 breaths/min; systolic blood pressure = 90mmHg; oxygen saturation <90%). Data wereabstracted from ambulance call reports (ACRs)and hospital records. All cases in which a do-not-resuscitate (DNR) order was documentedon the patient chart or ACR or whose in-hospital outcome (death or discharge) was un-known were excluded. Results. In all, 442 pa-tients met the above criteria. The mean (± stan-dard deviation) age was 73.0 (±13.9) years, and51.5% were women. The in-hospital mortalityrates did not differ for these patients: 17 of228 (7.5%) in the “before” group and 17 of 214(7.9%) in the “after” group (p = 0.72, ß 0.60to find an absolute difference of 6%). Althoughlacking power for statistical significance, ananalysis of the subgroup that had a hospital di-agnosis of chronic obstructive pulmonary dis-ease (COPD), CHF, or pulmonary edema (n =273) showed that mortality was lower in the“before” group (3/138, 2.2%) than in the “af-ter” group (8/135, 5.9%) (Fisher’s exact test, p= 0.13). No patients in either group were intu-bated in the prehospital setting, and in-hospitalintubation rates were similar for both groups(11.6% vs. 9.7%; chi-square = 0.26, p = 0.61).Conclusion. In contrast to previous studies, wewere unable to demonstrate either a decrease inintubation or mortality related to the use of pre-hospital CPAP. Our findings may be specific toour EMS system but suggest that further large-scale randomized controlled trials may be war-ranted to firmly establish the benefit of prehos-pital CPAP.

112. A BEFORE–AFTER STUDY TO EVALUATE THEEFFECTIVENESS AND USEFULNESS OFPREHOSPITAL NONINVASIVE VENTILATION IN ANURBAN SETTING

Andrew Willmore, Richard Dionne, IanStiell, University of Ottawa

Background. Noninvasive ventilation (NIV) iscommonly used in the treatment of acute de-compensated heart failure (CHF) and chronicobstructive pulmonary disease (COPD) exac-erbations. In-hospital evidence is robust: NIVhas been shown to improve respiratory sta-tus and reduce intubation rates. There is lessevidence on prehospital NIV, although emer-gency medical services adoption of this modal-ity is increasing. The objectives of this studywere as follows: 1) to measure the impact ofprehospital NIV on morbidity, mortality, andtransport times; and 2) to audit the selection of

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


patients by medics for appropriateness andsafety. Methods. We conducted a before–afterstudy from August 1 to October 31 in 2010 and2011, before and after the implementation ofprehospital NIV in a city of 1 million peoplewith large rural areas. Medics were trained toapply NIV to patients with respiratory distressand a presumed diagnosis of CHF or COPD.Charts were selected using the search criteriaof chief complaint of shortness of breath, emer-gent transport to hospital, and receipt of NIV inthe field. Data were extracted from ambulanceand hospital records and were analyzed withappropriate univariate statistics. Results. Weenrolled 373 patients (186 in the pre-NIV group,187 in the post-NIV group), with mean age71.5 years, female 51.4%, and final diagnosesof CHF 18.9%, and COPD 21.9%. Characteris-tics and transport times were similar betweengroups. In the post-NIV group, of 84 patientsmeeting NIV criteria, 41.6% actually receivedit; of 102 patients not meeting criteria, 5.2%received NIV. There were 12 adverse eventsdocumented from a total of 36 NIV applica-tions (33.3%), all of them minor. Comparing thepost-NIV with the pre-NIV groups, there werehigher rates of ED NIV administration (20.0%vs. 13.4%, p < 0.0001) and higher overall mor-tality (18.8% vs. 14.9%, p < 0.0001). There wasno difference between groups in rates of ED in-tubation (2.1 vs. 2.3%, p < 0.001) or hospitallength of stay (6.8 vs. 8.7 days, p = 0.24). Con-clusion. NIV was applied to a relatively smallproportion of patients meeting criteria. No pa-tient safety issues were identified. With respectto morbidity, mortality, and hospital length ofstay, prehospital NIV failed to show benefit inthis urban setting.

113. THE PREHOSPITAL ADMINISTRATION OFSTEROIDS IN MODERATE TO SEVERE ASTHMAAND COPD EXACERBATIONS DOES NOT REDUCEPATIENT ADMISSIONS

Andrew Stevens, Chelsie Baughman,Michael Daum, Mary Ann Kozak, BruceTilson, Dan O’Donnell, Indiana UniversitySchool of Medicine

Background. Asthma and chronic obstructivepulmonary disease (COPD) exacerbations area common reason for emergency medical ser-vices (EMS) activations. A small set of literaturesuggests decreased admission in asthma pa-tients given steroids prehospital. These studieswere limited by sample size and included onlyasthma. To our knowledge, no research existsexamining the effect of prehospital steroids onasthma and COPD exacerbations. We hypoth-esized that prehospital utilization of steroidsin asthma and COPD patients reduces hospitaland intensive care unit (ICU) admissions. Ob-jective. Determine the role of systemic steroidsin moderate to severe prehospital asthma andCOPD exacerbations. Methods. We performeda before-and-after observational two-group co-hort study. This study was conducted using alarge urban EMS system with patients trans-ported to two academic medical centers withover 100,000 emergency department (ED) vis-its per year. Patients were included in thestudy with a history of asthma or COPD andrequiring >1 nebulized medication or contin-uous positive airway pressure. Exclusion crite-ria were history of congestive heart failure ordyspnea attributed to other causes. Our controlgroup was all patients not receiving steroids forthe period January to July 2011. Our experimen-tal group, conducted during the trial period,January to July 2012, included patients receiv-ing prednisone or methylprednisolone basedon new prehospital protocols implemented Jan-uary 1, 2012. Primary outcome was rate of hos-pitalization before and after the implementa-tion of prehospital steroids. A secondary data

point was ICU admissions. Results. The twogroups were similar in respect to age (average55 years), gender, race, and respective asthmaand COPD histories. In our before group, 131patients were identified and 87 (66.4%) wereadmitted. Of these, 25 (19.1%) were admit-ted to the ICU. In the after group, 66 patientswere treated and 50 (75.8%) were admitted.Of these, seven (10.6%) were admitted to theICU. The overall admission rate before and af-ter increased 9.4% (p = 0.179), with a non-ICUadmission rate increase of 17.9% (p = 0.019).The ICU admission rate decreased 8.5% (p =0.127). Conclusion. Patients with moderate tosevere asthma and COPD exacerbations receiv-ing prehospital steroids have increased hospi-talization but lower ICU admission. This maybe attributable to prehospital providers’ iden-tifying patients with more severe conditionsas needing steroids and the intervention’s de-creasing the need for intensive care.

114. BLOOD IS BLOOD: VALIDATION OFSCAPULAR CAPILLARY BLOOD GAS (SCAPGAS)SAMPLING TO DETERMINE SERUM PH

William Bozeman, Patti Kriegel, JennyKneezel, Christine Hall,Wake Forest University School of Medicine

Background. Assessment of metabolic statusby measurement of serum pH is essential inmany critically ill patients. However, standardarterial blood gas (ABG) analysis sampling canbe hazardous or impossible in combative orseizing patients, and is typically unavailable inthe EMS setting. Capillary blood gas samplingis a validated alternative method to standardABG sampling that is commonly employed inpediatric populations. Scapular region capil-lary blood gas (ScapGas) sampling is easily per-formed in restrained but uncooperative adultsin the hospital or prehospital setting, and haspreviously been validated in healthy volun-teers. Methods. A prospective controlled trialwas performed to determine whether the pHmeasured by analysis of ScapGas samples wasin clinical agreement with the pH on standardABG samples. Subjects were adult emergencydepartment patients who received ABG analy-sis as part of their clinical care. Samples wereobtained within 5 minutes of each other; point-of-care (POC) analysis was performed. Clini-cally acceptable agreement was defined a pri-ori as pH values within ±0.05 of each other. Re-sults. Forty-one patients with matched pairs ofScapGas and standard ABG samples were en-rolled. The mean difference in pH was 0.022(standard deviation [SD] 0.043, 95% confidenceinterval 0.009–0.036). Agreement of pH valueswithin 0.05 was present in 33 of 41 cases (80.5%;95% confidence interval [CI] 65.1–91.2). Lactatevalues were also compared in a subgroup of 15subjects with matched pairs of lactate measure-ments. These were also similar between the twosampling techniques, with a mean difference of0.39 (SD 0.40, 95% CI 0.17–0.61). Conclusions.ScapGas sampling and POC analysis producespH measurements with clinically acceptableagreement compared with standard ABG sam-pling and analysis. This sampling techniquemay have utility in both hospital and prehos-pital settings to determine serum pH and guideclinical care in combative or uncooperative pa-tients.

115. DO ADULT PATIENTS HAVE MOREPREHOSPITAL DATA REPORTED THAN CHILDREN?

Lara Rappaport, Jesse Hawke, JosephDarmofal, Kathleen Adelgais,Children’s Hospital Colorado

Background. The Institute of Medicine Emer-gency Services Report and guidelines devel-oped by the American College of Surgeons ad-

vise that triage should begin with measure-ment of vital signs (VS). Local and state emer-gency medical services (EMS) pediatric pro-tocols indicate VS measurement and monitor-ing for many common conditions. Previousstudies have demonstrated that incomplete VSmonitoring during prehospital care is com-mon in children with acute illness and in-jury such as trauma and traumatic brain in-jury (TBI). Objective. The purpose of this studywas to compare rates of obtaining prehospi-tal VS data (oxygen saturation, blood pres-sure [BP] monitoring, heart rate [HR], respi-ratory rate [RR], capnography, and GlasgowComa Scale [GCS] scoring) and intravenous(IV) line placement attempts between adultsand children. Methods. Emergency medicalservices (EMS) records from 2010 to 2011 wereextracted from the Online Matrix system inthe Colorado State Reported EMS system andanalyzed. Records were excluded for the fol-lowing reasons: no treatment required, scenerefusal, or call canceled. Outcome measure-ments were proportions of adults and chil-dren (defined as <15 years) with the follow-ing: IV placement, HR, BP, RR, oxygen satu-ration, GCS score, and capnography monitor-ing. Comparisons between age groups and be-tween the state’s 11 Regional Emergency Med-ical and Trauma Advisory Councils (RETACs)using χ2. Results. During the study period,there were 475,501 patient encounters (adult:445,883, pediatric: 29,618). The proportions ofpatients with HR, RR, BP, and oxygen satura-tions were similar between adult patients andpediatric patients (48% vs. 52%). There werelower rates of GCS scoring and capnographyamong both adults and children, with no no-table differences between age groups. Signifi-cant variation was found between regions, withsome reporting VS in >85% of patients and oth-ers as low as 10%. Placement of IV was higherin adults compared with children (30% vs. 12%,p < 0.0001). Conclusions. There was no sig-nificant difference between adults and childrenin VS monitoring in the state of Colorado;however, there was a difference in the pro-portion of patients with IV placement. Therewere significant variations between the geo-graphic regions. Further studies are needed toassess factors associated with this large vari-ation as well as the large number of missingVS among patients transported by prehospitalproviders.

116. A NOVEL PEDIATRIC WEIGHT ESTIMATIONTOOL FOR EMS PROVIDERS

Ryan Jacobsen, Jennifer Watts, SusanAbdel-Rahman, Tiffany Hefner, DonnaO’Malley, Stacey Doyle, M. Dowd,Kansas City, Missouri, EMS/Truman MedicalCenter

Background. When caring for children inthe out-of-hospital environment, accurate andrapid weight estimation is critical. A novelweight-estimation method based on humerallength and mid–upper arm circumference(Mercy method) was recently developed. Thisstudy was designed to evaluate the accu-racy, precision, and speed of two Mercymethod–based devices (two-dimensional tape[2DT], three-dimensional tape [3DT]) com-pared with five other weight-estimation meth-ods (provider estimate [PE], Advanced Pedi-atric Life Support [APLS], Broselow tape [BT],Devised Weight-Estimation Method [DWEM],and Luscombe and Owens [LO]) when usedby prehospital providers. Methods. Emergencymedical services (EMS) providers were re-cruited for this prospective study from five re-gional EMS agencies. Each provider appliedall seven weight-estimation methods, in ran-dom order, to five children of varying ages

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


and weights. The speed with which the mea-sures were performed was recorded in sec-onds. Actual weight was determined usinga calibrated scale. Accuracy was assessed bymean error (ME), calculated as predicted mi-nus actual weight (kg), and mean percentageerror (MPE), calculated as ME divided by ac-tual weight × 100. Percentages of estimateswithin 10–30% of actual weight were also deter-mined. Results. Thirty-six EMS providers (15emergency medical technician [EMT]-Basic, 21EMT-Paramedic) participated; they were 37.4± 9.5 years of age and had 12.8 ± 8.9 yearsof experience. The children averaged 7.3 ±4.7 years of age, measured 123.1 ± 32.6 cmand weighed 31.8 ± 20.8 kg, and representedall five body mass index (BMI) percent cate-gories. Mean method speeds (seconds) were asfollows: PE (27.3 ± 14.5); APLS (15.7 ± 13.2); BT(43.1 ± 18.9); DWEM (50.7 ± 17.2); LO (15.2 ±6.8); 2DT (68.1 ± 21.4); and 3DT (63.1 ± 23.6).ME for each method: PE (0.1); APLS (–9.2); BT(–4.4); DWEM (–3.7); LO (–2.7); 2DT (–1.0); and3DT (–2.3). MPE (%) for each method: PE (4.1);APLS (–18.7); BT (–7.2); DWEM (–6.1); LO (1.5);2DT (–0.6); and 3DT (–6.1). Percentages of esti-mates within 10%, 20%, and 30%, respectively,of actual weight: PE (22, 42, 66); APLS (24, 41,62); BT (44, 69, 84); DWEM (37, 74, 83); LO (33,68, 80); 2DT (51, 75, 89); and 3DT (44, 73, 90).Conclusion. The 2DT and 3DT, while slower,appear to provide accurate weight estimation.Performance in an emergency setting and/orwith additional provider training needs to beevaluated.

117. CONTINUOUS CHESTCOMPRESSION–CARDIOCEREBRALRESUSCITATION IN NONPRIMARY CARDIACARREST

Jennifer Kimbrell, Joseph Salomone,Emergency Physicians of Springfield

Background. Continuous chest compression–cardiocerebral resuscitation (CCC-CCR) proto-cols have shown improved return of sponta-neous circulation (ROSC) and survival for pri-mary out-of-hospital cardiac arrest (OOHCA).Our protocol emphasizes continuous chestcompressions and passive oxygenation, andprevious reports demonstrated its efficacy.There are no reports for CCC-CCR protocolsfor nonprimary cardiac arrest (NPCA) patients.Objective. To describe the effect of a CCC-CCR protocol on ROSC and survival to hos-pital discharge for nonprimary OOHCA com-pared with a historical cohort. Methods. A ret-rospective, observational cohort extracted fromthe emergency medical services (EMS) car-diac arrest database was reviewed to iden-tify all adult (18 years and older), nontrau-matic, and nonprimary cardiac arrests fromJuly 2008 through June 2010 (CCC-CCR) andcompared with data from our EMS system us-ing the previous standard cardiopulmonary re-suscitation (CPR) from January 2003 to March2006 (standard). Examples of NPCA classeswere gastrointestinal bleeds, respiratory ar-rests, and toxic ingestions. Our system is a Mid-west, urban, all–advanced life support ambu-lance service with basic life support fire de-partment first response. Results. Overall, therewere a total of 203 and 156 NPCA patientsin the standard and CCC-CCR cohorts, respec-tively. The ROSC rates for ventricular fibril-lation/ventricular tachycardia (VF/VT) were28.6% (n = 2/7) standard and 28.6% (n = 4/14)CCC-CCR, relative risk (RR) 1.0, 95% confi-dence interval (CI) 0.24–4.20, p 1.00. The sur-vival rates for VF/VT were 0% (n = 0/7) stan-dard and 28.6% (n = 4/14) CCC-CCR, RR 0.0,95% CI 0.0–infinity, p 0.255. The ROSC rates forpulseless electrical activity (PEA) were 34.3%

(n = 49/143) standard and 38.9% (n = 21/54)CCC-CCR, RR 1.22, 95% CI 0.64–2.33, p 0.617.The survival rates for PEA were 10.5% (n =15/143) standard and 18.5% (n = 10/54) CCC-CCR, RR 1.91, 95% CI 0.80–4.56, p 0.155. TheROSC rates for asystole were 18.9% (n = 10/53)standard and 14.8% (n = 13/88) CCC-CCR, RR0.75, 95% CI 0.30–1.84, p 0.639. The survivalrates for asystole were 1.9% (n = 1/53) stan-dard and 0% (n = 0/88) CCC-CCR, RR notapplicable, 95% CI 0–infinity, p 0.376. Therewas no significant difference in mean age, by-stander CPR, gender, or response time. Con-clusion. This was a retrospective cohort studywith all of the inherent limitations. This co-hort was obtained from a single EMS sys-tem. The data pool was small, thus the studywas underpowered. There were no differencesin ROSC or survival for patients treated withthe standard protocol versus the CCC-CCRprotocol. Further investigation of NPCA isneeded.

118. AMBULANCE PATIENTS ARE MORE (OR LESS)LIKELY TO BE INSURED

Jeffrey Dixon, Annette Arthur, EmilyWilliams, Jeffrey Goodloe, Stephen Thomas,University of Oklahoma

Background. Ambulance diversion, whichis increasing in some emergency medicalservices regions, has a number of knowndisadvantages. From the hospital perspective,one potential disadvantage that may differ invarying regions is the loss of patients who aremore likely to be insured. If patients arriving tothe emergency department (ED) by ambulanceare less likely to be uninsured (“self-pay”) thannonambulance patients, ED leaders can usethis information to lobby for resources neededto avoid diversion. The purpose of this studywas to determine whether, for a single ED, am-bulance patients were more (or less) likely tobe uninsured as compared with nonambulancepatients. Methods. An administrative databasefrom January 2011 through March 2012 wasused from the study center (700-bed teachinghospital ED with annual census roughly50,000) to assess numbers of patients and modeof arrival to the ED. No other informationwas assessed and no identifiers were assayed.Binomial exact 95% confidence intervals (CIs)were calculated for the proportion of uninsuredpatients in the ambulance and nonambulancecohorts. The statistical association between ar-rival mode (ambulance versus nonambulance)and dichotomous insurance status was con-ducted using chi-square testing and the cohortstudy epidemiology function in STATA 12MP(StataCorp, College Station, TX). The measureof relative risk was the risk ratio (RR), reportedwith 95% CI. Results. For the 15-month studyperiod, there were 18,072 ambulance patientsand 58,594 nonambulance patients. The overallproportion of noninsured status for ambulancepatients was lower than the proportion ofnoninsured status for nonambulance patients.For ambulance cases, the uninsured ratewas significantly (p < 0.0001) lower than fornonambulance cases: 23.4% (95% CI 22.8% to24.0%) versus 28.7% (95% CI 28.2% to 29.0%).Ambulance cases were nearly 20% less likelyto be uninsured (RR 0.82, 95% CI 0.79 to 0.84, p< 0.0001). Conclusion. For the specific hospitalstudied, ambulance patients were significantlyless likely than nonambulance patients to beuninsured. These data are being used in ongo-ing conversations with hospital administrationregarding providing the ED with sufficientresources to avoid diversion status. Other cen-ters may benefit from performance of similaranalysis.

119. RELATIVE INFLUENCE OF DIFFERENTSTRESSES ON PTSD IN A CANADIAN EMSSERVICE

Elizabeth Donnelly, Paul Bradford, RandyMellow, Cathie Hedges, Peter Morassutti,University of Windsor

Background. Emergency medical services(EMS) providers are regularly exposed to avariety of stressors endemic to the profession,all of which may contribute to stress reactionssuch as posttraumatic stress disorder (PTSD).These stressors may be related to the provisionof patient care (critical incident stress), theorganization and the culture in which theresponder is working (organizational stress),or the stresses associated with working onan ambulance (operational stress). Previousresearch has identified a relationship betweenoperational stress, organizational stress, criticalincident stress, and PTSD; however, it isunclear whether this relationship persists inthe Canadian context. The objective of thisstudy was to investigate how different typesof occupationally related stress may contributeto stress reactions for paramedics working ina county-based service in southwest Ontario.Methods. All paramedics in a municipally op-erated service (annual call volume 80,000) wereinvited to complete a 167-item online surveyexamining self-reported levels of operationalstress, organizational stress, critical incidentstress, posttraumatic stress symptomatology(PTSS), and demographic characteristics.Pearson correlation coefficients were used toestimate linear dependence between stressvariables. Ordinary least-squares regressiondetermined predictor variables independentlyassociated with PTSS. Results. One hundredforty-five paramedics (a 54% response rate)completed the questionnaire. Analysis revealeda significant relationship between operationalstress (r = 0.508; p < 0.001), organizationalstress (r = 0.419, p < 0.001), and criticalincident stress (r = 0.433, p < 0.001) andPTSS. When controlling for demographicfactors, operational stress was independentlyassociated with PTSS (p < 0.001); an interactioneffect between operational stress and criticalincident stress (p = 0.001) created a robust finalmodel with an R2 of 0.391. Conclusion. In theCanadian context, exposure to a multiplicityof stressors increases the risk of paramedics’developing a posttraumatic stress reaction.Operational stress independently increases therisk for a posttraumatic stress reaction; criticalincident stress interacts with operational stressto further exacerbate the risk. These findingsindicate that health and wellness initiativesshould address the impact of both criticalincident stress and chronic work-relatedstress. Further, these findings illustrate theneed for the development and validation ofevidence-based interventions addressing themultiplicity of factors that can contribute to thedevelopment of stress reactions in paramedics.

120. EMS HELICOPTER EFFECTIVENESS: CANAUTO-DISPATCH IMPROVE EFFECTIVENESS WHILEMAINTAINING APPROPRIATE UTILIZATION OFEMS FLIGHT SERVICES?

Noel Wagner, Philip Sloan,CMU Healthcare

Background. It is well known that decreasingthe time from injury to definitive care corre-lates with a decrease in adverse outcomes fortrauma patients. Several changes have been im-plemented over the past several decades thathave been aimed at decreasing this time. Theuse of emergency medical services (EMS) sceneflights for trauma is one change that has beensuccessful in minimizing time to arrival atdefinitive care. Despite this success, there are

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


still areas in the typical EMS notification, dis-patch, and response system where adjustmentscan be made for further benefit. Our study fo-cuses on the time required to initiate the dis-patch of flight services to trauma scenes andcriteria that can be used to simultaneously dis-patch both ground EMS and flight services tothose scenes where potential exists for injuriesthat require rapid transport to definitive care fa-cilities. We theorize that by using prearrangedcriteria, dispatchers can simultaneously dis-patch these complementary services to appro-priate scenes, resulting in more prompt EMSflight arrival and a shorter time to definitivecare without resulting in an increased numberof unnecessary EMS flights. Methods. We putin place a set of criteria that would prompt dis-patch to initiate simultaneous launch of groundand flight EMS units. These criteria were acti-vated on January 1, 2011. In a retrospective re-view of records, data were collected from Jan-uary 1, 2010, to December 31, 2010, and fromJanuary 1, 2011, to December 31, 2011. Thedata were organized into several time intervals(call–dispatch, dispatch–arrival, etc.) and com-pared to evaluate the effect of the autolaunchprotocols. Times were averaged for the sepa-rate periods and compared using t-test for sta-tistical significance. Results. Fifty percent of the2011 flights met autolaunch criteria. The aver-age time lapse from call to flight dispatch wasdecreased by 3 minutes (p = 0.025). There wasno significant decrease in overall injury severityof patients transported (Glasgow Coma Scalescore; p = 0.307) (Revised Trauma Score; p =0.748). There was not significant change in thenumber of canceled flights (p = 0.499). Conclu-sions. Autolaunch criteria can successfully beused to simultaneously dispatch ground andflight EMS units without significant overuti-lization of flight EMS. Future studies may aidin identifying other areas of the typical dispatchprocedure where further time-saving protocolscan be initiated.

121. DEVELOPING A DYNAMIC RESTATIONINGAMBULANCE MODEL AND PILOT STUDY

Yu Jin Lee, Won Chul Cha, Ki Jeong Hong,Chu Hyun Kim, Kyoung Jun Song, Sang DoShin, Researcher

Background. Since requests for ambulance ser-vice are increasing every year, it is becomingmore difficult to sustain the appropriate re-sponse interval. Moreover, difference of work-load among ambulance teams is making it moredifficult to keep the appropriate service level.In this study, we aimed to test a new dispatchsystem with computer simulation, and thenvalidate it. Methods. This was a before-and-after study, comparing the stationing model ofambulances. The study area was a rural areawith small town in the center. The size wasabout 353 km2 with a population of 190,000.There were four stations and six ambulancesdispatched by a centralized dispatch center ina single-tiered system. Each ambulance had afixed base. First, we developed a dynamic resta-tioning strategy by which ambulances do nothave a fixed station. When any of four stationsbecame empty, available ambulances were re-stationed to fill the positions. After testing thestrategy with discrete event simulation, we per-formed an in vivo test from September 15 toOctober 29, 2010. Ambulance log data such ascall time, start time, scene arrival time, hospi-tal arrival time, and back time were extractedto measure outcome values. The control phasecomprised the same dates in 2009. The primaryoutcome was cases with coefficient of varia-tion (CV), which means discrepancy of work-load among ambulances. The secondary out-come was average response interval. Results.A total of 1,751 cases were collected: 896 cases

during the study phase and 855 cases duringthe control phase. The mean age of the patientswas 48.4 years (p = 0.79), and 61.0% were male(p = 0.22). The mean response time was 10.6minutes (standard deviation [SD] = 6.5) in thestudy phase and 10.1 minutes (SD = 6.3) inthe control phase (p = 0.16). The CV decreasedfrom 0.35 to 0.15. In the same period of 2009,the mean transfer rate was 3.4 cases/day (SD =1.2) and the CV was 0.35, and the mean worktime was 3.9 hours/day (SD = 0.68) and the CVwas 0.17. In the pilot study period, the meantransfer rate was 3.6 cases/day (SD = 0.2) andthe CV was 0.06, and the mean work time was3.5 hours/day (SD = 0.1) and the CV was 0.03.Conclusion. It is significant that this new resta-tioning strategy was effective to equalize work-load while keeping the same response time.

122. MEASURING THE QUALITY OF EMERGENCYMEDICAL SERVICES CARE USING COMPOSITESCORES

Patrick Chow-In Ko, Yi-Cheng Lu, Fen-RuChen, Wen-Chu Chiang, Matthew Huei-MingMa, Tsung-Tai Chen,National Taiwan University Hospital

Background. Policymakers may need to ratethe quality of emergency medical services(EMS) in terms of its performance. However,the quality of health care is usually a multi-dimensional construct that may not be mea-sured directly. Thus, multiple indicators areused to construct a composite score for qual-ity measurement. We investigated the associ-ation between the different composite scoresderived from EMS process measures and dis-charge mortality. Methods. Data were collectedfrom an urban out-of-hospital cardiac arrest(OHCA) registry from the period January 1,2006, to December 31, 2009. The compositescores in our study were derived from the fol-lowing two methods: 1) the raw sum score and2) the all-or-none score. These composite scoresof each EMS ambulance team were calculatedbased on two process measures (EMS responsetime <5 minutes and achieving prehospital re-turn of spontaneous circulation [ROSC]). Fi-nally, the association between the compositescores and the risk-adjusted discharge mor-tality was investigated using a mixed-effectsmodel. Results. A total of 4,000 adult non-trauma OHCA patients resuscitated and trans-ported by 44 EMS ambulance teams were an-alyzed. The all-or-none score of the EMS am-bulance team demonstrated a highest inverserelationship with risk-adjusted discharge mor-tality (–0.85, p < 0.01) compared with the rawsum score and the individual process measureof the EMS ambulance team. The two pro-cess measures (EMS response time <5 min-utes and achieving prehospital ROSC) showedlower inverse relationships with risk-adjusteddischarge mortality (–0.39 and –0.76, respec-tively). Conclusions. Applying the compositescores, especially the all-or-none score, to mea-sure the quality of care for EMS ambulanceteams shows higher validity. Thus, the compos-ite scores constructed by a number of EMS pro-cess measures may be utilized as an alternativeapproach to access and evaluate EMS qualityand performance.

123. RESCUE SHOCK TIMING AND OUTCOMESDURING THE METABOLIC PHASE OFVENTRICULAR FIBRILLATION

Ryan Coute, Timothy Mader, Scot Millay,Adam Kellogg,Baystate Medical Center/Tufts UniversitySchool of Medicine

Background. To determine whether 3 min-utes of cardiopulmonary resuscitation (CPR)and a single dose of epinephrine prior to the

first rescue shock (RS) are sufficient to achievereturn of spontaneous circulation (ROSC) af-ter 12 minutes of untreated ventricular fibril-lation (VF). Methods. This was a secondaryanalysis of prospectively collected data froman Institutional Animal Care and Use Com-mittee (IACUC)-approved protocol. Fifty-threeYorkshire swine (weighing 25–30 kg) weresurgically instrumented under anesthesia andVF was electrically induced. After 12 min-utes of untreated VF, CPR was initiated (andcontinued as needed [prn]) and a standarddose of epinephrine (SDE) (0.01 mg/kg) wasgiven (and repeated every 3 minutes prn).The first RS was delivered after 3 minutesof CPR (and every 3 minutes thereafter prn).Each failed RS was followed (in series) by va-sopressin (VASO [0.57 mg/kg]); amiodarone(AMIO [4.3 mg/kg]); and sodium bicarbonate(BICARB [1 mEq/kg]) prn. Resuscitation at-tempts continued until ROSC was achieved or20 minutes elapsed without ROSC. The pri-mary outcome measures were ROSC (systolicblood pressure [SBP] >80 mmHg for >60 sec-onds) and survival (SBP >60 mmHg for 20 min-utes). Coronary perfusion pressure (CPP) val-ues for the first two RS attempts were also cal-culated. Data were analyzed using descriptivestatistics. Results. ROSC was achieved in 30 ofthe 53 (57%) animals. Survival occurred in 28of the 53 (53%) animals. The mean preshockCPP was 27.9 mmHg (95% confidence inter-val [CI] 23.2–32.5) for the first RS (RS1) and49.8 (95% CI 43.4–56.2) for RS2. ROSC oc-curred in one animal following RS1 (3.3%; 95%CI 0.6–16.7), 17 animals following RS2 (56.7%;95% CI 39.2–72.6), five animals following RS3(16.7%; 95% CI 7.3–33.6), and seven animalswho received ≥4 RSs (23.3%; 95% CI 11.8–40.9).Survival was achieved in one animal follow-ing RS1 (3.6%; 95% CI 0.6–17.7), 15 animals fol-lowing RS2 (53.6%; 95% CI 35.8–70.5), five an-imals following RS3 (17.9%; 95% CI 7.9–35.6),and seven animals who received ≥4 RSs (25%;95% CI 12.7–43.4). Conclusion. Our data sug-gest that following 12 minutes of untreated VF,3 minutes of CPR and one SDE may be insuffi-cient to achieve ROSC on the first RS attempt.A longer duration of CPR and/or administra-tion of additional vasopressors may result inmore favorable conditions for successful defib-rillation on the first attempt.

124. IMPROVEMENT IN SURVIVAL FOROUT-OF-HOSPITAL CARDIAC ARREST INSINGAPORE OVER 10 YEARS

Marcus Ong, Caroline Choong, E. ShaunGoh, Benjamin Leong, Han Nee Gan, DavidFoo, Lai Peng Tham, Rabind Charles,Stephanie Fook-Chong,Singapore General Hospital

Background. Many efforts have been putin place to improve resuscitation for out-of-hospital cardiac arrests (OHCAs) in Singa-pore over the past 10 years. We aimed tostudy whether survival from OHCA has in-creased and which factors contributed to sur-vival. Methods. This was a cohort study thatcompared OHCA cases from the Cardiac Ar-rest and Resuscitation Epidemiology (CARE)project (October 2001–October 2004) with thosefrom the Pan-Asian Resuscitation OutcomesStudy (PAROS) project (April 2010–June 2011).Survival outcomes were adjusted for age, gen-der, and history of heart disease and were ex-pressed in terms of the odds ratio (OR) andthe corresponding 95% confidence interval (CI).Differences in resuscitation efforts were ex-pressed in terms of the p-value. Results. A to-tal of 2,428 cases from the CARE data and 1,514cases from the PAROS data were used for theanalysis. New interventions introduced overthe period included public-access defibrillation

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.


(PAD), motorcycle first responders, laryngealmask airways, intravenous epinephrine, me-chanical cardiopulmonary resuscitation (CPR),and therapeutic hypothermia. Survival to ad-mission increased from 9.0% to 15.8% (ad-justed OR 3.0; 95% CI 2.1–4.2) and overallsurvival to discharge increased from 1.6% to2.4% (adusted OR 4.1; 95% CI 1.7–11.5). Ut-stein survival (witnessed ventricular fibrilla-tion/ventricular tachycardia) increased from2.5% to 8.3% (adjusted OR 7.6; 95% CI 1.8–51.5).Ambulance response times decreased from 9.0to 7.8 minutes (p ≤ 0.001), but bystanderCPR rates remained similar (19.7 vs. 21.3, p= 0.2). Factors favoring survival included by-stander CPR (OR 2.3; 95% CI 1.2–4.4), PAD(OR 11.1; 95% CI 1.2–51.6), motorcycle firstresponders (OR 4.6; 95% CI 1.6–11.1), ambu-lance defibrillation (OR 6.0 95% CI 3.5–10.7),and therapeutic hypothermia (OR 28.9; 95% CI4.7–131.0). Conclusion. Survival from OHCAhas increased over the past 10 years. However,more can be done to increase bystander CPRand the use of PAD and improve resuscitationefforts.

125. CARDIOLOGIST INTERPRETATION OFPREHOSPITAL ECG IMPACTS ACCURACY OF EMSSTEMI ACTIVATIONS

Louis Gonzales, Kayla Riggs, Frank Zidar,Robert Wozniak, Osvaldo Gigliotti, JoseCabanas, Paul Hinchey,Office of the Medical Director, Austin/TravisCounty EMS System

Background. Prehospital identification ofST-segment elevation myocardial infarction(STEMI) is essential for an effective STEMIsystem of care. The accuracy of STEMI elec-trocardiogram (ECG) identification is essentialto appropriate activation of the catheteriza-tion team, maintaining cardiology processcommitment, and improving the performance

of emergency medical services (EMS) STEMIcare. However, an important contributingfactor of EMS STEMI ECG accuracy is thereliability of the cardiologist interpretation ofthe ECG. The objectives of this study wereto define EMS ECG interpretation accuracyand describe the interventional cardiologist’sinterpretation of paramedic STEMI activationsutilizing a regional STEMI definition. Methods.From February 1, 2011, to June 30, 2011, weperformed a prospective study using ECGsfrom all paramedic STEMI activations in anurban/suburban EMS system. Paramedicsutilized STEMI criteria agreed upon by theAmerican Heart Association regional “Mission:Lifeline” workgroup. We sought to determinethe accuracy of ECGs meeting STEMI criteriautilizing a panel of three interventional cardi-ologists blinded to the catheterization results.For each ECG, the cardiologists were asked todetermine whether the activation met definedregional STEMI criteria. Data were collectedfor descriptive analysis and level of agreementamong the three reviewers. Results. A total of106 STEMI activations were included in theanalysis. Two or three cardiologists determinedthat STEMI criteria were met for 71 of 106 ECGs(66%). In 51 cases (48%), all three cardiologistsagreed that the ECG met STEMI criteria.Paramedic accuracy increased to 83% whenonly one or more cardiologists agreed that theECG met STEMI criteria. Overall agreementamong the three cardiologists was low (κ =0.2562; 95% confidence interval 0.1047–0.4077).Conclusion. We found low agreement betweencardiologists when determining the accuracyof paramedic STEMI activations. Paramedicaccuracy is dependent on the specific cardiol-ogist who reviews the ECG. Further studiesare needed to define an accuracy standard forEMS systems and to understand the reasonsfor cardiologist variation. Difficulty in defininga standard for EMS performance measurementremains.

126. HIGH RATE OF UNRECOGNIZED METABOLICSYNDROME IN YOUNG PARAMEDICS

Shalini Bobra, Michael Miedema, RossGarberich, David Hildebrandt, TimothyHenry,Minneapolis Heart Institute Foundation

Background. The increasing prevalence ofmetabolic syndrome and type II diabetes isa major U.S. cardiovascular health challenge,and these conditions are frequently unrecog-nized. We evaluated risk factors in paramedics,a group of young adults who directly witnessthe devastating effects of coronary artery dis-ease. Methods. Paramedics were offered freecardiovascular screening at an educational con-ference: a health screening questionnaire wasfilled out, blood pressure and waist circumfer-ence were measured, and a fasting lipid paneland glucose values were obtained. The cur-rent National Cholesterol Education ProgramAdult Treatment Panel III (NCEP/ATP III) cri-teria were used to define the components ofmetabolic syndrome. Results. Of 98 attendees,50 young adults (median age 40.8 years, range21–60) participated in the screening. Only fourreported a history of hypertension and nonehad diabetes. The components of the metabolicsyndrome were highly prevalent (Fig. 1), with34 (68%) having hypertension (≥130 beats/minsystolic or ≥85 beats/min diastolic), 29 (58%)an elevated level of triglycerides, 16 (32%)an abnormal level of high-density lipoprotein(HDL), and 29 (58%) an abnormal glucose level.In total, 26 (52%) of the participants met cri-teria for the metabolic syndrome. Seven (14%)had a fasting glucose level >125 mg/dL, poten-tially meeting criteria for type II diabetes, six ofwhom were ≤41 years old. Conclusions. In thissample of young paramedics, >50% had un-recognized metabolic syndrome. More aggres-sive screening and education regarding healthylifestyle choices are needed even among thosefrequently exposed to the cardiovascular im-pact of these risk factors.

Preh

osp

Em

erg

Car

e D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y 21

6.61

.187

.254

on

12/2

0/12

For

pers

onal

use

onl

y.

Date post:	04-Jan-2017
Category:	Documents
Upload:	dangminh
View:	222 times
Download:	0 times

NAEMSP ABSTRACTS

Documents