Appendix to:
The association between hospital capacity strain and patient outcomes: A systematic review
Table of contents
Page- Key questions
- Inclusion and exclusion criteria
- Search strategy
- Characteristics of included studies
- Measures and thresholds to describe strain (in studies of the association between strain and mortality)
- Quality assessment of observational studies
- Quality assessment of experimental studies
1. Key questions
Key questions
- What are the effects of inpatient capacity strain on intermediate outcomes and health outcomes?
- Which health systems interventions improve intermediate outcomes and health outcomes for children receiving inpatient care during times of capacity strain?
2. Inclusion and exclusion criteria
Population / Patients receiving inpatient hospital care in highly developed countries during times of hospital capacity strain (may also be called “surge”, “overcrowding”, “hospital crowding”). May also be defined by specific effects of inpatient capacity strain, like Emergency Department boarding (sometimes called “access block”).Intervention /
- Health system interventions, implemented before or during capacity strain, for the purpose of increasing the likelihood of optimal clinical quality of care during times of capacity strain (examples include increasing staff, instituting early discharge programs)
- We will also include observational studies
Comparator / Experimental studies: No intervention
Outcome /
- Health outcomes (e.g., morbidity and mortality)
- Intermediate outcomes (e.g., proportion of patients treated according to clinical guidelines, number of medication errors)
Time period / 1999-2015
Setting /
- Inpatient medical or surgical (not psychiatric) units at acute care hospitals in countries with “very high human development” according to United Nations Human Development Index.
- Emergency departments, if study specifically examines effects of inpatient capacity strain (including ED boarding or access block) on patients who will be admitted to the hospital
Other criteria /
- Admissible designs: randomized controlled trials, including cluster randomized controlled trials; non-randomized trials, including interrupted time series studies with comparison groups; prospective and retrospective cohort studies; case-control studies.
- Non-admissible designs: descriptive studies lacking comparison groups, including case reports
- Language: English
- Exclude: Studies focusing on staffing, unless they specifically examine patient-driven (rather than staffing-driven) capacity strain
Countries included
(“Very high human development” in 2014 per United National Human Development Index, available at Andorra, Argentina, Australia, Austria, Bahrain, Belgium, Brunei Darussalam, Canada, Chile, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hong Kong, China (SAR), Hungary, Iceland, Ireland, Israel, Italy, Japan, Korea (Rep. of), Kuwait, Latvia, Liechtenstein, Lithuania, Luxembourg, Malta, Montenegro, Netherlands, New Zealand, Norway, Poland, Portugal, Qatar, Saudi Arabia, Singapore, Slovakia, Slovenia, Spain, Sweden, Switzerland, United Arab Emirates, United Kingdom, United States
Examples of characteristics of excluded studies
Ineligible publication types: Reviews, policy statements, editorials
Ineligible study designs: Ecological studies, descriptive studies without comparison groups
Ineligible settings: Entirely Emergency Department-based studies
No eligible exposures: No exposure measures reflecting inpatient capacity strain
No eligible outcome: No health outcomes, no outcomes for hospitalized patients
3. Search strategy
Medline search
# / Searches1 / bed occupancy/
2 / ((surge or surges or surging or strain$ or exceed$ or overextend$) adj3 capacit$).mp.
3 / high census.mp.
4 / 1 or 2 or 3
5 / (crowd$ or overcrowd$).mp.
6 / exp Hospitals/ or exp Hospital Administration/ or exp Hospitalization/ or health systems.mp.
7 / 5 and 6
8 / ((emergency or hospital$ or inpatient$) adj7 ((admission$ or admit$) adj3 delay$)).mp.
9 / ((emergency or hospital$ or inpatient$) adj5 (boarding or boarder$)).mp.
10* / access block.mp.
11 / 4 or 7 or 8 or 9 or 10
12 / limit 11 to yr="1999 -Current"
*Access block is defined as “the situation where patients who have been admitted and need a hospital bed are delayed from leaving the Emergency Department (ED) because of lack of inpatient bed capacity”, often operationalized as an ED length of stay greater than 8 hours. (Australasian College for Emergency Medicine Statement on Access Block (2014). October 27, 2016.
CINAHL
# / QueryS1 / (MH "Bed Occupancy")
S2 / (surge or surges or surging or strain or exceed or overextend) n3 capacity
S3 / (MH "Hospitals+")
S4 / (MH "Hospitalization+")
S5 / (MH "Health Facility Administration")
S6 / health systems
S7 / High census
S8 / S3 OR S4 OR S5 OR S6
S9 / crowding or overcrowding
S10 / S8 AND S9
S11 / S1 OR S2 OR S7 OR S10
S12 / S11 Limiters – Published Date: 19990101-, Exclude MEDLINE records
Cochrane Library and clinicaltrials.gov searches utilized similar terms and logic
Page A1
4. Characteristics of included studies
Author, year of publication, countryStudy year(s) / Number and type of patient
Setting/hospital characteristics
Primary diagnosis (if specified) / Metric to define capacity strain
Timing of capacity strain / Key Results
Outcome: Mortality
Ball, 2006, USA
Study year(s): 2003-2004 / 861 adults
1 trauma center
Trauma victims, injury severity score >12 / Metric: Mass casualty incident: 3 or more trauma patients admitted within 3 hours.
Timing: Time of presentation to hospital / Patients treated during mass casualty incidents did not have higher hospital mortality than trauma patients admitted during other times.
Bekmezian, 2012, USA
Study year(s): 2007-2008 / 1,792 children
1 tertiary care hospital / Metric:ED boarding time (time from admission decision to departure for an inpatient bed)
Timing:Time of admission request / No association between ED boarding time and hospital mortality.
Chalfin,2007, USA
Study year(s): 2000-2003 / 50, 322 adults requiring ICU admission
120 ICUs in 90 hospitals / Metric:≥6 hours spent in the ED after decision to admit to ICU.
Timing:N/A / 29% lower odds of hospital survival among patients with delayed admission to ICU.
Clark, 2007, USA
Study year(s): 2001-2003 / 1,200 adults requiring ICU admission
1 tertiary care hospital / Metric: Wait time from admission order until leaving ED
Timing: At time of admission order / 1.5% increase in hospital mortality for every 10% increase in waiting time between decision to admit to ICU and ICU admission.
Clark, 2012, USA
Study year(s): 2004-2005 / 1,433 adults requiring ICU admission
1 tertiary care hospital / Metric:ICU occupancy
Timing:At time of decision to admit to ICU / 9% increase in hospital mortality for every unit increase in ICU occupancy.
Derose, 2014, USA
Study year(s): 2008-2010 / 136,740 adults
13 hospitals / Metric: ED boarding time, both for index patient and non-index patients
Timing: ED registration / No association between increased ED boarding time and inpatient mortality.
Evans, 2006, USA
Study year(s): 1996-2000 / 9.9 million discharges (outcomes analyzed for 28,561 to >8M)
~400 hospitals / Metric: Admissions during the two days after a patient's admission (all index admissions on Thursdays). 1. Indexed to hospital's 8-week moving average. 2. Indexed to predicted admissions based on regression model.
Timing: Next 2 days after a Thursday admission / No association between high number of admissions on Friday and Saturday following a Thursday admission and hospital mortality.
Gabler, 2013, USA
Study year(s): 2001-2008 / 264,401 adults requiring ICU admission
155 ICUs in 107 hospitals / Metric: 1. ICU census, indexed to hospital’s mean daily census and standardized deviation. 2. ICU acuity, measured using mortality prediction model MPM0-III. 3. Number of ICU admissions.
Timing: 1. Day of ICU admission. 2. Average over first 3 days of ICU admission. / 2% increase in odds of in-hospital death for each standardized unit increase in ICU census. 2% decrease in odds of in-hospital death for each 10% increase in number of admissions. No effect of ICU acuity on mortality. Similar results for ICU mortality. Similar results for ICU capacity strain averaged over first 3 days of patients' ICU stays.
Gilligan, 2008, Ireland
Study year(s): 2004-2005 / 13,357 patients (3.6% < 18 years)
1 tertiary care hospital / Metric: ED boarding: Number of admitted patients boarding in ED at 9am.
Timing: Day of ED presentation / No association between number of ED boarders at 9am and hospital mortality.
Iapichino, 2004, Europe
Study year(s): 1994-1995 / 12,615 adults requiring ICU admission
89 ICUs in 12 European countries / Metric: Occupancy rate > 80%
Timing: Averaged over entire study period / Admission to an ICU with average occupancy >80% was associated with 32-35% increase in odds of hospital mortality compared to ICUs with average occupancy ≤80%.
Intas, 2012, Greece
Study year(s): 2009 / 200 adults requiring ICU admission after intubation in ED
1 hospital / Metric: ED boarding:>6 hours in ED after decision to admit to ICU
Timing:Time of admission / Patients with ED boarding times >6 hours had 5.7 times higher hospital mortality than those with shorter boarding time.
Iwashyna, 2009, USA
Study year(s): 2002-2005 / 200,499 adults requiring ICU admission
48 hospitals / Metric:ICU census, indexed to hospital’s mean daily census, divided into deciles
Timing:Day of ICU admission / No difference in mortality with increasing census on the day of admission.
Jenkins, 2015, USA
Study year(s): 2010-2011 / 230,621 adults
156 hospitals with trauma centers
Trauma victims / Metric: 1. Trauma Surge Index (TSI): Aggregate Injury Severity Score of patients admitted within 24 hours (before or after) of index patient, indexed to hospital's median and IQR; TSI >3 cutoff based on univariate analyses. 2. Mass casualty event (MCE): 10 trauma admits during 24-hour period, or 3 patients with ISS >15 over 3-hour period.
Timing: 24 hours before and after admission / 2-fold increase in hospital mortality for patients admitted during times of high trauma surge (TSI >3). No significant association between mass casualty event and mortality.
Madsen, 2014, Denmark
Study year(s): 1995-2012 / 2,651,021 adults
72 hospitals / Metric:Median bed occupancy
Timing: During first 24, 48, and 72 hours after admission / 1.2% increase in relative risk of inpatient and 30-day mortality per 10% increase in median bed occupancy rate. Occupancy ≥110% associated with 9% increase in mortality compared to occupancy rate <80%.
Marcin, 2004, USA
Study year(s): 1995-1999 / 102,008 adults
39 hospitals
Trauma victims / Metric: Number of trauma admissions, indexed to baseline admissions.
Timing:Annual, quarterly, and monthly / No association between high annual, quarterly, or monthly number of trauma admissions and hospital mortality.
O'Callaghan, 2012, UK
Study year(s): 2003-2007 / 1,609 adults requiring ICU admission
1 hospital / Metric:ED boarding:>3 hours from admission referral to admission
Timing:time of referral to ICU / Delay in admission to the ICU from the ED was not associated with ICU mortality.
Pascual, 2014, USA
Study year(s): 2005-2010 / 8,626 adults requiring surgical ICU admission
1 tertiary care hospital with multiple ICUs / Metric:ICU boarding: Surgical ICU patients boarded in an alternate overflow ICU when surgical ICU at capacity; boarding ICU distance from home ICU / No difference in ICU mortality between patients boarding in an overflow ICU and non-boarded patients.
Plunkett, 2011, Ireland
Study year(s): 2002-2008 / 23 ,114 adults
1 secondary care hospital / Metric:ED boarding: Wait time from admission team referral to bed placement
Timing: Time of presentation to ED / Compared to patients who waited <1 hour between admission team referral and bed placement, each category of increasing delay (1-2.5h, 2.5-6h, 6-14h, >14h) was associated with an additional 7% increase in 30-day hospital mortality.
Robert, 2012, France
Study year(s): unclear / 1,332 adults referred for ICU admission
10 hospitals / Metric: Refusal of ICU admission because of full ICU
Timing: Time of referral to ICU / 1.8-fold increased adjusted odds of 60-day mortality in patients admitted to ICU after subsequent referral compared to patients admitted immediately; no significant effect on 28-day mortality.
Rubinson, 2013, USA
Study year(s): 2009 / 4,549,623 total admissions (number with each diagnosis not reported)
661 hospitals
Acute MI, CHF, stroke, traumatic injury, pediatric patients with traumatic injury or chronic comorbidies / Metric: 1. Standardized number of weekly admissions (hospitals categorized high, medium, and no surge). 2. Weekly census: bed ratio.
Timing: Weekly / Increased hospital mortality for patients with stroke (15% increase) and acute myocardial infarction (20% increase) at hospitals experiencing high surge during influenza pandemic compared with no-surge hospitals. No effect of high surge on mortality in other patients.
Schilling, 2010, USA
Study year(s): 2003-2006 / 166,920 adults
39 hospitals
Acute MI, CHF, stroke, pneumonia, hip fracture, gastrointestinal bleeding / Metric:1. Daily hospital occupancy, in tertiles. 2. Widespread or regional influenza activity
Timing:Day of admission / 6% increase in hospital mortality among patients admitted on high occupancy days. 12% increase in mortality among patients admitted during widespread or regional influenza activity.
Schwierz, 2011, Germany
Study year(s): 2004 / 937,360 adults (unclear how many analyzed for outcomes)
72 acute care hospitals / Metric:Unexpected, high hospital department census (after controlling for expected hospital department census based on day of week, month, and public holiday)
Timing:Day of admission / No increase in 24-hour mortality among patients admitted on days with unexpectedly high census (after adjusting for changes in unobserved risk characteristics on high-census days).
Serafini, 2015, Italy
Study year(s): 2012 / 3,828 adults
1 acute care hospital
Conditions requiring medical admission / Metric:Admission to outlying unit (Surgical) instead of home medical unit
Timing:Time of admission / Among medical patients boarding in surgical wards, 1.8-fold increased adjusted odds of hospital mortality.
Singer, 2011, USA
Study year(s): 2005-2008 / 41, 256 adults
1 suburban teaching hospital / Metric:ED boarding: ≥2 hours between admission request and admission
Timing:Time of admission request / Compared to patients who did not board in the ED, patients who boarded ≥12 hours had 23-43% increased hospital mortality.
Sprivulis, 2006, USA
Study year(s): 2000-2003 / 62, 495 adults
3 hospitals / Metric: 1. Percent occupancy at midnight. 2. Percentage of ED capacity occupied by patients experiencing access block (>8h wait for inpatient bed). 3. Overcrowding hazard scale: Composite measure of hospital crowding (level 1, 2, 3) multiplied by access block (level 1, 2, 3).
Timing: 1. Midnight on day of ED presentation. 2. At time of ED presentation. / Compared to patients admitted while hospital occupancy was <90%, patients admitted while occupancy was ≥100% had 30% increase in 7-day mortality hazard. Overcrowding Hazard Scale >2 associated with 20-30% increased Day 2, Day 7, and Day 30 mortality.
Stowell, 2013, USA
Study year(s): 2010 / 483 adults
1 teaching hospital / Metric: Outlying patients: Patients admitted to outlying ward because of lack of vacant beds in home specialty ward.
Timing: Time of admission / No association between admission to outlying wards instead of usual specialty ward and unadjusted mortality at 24 hours, 28 days, and 90 days.
Tarnow-Mordi, 2000, UK
Study year(s): 1992-1995 / 1,050 adults
ICU admissions at 1 ICU
none specified / Metric:1. Census: ICU occupancy. 2. Workload (census + acuity): ICU nursing requirement (based on expert recommendations). 3. Composite measure: Average nursing requirement per occupied bed (high/low) and peak occupancy (high/low), 4 categories.
Timing:1. During patient's first ICU shift. 2. Average and peak values during patient's ICU stay. / Highest quartile of each workload measure (initial occupancy, initial workload, peak occupancy, average occupancy, average workload, average acuity) associated with 90-130% increased odds of mortality.
Tucker, 2002, UK
Study year(s): 1998-1999 / 13,334 newborns requiring ICU admission
186 neonatal ICUs / Metric: Census: Percentage of maximum occupancy.
Timing: 1. Noon or midnight immediately before admission. 2. Averaged throughout ICU stay. / For every 10% increase in percentage of maximum occupancy on day of admission, odds of hospital mortality increased 9%; infants admitted at 50% occupancy had about 50% lower odds of mortality compared to those admitted at maximum occupancy.
Wagner, 2013, USA
Study year(s): 2001-2008 / 200,730 adults discharged from ICU
155 ICUs in 107 hospitals / Metric:1. ICU census: All patients spending at least 2 hours in ICU each day. 2. ICU acuity, measured using mortality prediction model MPMo-III. 3. Admissions: Proportion of the daily census composed of new admissions.
Timing:On the day of discharge from the ICU / No capacity strain variable associated with increased odds of in-hospital death; discharge on days with increased admissions associated with 3% lower odds of in-hospital death.
Yergens, 2015, Canada
Study year(s): 2006-2009 / 1,770 adults
3 hospitals
Sepsis / Metric:ICU occupancy rate
Timing: Time of first ED physician assessment / Compared to patients admitted to the hospital when ICU occupancy was <80%, patients admitted when ICU occupancy was ≥90% had 72% increased odds of hospital mortality.
Outcome: Adverse Events
Ahyow, 2013, UK
Study year(s): 2006-2008 / 93,190 adults
3 hospitals / Metric:Census: Proportion of available beds that were occupied at midnight in every ward
Timing:Every day from admission to 2 days before Clostridium difficile diagnosis (each patient-day analyzed independently) / Patients on wards with ≥80% occupancy had 52-56% higher rates of C. difficile infection than patients on wards with occupancy <70%.
Gilligan, 2008, Ireland
Study year(s): 2004-2005 / 13,357 patients (3.6% < 18 years)
1 tertiary care hospital / Metric: ED boarding: Number of admitted patients boarding in ED at 9am.
Timing: Day of ED presentation / No association between number of boarders at 9am and MRSA diagnosis.
Howie, 2008, UK
Study year(s): 2006 / 619 patients requiring ICU admission
1 urban general hospital
MRSA infection / Metric: Census: Bed occupancy >87.5%
Timing:Daily / Frequency of MRSA acquisition was 53% higher on days when average ICU bed occupancy was above 87.5%.
Liu, 2011, USA
Study year(s): 2004-2005 / 1,431 adults
2 teaching hospitals
Chest pain, pneumonia, or cellulitis / Metric:1. ED boarding > 2 hours after admission decision. 2. Boarding time (continuous and intervals)
Timing:Time of admission / No association between ED boarding and adverse events.
Pascual, 2014, USA
Study year(s): 2005-2010 / 8,626 adults requiring surgical ICU admission
1 tertiary care hospital with multiple ICUs / Metric:ICU boarding: Surgical ICU patients boarded in an alternate overflow ICU when surgical ICU at capacity; boarding ICU distance from home ICU / Among patients boarding in alternate units, increase in aspiration pneumonia from 2.2% to 3.4%, and in any complication (excluding death) from 14.3% to 15.6%.
Tibby, 2004, UK
Study year(s): 2001-2002 / 816 children requiring pediatric ICU admission
1 urban hospital / Metric:1. Census: Bed occupancy at start of shift, 2. Turnover: Number of admissions and discharges during shift, 3. Acuity: Average patient dependency during shift.
Timing: Twice daily (each shift analyzed independently) / Combination of high pediatric ICU bed occupancy and high dependency associated with 1.9-fold increased odds of patient-related adverse event.
Tucker, 2002, UK
Study year(s): 1998-1999 / 13,334 newborns requiring ICU admission
186 neonatal ICUs / Metric: Census: Percentage of maximum occupancy.
Timing: 1. Noon or midnight immediately before admission. 2. Averaged throughout ICU stay. / No association between occupancy and nosocomial bacteremia.
Weissman, 2007, USA
Study year(s): 2000-2001 / 6,841 adults
4 hospitals / Metric: 1. Census: Occupancy. 2. Turnover: Admissions and discharges per day. 3. Intensity: Severity-weighted census. 4. Weekday/elective admissions.
Timing:Daily (each patient day analyzed independently) / In one hospital with 97% median occupancy, each admission increased risk of adverse event by 0.5%. No consistent effect of strain in remaining three hospitals, all with lower median occupancy (62-88%).