CLINICAL PRACTICE

Emergency Department Management ofAcute Pain Episodes in Sickle Cell DiseasePaula Tanabe, PhD, RN, MPH, Randall Myers, MS, MD, Amy Zosel, MD, Jane Brice, MD, MPH, Altaf H. Ansari,MD, Julia Evans, MD, Zoran Martinovich, PhD, Knox H. Todd, MD, MPH, Judith A. Paice, PhD, RN

AbstractObjectives: To characterize the initial management of patients with sickle cell disease and an acute painepisode, to compare these practices with the American Pain Society Guideline for the Managementof Acute and Chronic Pain in Sickle-Cell Disease in the emergency department, and to identify factorsassociated with a delay in receiving an initial analgesic.

Methods: This was a multicenter retrospective design. Consecutive patients with an emergency depart-ment visit in 2004 for an acute pain episode related to sickle cell disease were included. Exclusion criteriaincluded age younger than 18 years. A structured medical record review was used to abstract data, includ-ing the following outcome variables: analgesic agent and dose, route, and time to administration of initialanalgesic. Additional variables included demographics, triage level, intravenous access, and study site.Mann–Whitney U test or Kruskal–Wallis test and multivariate regression were used to identify differencesin time to receiving an initial analgesic between groups.

Results: There were 612 patient visits, with 159 unique patients. Median time to administration of an initialanalgesic was 90 minutes (25th to 75th interquartile range, 54–159 minutes). During 87% of visits, patientsreceived the recommended agent (morphine or hydromorphone); 92% received the recommended dose,and 55% received the drug by the recommended route (intravenously or subcutaneously). Longer timesto administration occurred in female patients (mean difference, 21 minutes; 95% confidence interval = 7to 36 minutes; p = 0.003) and patients assigned triage level 3, 4, or 5 versus 1 or 2 (mean difference,45 minutes; 95% confidence interval = 29 to 61 minutes; p = 0.00). Patients from study sites 1 and 2 alsoexperienced longer delays.

Conclusions: Patients with an acute painful episode related to sickle cell disease experienced significantdelays to administration of an initial analgesic.

ACADEMIC EMERGENCY MEDICINE 2007; 14:419–425 ª 2007 by the Society for Academic EmergencyMedicine

Keywords: sickle cell disease, pain, analgesic use, guideline, emergency department, American PainSociety

Sickle cell disease (SCD) is associated with seriousphysiologic complications and acute pain episodesthat often require care in the emergency depart-

ment (ED).1,2 This experience has been found to be frus-trating for patients.3–5 This frustration may be a result of

From the Department of Emergency Medicine (PT, RM, AZ), In-

stitute for Healthcare Studies (PT), Mental Health Services and

Policy Program (ZM), and Division of Hematology-Oncology,

Feinberg School of Medicine (JAP), Northwestern University,

Chicago, IL; Department of Emergency Medicine, University of

North Carolina (JB), Chapel Hill, NC; and Department of Emer-

gency Medicine (AHA, JE) and Pain and Emergency Medicine

Institute (KHT), Beth Israel Medical Center, New York, NY.

Received September 22, 2006; revision received November 20,

2006; accepted November 21, 2006.

Presented at the annual meeting of the National Sickle Cell Dis-

ease Program, Clinical Care, Research and Education, Memphis,

ª 2007 by the Society for Academic Emergency Medicine

doi: 10.1197/j.aem.2006.11.033

treatment decisions caused by the belief held by manynurses and physicians that patients with SCD are opioiddependent, despite a lack of data supporting this belief.6–8

While data are limited, it is unknown if this perceptionof addiction held by some clinicians contributes to the

TN, April 2006; American College of Emergency Physicians

annual meeting, New Orleans, LA, October 2006; and Illinois

College of Emergency Physicians Academic Forum, Resident

Competition, Chicago, IL (awarded second place), September

2006.

Dr. Tanabe was partially supported by a grant from the Illinois

Department of Healthcare and Family Services to Northwestern

Memorial Hospital under the Excellence in Academic Medicine

Act.

Contact for correspondence and reprints: Paula Tanabe, PhD,

RN, MPH; e-mail: ptanabe2@nmff.org.

ISSN 1069-6563

PII ISSN 1069-6563583 419

420 Tanabe et al. � ED SICKLE CELL PAIN MANAGEMENT

inconsistent or inadequate treatment of acute pain epi-sodes for persons with SCD in the ED setting.

Frequent acute pain episodes have been associatedwith increased mortality, and therefore rapid evaluationis critical to ensure not only prompt relief of pain butto assess for life-threatening complications.9 Significantvariability in the management of acute pain episodes hasbeen reported; only 20% of emergency physicians re-port using a protocol when treating acute pain episodesassociated with SCD.10

The American Pain Society (APS) has published the‘‘Guideline for the Management of Acute and ChronicPain in Sickle-Cell Disease,’’ along with the ‘‘Guidelinefor the Management of Acute Pain in Sickle-Cell Disease,Quick Reference Guide for Emergency Department Clini-cians.’’11,12 The guideline recommends providing initialanalgesics within 15–20 minutes of arrival and use ofintravenous (IV) or subcutaneous (SQ, when IV accessis not available) morphine 0.1–0.15 mg/kg (or 5–10 mgif weight >50 kg) or hydromorphone 0.015–0.02 mg/kg.The current crisis of ED overcrowding presents a signif-icant challenge to providing timely pain relief; however,it is important to understand and examine pain manage-ment practices in the ED setting in the context of thecurrent ED overcrowding challenges.13

Despite the severity of acute pain episodes, the fre-quency of ED visits, and the frustration level of bothclinicians and patients, to date, no study has objectivelyevaluated the analgesic management practices for pa-tients who present to the ED with acute pain episodes.The aims of this study were to 1) characterize the initialmanagement of patients with SCD presenting with anacute pain episode in three ED settings, 2) compare thesepractices with the APS guideline, and 3) identify factorsassociated with a delay in receiving an initial analgesic.

METHODS

Study DesignA multicenter, retrospective study design was used. Theinstitutional review board at each setting approved theproject, and a waiver of written and verbal consent wasobtained at each site.

Study Setting and PopulationThree EDs from the East Coast and Midwest wereincluded. All sites were academic medical centers withemergency medicine residency programs. The majorityof faculty were board certified in emergency medicine.Sites reported a combined census of 140,000 individualED patient visits during the study period. Consecutivevisits for patients presenting to the ED with an acutepain episode associated with SCD during a one-yearstudy period were eligible for inclusion. Exclusion crite-ria were age younger than 18 years or having a chiefcomplaint unrelated to pain crisis.

Study ProtocolClinicians were unaware of the study. A structured med-ical review was conducted by specially trained data ab-stractors. Before individual site abstraction, a sample often blinded ED records from the coordinating centerwas sent to data abstractors from each site to establish

interrater reliability. The abstractors were required toabstract all data elements and enter the data into thestudy de-identified database. Sample case data entry wasreviewed by the study primary investigator, and dataentry discrepancies were discussed and resolved until100% agreement on all data elements was reached. Alldata elements were abstracted from the ED medicalrecord. One site used only electronic records, the secondsite used only paper records, and the third site switchedfrom paper to electronic records during the project;therefore, both electronic and paper records wereabstracted for this site.

A structured medical record review was used to ab-stract all data, including the following outcome variables:analgesic agent and dose, route, and time of adminis-tration of the initial analgesic. All pain scores were ab-stracted as documented in the medical record using ascale from 0 to 10. Time of administration was calculatedby subtracting the time the first analgesic agent wasprovided from the time of arrival in the ED.

Additional variables abstracted included time to admin-istration of initial analgesic, demographics (gender, age,and race), triage level, presence or absence of IV accessat any point during the ED visit, and study site. Triagepriority levels were recorded as 1–5 using the EmergencySeverity Index (ESI) triage system. Triage level 1 is thehighest priority and reserved for immediate life-threaten-ing situations. Level 5 indicates the lowest priority. ESIrecommends a triage level of 2 for patients with pain rat-ings R7 for whom pain cannot be managed at triage bythe triage nurse. Therefore, patients with SCD with anacute pain episode and pain rating R7 should receive anESI triage level of 2.14 IV access was coded as present ifobtained at any time during the ED visit. A final variable,‘‘emergency department visit frequency,’’ was generatedfrom the data set. The total number of ED visits for each in-dividual patient during the study period was calculated.‘‘Emergency department visit frequency’’ was then cate-gorized using the following three visit frequency groups(1–3, 4–12, or 13 or more), indicating the total number ofED visits for an individual subject. These intervals weredetermined post hoc based on the natural breaks in thedistribution of visits for the sample.

Analgesic management was compared with the prac-tices recommended in the APS guideline.11,12 The mediantime to administration of initial analgesic was calculatedfor each visit and compared with the recommendedtime of 15–20 minutes from ED arrival. Analgesic agentswere categorized as correct (morphine sulfate or hy-dromorphone) or incorrect (all other agents). Analgesicroutes were categorized as correct (IV or SQ) or incor-rect (oral or intramuscular). The initial analgesic dose,agent, and route were converted to standard intravenousmorphine sulfate equivalents (MSE) for analysis. The cor-rect dose was determined by comparing the initial dosein MSE with the recommended APS guideline–suggesteddose of morphine sulfate 5 mg IV. All doses were catego-rized as correct (R5 mg IV MSE) or incorrect (<5 mgIV MSE).

Data AnalysisData from each site were directly entered into an Accessdatabase (Microsoft Corp., Redmond, WA). Descriptive

ACAD EMERG MED � May 2007, Vol. 14, No. 5 � www.aemj.org 421

Table 1ED Patient Characteristics per Study Site

Site 1 Site 2 Site 3 Total Sample

No. of unique patients 62* 62 35 159Age, mean � SD (yr) 35 � 9 32 � 12 30 � 10 32 � 10Female, n (%) 32 (52) 33 (53) 22 (63) 87 (55)Race and ethnicity, n (%)

African American 59 (97) 62 (100) 26 (90) 147 (97)yWhite 1 (1.6) 0 0 1 (0.7)Hispanic 1 (1.6) 0 3 (10) 4 (2.3)

Frequency of visit/studyperiod/individual patient, n (%)

1–3 43 (69) 44 (71) 29 (83) 117 (73.1)4–6 8 (13) 8 (13) 2 (6) 18 (11.2)7–10 2 (3) 5 (8) 0 7 (4.3)11–14 2 (3) 2 (3) 3 (9) 7 (4.3)15–17 5 (8) 1 (1.5) 0 6 (3.9)18–20 1 (1.6) 1 (1.5) 1 (2) 3 (2)R21 1 (1.6) 1 (1.5) 0 2 (1.2)Total no. of patients 62 62 35 159

Data presented represent individual patients and are organized by study site.

* Subjects at site 1 were significantly older (p = 0.048).yTotal does not equal 159 subjects due to missing racial data.

statistics were used to report time to administration ofinitial analgesic and to describe agents, doses, and routesof administration. Due to the positive skew of the data,univariate analyses were performed using the Mann–Whitney U test or the Kruskal–Wallis test to identifydifferences in time to administration of initial analgesicbetween the following groups: gender, triage level, IV ac-cess, study site, and ED visit frequency. Due to the smallnumber of patients who were assigned a triage score of1, 4, or 5, the triage score was dichotomized and analyzedas 1 or 2 versus 3, 4, or 5. Means and 95% confidence inter-vals (CIs) of the differences between groups are reported.Spearman’s rank order correlation coefficient was calcu-lated to assess the relationship between age and time toadministration of initial analgesic. Multiple linear regres-sion was used to assess the relationship between all varia-bles and time to initial analgesia.

RESULTS

There were 612 patient visits for 159 unique patients(45% male; mean age, 33 years; range, 18–74 years). Themedian number of visits per individual patient during thestudy period was two (mean, 3.85; range, 1–31). Individ-ual patient characteristics are presented by site in Table 1.Approximately 73% of all patients had a total of one tothree ED visits during the study period (Table 1). Data de-scribing ED visit characteristics are presented in Table 2.During 15 visits, analgesics were not administered; dur-ing two of these visits, patients left before evaluation.Complete data documenting arrival and time of initial an-algesic were available for 529 patient visits.

Adherence to GuidelinesMedian time to administration of initial analgesic was 90minutes (25th to 75th interquartile range [IQR], 54–149minutes) and is presented in Figure 1 and Table 2. Timefrom arrival to room placement averaged 59 minutes(median, 30; IQR, 15–71) and time from room placement

to analgesic averaged 60 minutes (median, 45; IQR,30–75). During 87% of visits, patients received eithermorphine or hydromorphone as the initial analgesicagent; 92% received the recommended dose, and 55%received the drug by either the IV or SQ route. Initial an-algesic administration agents and routes are presented inTable 2 by study site and represent analgesic manage-ment by total ED visits during the study period.

Factors Associated with Longer Timeto Analgesic AdministrationUnivariate techniques revealed that longer times to ad-ministration in minutes (mean [�SD]) of the initial analge-sic agent occurred in female patients (125 [�89]) whencompared with male patients (103 [�80]; mean difference,21 minutes; 95% CI = 7 to 36 minutes; p = 0.003) and inpatients assigned triage level 3, 4, or 5 (127 [�91]) whencompared with patients assigned triage level 1 or 2 (82[�56]; mean difference, 45 minutes; 95% CI = 29 to 61minutes; p = 0.00; Figure 2). Pain ratings were similarfor patients triaged as level 1 or 2 (mean [�SD], 9.21[�1.05]) when compared with patients triaged as level 3,4, or 5 (mean [�SD], 8.66 [�1.42]). Those patients withoutIV access were more likely to have longer times toadministration (mean [�SD], 132 [�89] minutes) whencompared with patients with IV access (mean [�SD], 107[�83] minutes; mean difference, 24 minutes; 95% CI = 8to 40 minutes; p = 0.002). Study site 2 had longer times(mean [�SD], 135 [�91] minutes) when compared withsite 1 (mean [�SD], 109 [�84] minutes; mean difference,25 minutes; 95% CI = 5 to 46 minutes; p = 0.007), andwhen compared with site 3 (mean [�SD], 98 [�73] min-utes; mean difference, 36 minutes; 95% CI = 10 to 62 min-utes; p = 0.002). No statistically significant difference toanalgesic time between frequency of visit categorieswas found. There was no correlation between age andtime to analgesic administration. Multivariate regressionresults are presented in Table 3. Study sites 1 and 2,female patients, and patients triaged as level 3, 4, or 5

422 Tanabe et al. � ED SICKLE CELL PAIN MANAGEMENT

Table 2ED Visit Characteristics

Site 1 Site 2 Site 3 Total Sample

Median (interquartile range) time in minutesto administration of initial analgesic

80 (48–144) 107 (73–187) 77 (50–116) 90 (54–149)

Triage level, n (%)* 263 230 110 603y1 1 (0.4) 0 0 1 (0.2)2 100 (38) 58 (25) 4 (3.6) 162 (27)3 161 (61) 168 (73) 100 (91) 429 (71)4 1 (0.4) 2 (0.9) 6 (5.5) 9 (1.5)5 0 2 (9) 0 2 (.3)

Clinical characteristicsArrival pain score, mean � SD 8.6 � 1.3 8.9 � 1.3 9 � 1.5 8.8 � 1.4IV access obtained, n (%) 158 (58) 158 (69) 97 (90) 413 (68)

Initial analgesic agents, n (%)Morphine sulfate 24 (9.1) 75 (33) 32 (31) 131 (22)Hydromorphone 222 (84.1) 125 (55) 39 (38) 386 (65)Fentanyl 1 (0.4) 1 (0.4) 0 3 (0.3)Meperidine 0 3 (1.3) 28 (27) 31 (5)Ketorolac 3 (1.1) 2 (0.9) 0 5 (0.8)Ibuprofen 1 (0.4) 1 (0.4) 0 2 (0.3)Acetaminophen 4 (1.5) 0 1 (1) 5 (0.8)Codeine/acetaminophen 2 (0.8) 2 (0.9) 0 4 (0.7)Hydrocodone/acetaminophen 7 (2.7) 0 0 7 (1.2)Oxycodone/acetaminophen 0 19 (8) 4 (4) 4 (3.8)

Initial analgesic route, n (%)Intravenous 125 (47) 120 (53) 75 (71) 320 (54)Oral 42 (16) 22 (10) 5 (4.8) 69 (12)Intramuscular 97 (37) 78 (34) 24 (23) 199 (33)Subcutaneous 0 8 (3.5) 1 (1) 9 (1.5)

Data presented represent individual ED visits and are organized by study site.

* Triage levels were assigned on arrival by the triage nurse or nurse first evaluating the patient at the bedside. Triage level 1 is the highest priority and

reserved for immediate life-threatening situations. Patients with sickle cell disease and an acute pain episode should be triaged as level 2, the second

highest priority. Level 5 indicates the lowest priority.yn does not equal 612 due to missing data for triage categories.

remained significant predictors of longer times to provi-sion of initial analgesic in this model. The most significantpredictor remained triage level; patients assigned level 3,4, or 5 waited an average of 48 minutes longer than pa-

Figure 1. Histogram of time to initial analgesic administra-

tion. The histogram represents the number of subjects and

time from arrival at triage to receiving the initial analgesic.

tients assigned level 1 or 2. Finally, presence of IV accesswas no longer significant in the multivariate model andage remained an insignificant predictor.

DISCUSSION

To the best of our knowledge, this is the first study to re-port initial analgesic management practices for patientspresenting to an ED with an acute pain episode due toSCD. Data demonstrate lengthy delays in providing ini-tial analgesia, on average 70–75 minutes longer than rec-ommended by the APS guideline. Patients also receivedinitial analgesics by routes not recommended (oral andintramuscular) by the APS guideline. Actual analgesicagents and doses usually followed the APS guidelinerecommendations.

Emergency departments are challenged by a crisis inovercrowding. The recent Institute of Medicine reportclearly describes this crisis.13 New York and Californiahave also reported their experiences with ED overcrowd-ing.15,16 Although we were unable to directly measureovercrowding in our specific study sites, we acknowl-edge that overcrowding most likely contributed tosome of the delay that was experienced by patients.EDs are challenged to provide rapid care for patientswith chest pain, patients with stroke, multiple trauma vic-tims, and many other patients with serious physiologiccompromise. Where does the patient with SCD and an

ACAD EMERG MED � May 2007, Vol. 14, No. 5 � www.aemj.org 423

acute pain episode fit into this picture? While we ac-knowledge that the APS recommendation of 15–20 min-utes may be very difficult to achieve in the ED setting, adelay of 90 minutes leaves much room for improvement.Some patients in the study experienced delays in the ad-ministration of the initial analgesic of up to ten hours.This is not reasonable. Although no study to our knowl-edge has examined the time to administration of an initialanalgesic for patients with renal colic, we propose adopt-ing a similar model of rapid, aggressive analgesic man-agement, for both pain syndromes. Patients with both painsyndromes present with acute pain, have a documentedphysiologic cause, often present with accompanyingemotional distress, and may suffer from recurrent epi-sodes of acute pain. Data from this project suggest thereis room for improving time to administration of an initial

Figure 2. Triage level differences in time to initial analgesic

administration. Boxes represent median pain score and

interquartile range. Outliers and extreme outliers are indi-

cated by circles and asterisks, respectively. Patients triaged

as Emergency Severity Index (ESI) level 3, 4, or 5 experienced

significantly longer delays to initial analgesic when com-

pared with patients triaged as ESI level 1 or 2.

analgesic for patients with pain episodes associated withSCD, even in the context of severe overcrowding. EDsare encouraged to examine departmental-specific dataand triage policies and consider implementation ofstanding orders that would allow nurses the ability to ad-minister analgesic agents before physician evaluation.This could be accomplished by developing individual pa-tient care plans for patients known to an individual ED.

There are many reasons to encourage rapid analgesicmanagement. Severe pain from an acute pain episodeshould be considered a medical emergency. Acute painepisodes are often associated with other serious life-threatening physiologic complications of SCD, includingsepsis, acute chest syndrome, aplastic anemia, stroke,pulmonary embolus, and overt organ failure.9 By delay-ing treatment, physicians may delay their ability to diag-nose and treat these potentially catastrophic events. Ina classic study of adults with SCD, 22% of deaths wereassociated with an acute pain episode.9 Patients withfrequent ED visits deserve careful attention because anincreased frequency of hospitalization has been associ-ated with an increased risk of death.1 Finally, untreatedpain triggers additional negative physiologic responses.Sensory input from injured tissues reaches spinal cordneurons and may cause subsequent responses to beenhanced.17–19 This physiologic response occurs with de-lays in initial analgesia, making it more difficult to controland may exacerbate the pain episode. It is hypothesizedthat increased severity of pain is more difficult to controland has been associated with longer hospitalizations,immediate recurrence of pain episodes, and increasedfrequency of hospitalizations.

We identified several factors that contributed to delaysin receiving initial analgesics. Patients assigned a lowertriage priority (level 3, 4, or 5) waited an average of45 minutes longer before receiving their first analgesicwhen compared with patients assigned a higher triagepriority (level 1 or 2), despite pain scores being equal.The ESI implementation handbook states that all patientswho have a pain rating of R7 on the ten-point scaleshould be considered to meet ESI level 2 criteria.14

However, it is up to the discretion of the triage nurse todetermine whether the patients’ chief complaint, medical

Table 3Multivariate Linear Regression Analysis of Time to Administration of Initial Analgesic

Parameter b Coefficient Standard Error t-statistic p-value95% Confidence

Interval

Intercept 132.59 17.06 7.77 0.00 99.0, 166.1Site 1 vs. Site 3 (referent) 30.35 10.72 2.83 0.005 9.3, 51.4Site 2 vs. Site 3 (referent) 47.43 10.58 4.48 0.00 26.6, 68.2Frequency of visits (no. per study period) �0.25 0.45 �0.56 0.57 �1.1, 0.62Age (yr) �0.32 0.39 �0.80 0.42 �1.1, 0.46Male �22.05 7.26 �3.03 0.003 �36.3, �7.7ESI level 1 or 2 (referent) vs. ESI level 3, 4, or 5 �48.09 8.49 �5.66 0.000 �64.7, �31.4Intravenous access �13.07 8.38 �1.5 0.119 �29.5, 3.4

Time to initial analgesic is reported in minutes. Patients from sites 1 and 2 experienced longer times to initial analgesic by 30 and 47 minutes, respec-

tively, when compared with patients at site 3. Site 3 was used as the referent category as it had the lowest times to initial analgesic administration,

although the time still exceeded recommendations. Female patients experienced an average of 22 minutes longer to initial analgesic, and patients triaged

as lower priority (ESI level 3, 4, or 5) waited an average of 48 minutes longer when compared to patients triaged as ESI level 1 or 2.

ESI = Emergency Severity Index.

424 Tanabe et al. � ED SICKLE CELL PAIN MANAGEMENT

history, and pain rating meet ESI level 2 criteria. Patientswith an acute sickle cell pain crisis are clearly identifiedas an example of an ESI level 2 patient who should be tri-aged as level 2.14 In our three centers that use ESI as thetriage system, only 27% of patients were assigned ESIlevel 2. The triage nurse plays a critical role in determin-ing how quickly patients with an acute pain episode willbe placed in a room and evaluated by a physician. Thishas a direct effect on time to receiving analgesia. If a tri-age nurse assigns a lower triage priority level, patientswill wait significantly longer before receiving pain reliefin the current atmosphere of overcrowding. It is criticalthat all emergency clinicians recognize the potential seri-ous physiologic complications of inadequate pain reliefand the importance of providing rapid pain control forthis population. In the current atmosphere of ED over-crowding, it is not uncommon for the triage nurse tobe in a situation of triaging many ESI level 2 patientswith no open ED beds in which to place the patients.In these situations, the triage nurse may be tempted to‘‘undertriage’’ the patient with an acute pain episodefrom SCD. However, this can lead to serious, negativepatient outcomes and make it more difficult to controlthe pain episode. This in turn may lead to an increasedrate of hospitalization and a longer length of stay in theED. Additional studies are warranted to evaluate theseoutcomes.

In the univariate analysis, we also found that the lack ofIV access contributed significantly to delays in adminis-tration of an initial analgesic. This delay did not remainsignificant in the multivariate model; however, it is worthdiscussing. The delay may be secondary to multiple un-successful attempts because IV access is often difficultfor persons with SCD due to complications associatedwith the disease and many previous venous accessattempts. The delay associated with obtaining IV accesscould be avoided by instituting protocols that recom-mend immediate SQ administration of an opioid if IV ac-cess is unsuccessful with the initial attempt. However,patients and clinicians may be unfamiliar with the SQroute and education is warranted. Furthermore, patientsoften request intramuscular delivery and may expressresistance to the SQ route. Historically, patients withSCD received intramuscular meperidine for treatment ofacute pain episodes.20 Yet, the intramuscular route is as-sociated with unreliable absorption and the potential tocause muscle and soft tissue damage and is not recom-mended.21 This resistance to adapt to an improved routeof therapy may originate from a level of mistrust that ex-ists between many patients with SCD and clinicians.5

We also identified a gender disparity in time to initialanalgesia. Female patients in this sample waited longerthan their male counterparts for analgesic administra-tion, despite a lack of statistically significant differencesin initial pain scores. This was an unanticipated finding,and the meaning is unclear. Although numerous studieshave demonstrated behavioral differences in how menand women experience pain as well as physiologic differ-ences in how they respond to opioid analgesia, recentresearch has shown that men and women generallyreport similar pain scores.22–26 Men with SS genotypereported more days with an acute pain episode and asubsequent higher utilization of health care services

associated with these episodes when compared withwomen.26 It is unclear if the patients in our studyactually presented to the ED with differing pain exp-eriences or if the delay in administration of opioidsmay be attributed to gender bias. Gender disparities havebeen demonstrated in other areas of medicine. Specifi-cally, gender differences have been known to existin the recognition and treatment of coronary arterydisease.27,28 Future research on the treatment of SCDin the ED and other settings should further explore thisissue.

It was noted that repeat visits by an individual patientdid not result in longer wait times for initial analgesic ad-ministration when compared with patients who did nothave multiple ED visits during the study period. Patientswith frequent ED visits may be at risk for being perceivedas ‘‘drug seekers,’’ and it is unknown if emergency clini-cians may subconsciously delay administration of theinitial analgesic. This did not occur in our sample.

There were several strengths to our study. Three sitesrepresenting geographically distinct sections of theUnited States were included. Additionally, visits were in-cluded during an entire one-year period, representingcomplete data for more than 500 visits.

LIMITATIONS

Limitations to the study included use of a structuredretrospective chart review. In general, recording ofpharmacologic agents and doses in the medical recordis usually reliable, whereas documentation of the timethat care was provided is often missing. Specific to thisstudy, we found excellent documentation of analgesicagents, routes, and doses; however, documentation ofarrival times and/or analgesic administration times wasmissing for 83 visits. These were not included in the anal-ysis. It is possible, although not likely, that analgesicagents were provided more rapidly for those patientswith missing time data. It was not possible to addressthe issue of patient preference of route of administration,and it is possible that some subjects preferred and re-quested the oral or intramuscular route. Similarly, it ispossible that IV access was not obtained because emer-gency physicians preferred the oral or intramuscularroute for analgesic administration, not because IV ac-cess was difficult or not possible. Additionally, patientweights were not available to the abstractors. It is possi-ble that patients weighing <70 kg actually received a doseconsistent with the APS guideline; however, because thiswas an adult sample, it is unlikely that this was a signifi-cant limitation. Finally, we were not able to directlymeasure the contribution of overcrowding to the delay inthe administration of the initial analgesic agent.

CONCLUSIONS

Patients with an acute painful episode related to SCD ex-perienced significant delays when seeking relief in theED. Patients assigned a lower triage priority level and fe-male patients experienced the longest delays. Further re-search is required to identify strategies and/or protocolsthat will result in the delivery of rapid analgesia for pa-tients with an acute pain episode from SCD in the ED.

ACAD EMERG MED � May 2007, Vol. 14, No. 5 � www.aemj.org 425

The authors thank James Newmark for his assistance with dataabstraction.

References

1. Ballas SK, Lusardi M. Hospital readmission for adultacute sickle cell painful episodes: frequency, etiology,and prognostic significance. Am J Hematol. 2005; 79:17–25.

2. Platt OS, Thorington BD, Brambilla DJ, et al. Painin sickle cell disease. Rates and risk factors. N Engl JMed. 1991; 325:11–6.

3. Philpott S, Mason J, Aisiku IP. Patient satisfaction inthe emergency department management of acutesickle cell pain [abstract]. Acad Emerg Med. 2005;12(Suppl 1):158.

4. Todd KH, Green C, Bonham VL Jr, et al. Sickle celldisease related pain: crisis and conflict. J Pain. 2006;7:453–8.

5. Maxwell K, Streetly A, Bevan D. Experiences of hospi-tal care and treatment seeking for pain from sickle celldisease: qualitative study. BMJ. 1999; 318:1585–90.

6. Waldrop R. Health professional perceptions of opioiddependence among patients with pain. Am J EmergMed. 1995; 13:529–31.

7. Pack-Mabien A, Labbe E, Herbert D, Haynes J Jr.Nurses’ attitudes and practices in sickle cell painmanagement. Appl Nurs Res. 2001; 14:187–92.

8. Shapiro BS, Benjamin LJ, Payne R, Heidrich G. Sicklecell-related pain: perceptions of medical practi-tioners. J Pain Symptom Manage. 1997; 14:168–74.

9. Platt OS, Brambilla DJ, Rosse WF, et al. Mortality insickle cell disease. Life expectancy and risk factorsfor early death. N Engl J Med. 1994; 330:1639–44.

10. Silbergleit R, Jancis MO, McNamara RM. Manage-ment of sickle cell pain crisis in the emergencydepartment at teaching hospitals. J Emerg Med.1999; 17:625–30.

11. Benjamin LJ, Dampier CD, Jacox A, et al. Guidelinefor the Management of Acute and Chronic Pain inSickle-Cell Disease. Glenview, IL: American PainSociety, 1999.

12. Benjamin LJ, Dampier CD, Jacox A, et al. Guideline forthe Management of Acute Pain in Sickle-Cell Disease,Quick Reference Guide for Emergency DepartmentClinicians. Glenview, IL: American Pain Society, 2001.

13. Institute of Medicine. The Future of Emergency Carein the United States Health System. Washington, DC:National Academies Press, 2006.

14. Gilboy N, Tanabe P, Travers D, et al. EmergencySeverity Index, Version 4: Implementation Handbook.

Rockville, MD: Agency for Healthcare Quality andResearch, 2005.

15. Schneider S, Zwemer F, Doniger A, Dick R, Czapran-ski T, Davis E. Rochester, New York: a decade ofemergency department overcrowding. Acad EmergMed. 2001; 8:1044–50.

16. California HealthCare Foundation. Emergency De-partments in the Health Care System, Use of Servicesin California and the United States. 2002. Availableat: http://www.chcf.org/topics/hospitals/index.cfm?itemID=20203. Accessed Nov 23, 2006.

17. Hunt SP, Pini A, Evan G. Induction of c-fos-likeprotein in spinal cord neurons following sensorystimulation. Nature. 1987; 328:632–4.

18. Jacob E, Beyer JE, Miaskowski C, Savedra M, Tread-well M, Styles L. Are there phases to the vaso-occlu-sive painful episode in sickle cell disease? J PainSymptom Manage. 2005; 29:392–400.

19. Fitzgerald M. c-fos and the changing face of pain.Trends Neurosci. 1990; 13:439–40.

20. Perlman KM, Myers-Phariss S, Rhodes JC. A shiftfrom Demerol (meperidine) to Dilaudid (hydromor-phone) improves pain control and decreases admis-sions for patients in sickle cell crisis. J Emerg Nurs.2004; 30:439–46.

21. American Pain Society. Principles of Analgesic Use inthe Treatment of Acute Pain and Cancer Pain, 5th ed.Glenview, IL: American Pain Society, 2003.

22. Unruh AM, Ritchie J, Merskey H. Does gender affectappraisal of pain and pain coping strategies? Clin JPain. 1999; 15:31–40.

23. Sarton E, Olofsen E, Romberg R, et al. Sex differ-ences in morphine analgesia: an experimental studyin healthy volunteers. Anesthesiology. 2000; 93:1245–54.

24. Pleym H, Spigset O, Kharasch ED, Dale O. Genderdifferences in drug effects: implications for anesthesi-ologists. Acta Anaesthesiol Scand. 2003; 47:241–59.

25. Craft RM. Sex differences in opioid analgesia: ‘‘frommouse to man.’’ Clin J Pain. 2003; 19:175–86.

26. McClish DK, Levenson JL, Penberthy LT, et al.Gender differences in pain and healthcare utilizationfor adult sickle cell patients: The PiSCES Project.J Womens Health (Larchmt). 2006; 15:146–54.

27. Takakuwa KM, Shofer FS, Hollander JE. The influ-ence of race and gender on time to initial electrocar-diogram for patients with chest pain. Acad EmergMed. 2006; 13:867–72.

28. Diercks DB, Kirk JD, Lindsell CJ, et al. Door-to-ECGtime in patients with chest pain presenting to theED. Am J Emerg Med. 2006; 24:1–7.