Quality & Safety Issues: How they impact practice in the ICU
Hildy Schell, RN, MS, CCNSMichael Gropper, MD, PhDUCSF Medical Center
Definition: Medical Quality
Medical quality is the degree to which health care systems, services and supplies for individuals and populations increase the likelihood for positive health outcomes IOM, 1990
Quality care is safe, timely, effective, efficient, equitable, and patient-centered IOM, 2001
1999: IOM Report: To Err is Human
As many as 98,000 Americans die each year from medical errorsMore people die in a given year as a result of medical errors than from motor vehicle accidents (43,458), breast cancer (42,297), or AIDS (16,516)Costs are estimated to be between $17 billion and $29 billionRecommended the establishment of the Agency for Healthcare Research and Quality (AHRQ)
The “Swiss Cheese Model” of Major Accidents & Errors
No Code Blue ID Protocol, safety
systems
No teamwork or “fog of war” training
“Culture of Low Expectations”
Nurse does not express concerns
Patient Mistakenly
Not Resuscitated
James Reason, Human Error
May 2006
Time Magazine 5/1/06
An attending neurologist said one drug should be started immediately, that "time is of the essence." That was on a Thursday morning at 10 a.m. The first dose was given 60 hours later, on Saturday night at 10 p.m. "Nothing I could do, nothing I did, nothing I could think of made any difference," Berwick said in a speech to colleagues. "It nearly drove me mad." One medication was discontinued by a physician's order on the first day of admission and yet was brought by a nurse every single evening for 14 days straight. "No day passed--not one--without a medication error," Berwick remembers. "Most weren't serious, but they scared us." Drugs that failed to help during one hospital admission were presented as a fresh, hopeful idea the next time. If that could happen to a doctor's wife in a top hospital, he says, "I wonder more than ever what the average must be like. The errors were not rare. They were the norm."
Quality Improvement
Just trying harder is the worst quality improvement plan of all
Deming
Quality Goals
Improve patient care and outcomesUtilize evidence to drive and/or revise current practiceMaintain a reasonable workflow pattern to meet the patient/family needs while considering impact on care providersOptimal/appropriate use of resources
What triggers practice review/change?
New evidence– Research– Best practice
High mortality High morbiditySafety issuesResources– Time– Personnel– Cost (supplies, equipment, procedure)
Performance Improvement Process
IdentifyMeasureAnalyzeDevelopImplementMeasure
Team organizationResourcesData AcquisitionAnalystsToolsStrategies for changeMaintenance
Rapid-Cycle Change Process
Involves staff in identification and implementationSmall scale, clinically relevant changesRe-evaluation and ongoing improvementLess resistance to change
The PDSA Cycle for Learning and Improvement
Act• What changes
are to be made?
• Next cycle?
Plan• Objective• Questions and
predictions (why)• Plan to carry out the cycle
(who, what, where, when)• Plan for data collection
Study• Complete the
analysis of the data• Compare data to
predictions• Summarize
what waslearned
Do• Carry out the plan• Document problems
and unexpectedobservations
• Begin analysisof the data
Hot Topics
Prevention of hospital-acquired infections– VAP & HAP– CR-BSI– SSI
“Rescuing” patients -RRTeamGlycemic control (“tight glycemic control”) Early Goal Directed Therapy: Sepsis identification & resuscitation
Hot Topics
DVT prophylaxisStress ulcer prophylaxisPerioperative Beta-BlockadeAMI EB care (time to intervention)Prevention of adverse drug eventsPressure Ulcer PreventionFalls Prevention
“Bundle” phenomena
Groupings of best practices to improve careEvidenced-basedImprove quality when applied together
UCSF Adult Critical Care
VAP preventionGlycemic controlCR-BSI preventionRRTeamEGDT for sepsisHypothermia post-cardiac arrestICU discharge checklistPressure ulcer prevention
“Tight” Glycemic Control
Hyperglycemia is common in critically ill patients, and has been associated with:
1. Neuropathy2. Skeletal muscle wasting3. Increased growth hormone concentrations4. Increased susceptibility to infection5. Prolonged mechanical ventilation6. Hyperglycemia impairs neutrophil phagocytosis
Hypothesis: Glycemic control with insulin infusion will decrease the incidence of the above complications
Tight Glycemic Control
van den Berghe, et al (2001) NEJM– Randomized, prospective trial– N = 1548 SICU pts started on insulin infusions– Conventional (180-200) vs Tight Control (80-110)
van den Berghe, et al (2006) NEJM– Randomized, prospective trial– N = 1200 MICU pts (767 with LOS > 3 days)– Conventional (180-215) vs Tight Control (80-110)
Intensive Insulin Therapy in Critically Ill Patients (van den Berghe et al. N Eng J Med 2001)
Randomization
Conventional Intensive
>215 mg/dL
180 to 200 mg/dL
>110 mg/dL
80 to 110 mg/dL
Blood glucose level when insulin infusion
was started
Infusion adjusted to maintain blood
glucose
-60
-50
-40
-30
-20
-10
0In hospitalmortality
Blood streaminfections
ARF requiringdialysis
Red celltransfusions
Critical illnesspolyneuropathy
Benefits of intensive insulin therapy compared to conventional insulin therapy
34
4641
5044
Tight Glycemic Control in MICU patients (van den Berghe,2006)
Mortality– All patients (n = 1200), there was NO difference in
mortality
– TG group patients in ICU > 3 days (n = 767) had reduced mortality (43% vs 52.5%, p=0.009) & had significantly reduced morbidity.
IMADIM: Measure
A pilot test utilizing a revised insulin infusion protocol which included a rate of change methodology and target BG level of 80-120 mg/dL was done in 2 of the adult ICUs.Elements measured: BG level, Insulin dose,
Dextrose source, Patient demographics & partial Hx.
December 2002 – Jan 2003
IMADIM: Analyze
Pilot test: 12 patientsAverage BG level: 127 +/- 43 mg.dLMedian BG level: 119 mg/dLHypoglycemic(< 60 mg/dL) episodes: 1.3%
IMADIM: Design
PI team convenedInsulin infusion protocol revised based on clinician feedback, CNS observations & monitored data.Developed educational presentation.Developed competency tool.Developed monitoring tool.
IMADIM: Implement
Inservice education to nursing staff by CNSs.Case based competency test to all RN staff.Inservice education to CC Medicine physicians by CCM director & CNS.Inservice to primary teams by Endocrine MD.Monitoring implemented by CNSs.
IMADIM: Measure
33 Medical & surgical critically ill pts69% with DM & 31% without DM2920 hours (avg 88.5 hrs on protocol)2265 BG checks performed9.4 hrs avg time to target BG level136 mg/dL average BG level119 median BG level19 (0.08%) episodes of hypoglycemia that were treated & without clinical harm
Analysis: Blood Glucose Monitoring in Adult ICUs (IV insulin therapy)
Sep Oct Nov Dec Jan Feb
Total # BG checks 401 428 407 204 250 200
Average BG (mg/dL) 147 148 143 126 135 133
Standard Deviation + 60 + 56 + 80 + 28 + 45 + 52
Median BG (mg/dL) 136 135 124 123 128 120
# BGs < 60 mg/dL 2(0.5%) 3(0.7%) 2(0.4%) 0 2(0.8%) 0
# insulin dose calculation errors
97 (24%)
58 (14%)
75 (18%)
0 3 (1.2%)
0
# missed BG checks 58 (14%)
39 (9%) 53 (13%)
20 (10%)
2 (0.8%)
2 (1%)
IMADIM: Evaluation
Increased nursing resources (time to obtain sample, perform test & intervene: 2 pts = 2 hrs/12 hr shift)Identified need for more monitoring supplies & equipmentPatient safety: invasive access for samplingPatient comfort, sleep, & satisfactionIdentified need for guidelines for use to include indications, review prompts & discontinuation/transition planOngoing monitoring (10 pts/month for 48 hrs of therapy). Identified needs for inservicing & compliance feedback
ICU Glycemic Control Report
2000– Insulin Infusion: mean BG 206 mg/dL– Insulin SQ: mean BG 264 mg/dL
2004– Insulin Infusion: mean BG 146 mg/dL– Insulin SQ: mean BG 186 mg/dL
2006– Insulin Infusion: mean BG 136 mg/dL– Insulin SQ: mean BG 150 mg/dL
UCSF Medical Center 12.2006
January 2008
What should we do now?
Glycemic Control
Mar-Apr 2008 all ICU patients x 5 daysPatient day assessments 247Total # BG checks 1604Average BG level 141 mg/dL
0
20
40
60
80
100
120
140
160
180
Total 9ICU 10ICC 11ICU
All AvgIV AvgSQ Avg
Blo
od G
luco
se L
evel
2008 UCSF Adult Critical Care (Units)
Glycemic Control (insulin & no insulin)
0
50
100
150
200
250
300
2000 2004 2006 2008
IV InfusionSQ
Blo
od G
luco
se L
evel
UCSF Adult Critical Care (2008: 5 day prevalence)
2008 Findings
Use of SQ insulin protocol increased in all units from 33% to 49% patients on insulin.Use of IV insulin protocol decreased from 19% to 11% (9ICU decreased from 29% to 9%)
Hypoglycemia (BG < 60) did not occur in any patients on insulin therapy
2008 Findings
One pt with no BG monitoring (improved)
Patients with NO insulin orders53/100 (53%) had at least 1 BG > 120 mg/dL21/100 (21%) had at least 1 BG > 150 mg/dL– Medicine, Cards/HF, NV, NS, Neuro, LTU, KTU
Recommendations
Identify threshold for review and change in insulin dose when patients on therapy
– IV Insulin protocol orders (call if BG > 400)
Pt on NO insulin & BG > 150 x 2, notify team to consider insulin coveragePt on IV or SQ insulin & BG > 150 x 2, notify team to consider increase in dose/additional dose IVPt on IV insulin and BG > 250, notify to consider additional doseShare data with primary and CCM services
Protocol-based Care
Evidence-basedSafety-basedGuides decision makingEnhances consistency Reduces omissions & potential errorsImpacts patient outcomesPotential iterations
Protocol-based computer reminders, the quality of care and the non-perfectibility of man McDonald, CJ, NEJM 1976
Purpose: To determine whether clinical errors could be reduced by prospective computer suggestions about the management of simple clinical eventsMethods: Evaluated responses of 9 MDs to computer suggestions in protocols and usual care without protocolsResults: 51% response to protocol and 22% response to no protocolProspective reminders do reduce errors“Many of these errors are probably due to man’s limitations as a data processor rather than to correctable human deficiencies”
ICU Protocols
ARDS VentilationSedation Wake UpSedation ManagementTight Glycemic ControlEGDT for SepsisWeaning & extubation
AMI CareVAP PreventionCR-BSI PreventionDVT prophylaxisElectrolyte ReplacementPost-cardiac arrest hypothermia
Protocol Cautions
Check box phenomenon– Familiarity/knowledge– Critical thinking
Inappropriate generalization– Use beyond indicated population
Managing multiple protocols– Conflicting interventions– Confounding assessments– Omission of “other” work
Keys for successful change
Use data to define & solve problemFind champions to lead and implement– Need a “Process Owner”– Help maintain the long-term integrity of the
effort/change processAction oriented Start small & then spreadEducate/orient all team membersMatch change measures to improvement goalProcess feedback from clinicians
PATIENT
Nurse
Physician Respiratory Therapist
Nutritionist
PharmacistRehab Therapist
The System
Family
The Multidisciplinary TEAM
Questions?
UCSF San Francisco, CA