Geographic Variation In Critical Care Bed Supply And Intensive Care Unit Case Mix: Does Bed Supply Drive Utilization?

Abstract

Rationale: Variation in the regional supply of critical care beds potentially creates differences in the ability to admit patients to the intensive care unit (ICU) who do not require life-sustaining interventions. Objective: We sought to determine the cross-sectional relationship between regional critical care bed supply and ICU case-mix. We hypothesized that ICUs in regions with a greater supply of critical care beds would admit a smaller proportion of mechanically ventilated patients. Methods: We analyzed the administrative claims for all adult patients (age >14 years) admitted to intensive/coronary care units in acute care hospitals across four states (MA, NC, NY, WA) participating in the Healthcare Cost and Utilization Project's State Inpatient Database in 2007. Geographic boundaries were defined using Dartmouth Atlas hospital referral regions (HRR). We estimated intensive care unit beds per-capita within each HRR from Medicare’s Healthcare Cost Report Information System linked to 2007 US population estimates. Mechanical ventilation was defined using ICD-9-CM procedure codes (96.7X). We used generalized linear mixed models to estimate the relationship between the odds of mechanical ventilation and ICU bed supply, adjusting for patient (e.g. demographics, comorbidities), hospital (e.g. total beds), and regional (e.g. physician and nurse supply, disease prevalence) characteristics. Results: We identified 563,685 ICU admissions in 338 acute care hospitals within 28 HRRs. The mean age of the cohort was 65 (SD: 17) years. The median number of ICU beds per 10,000 adult residents was 2.9 (IQR 2.4-3.3). The crude percent of ICU admissions that were mechanically ventilated was 12.2% ( range: 2.0 to 18.6% across HRRs). On adjusted analysis, an increase in the regional supply of critical care beds by 1 bed per 10,000 adult-population corresponded to a 23% decrease in the odds that an ICU admission was mechanically ventilated (OR 0.77, 95% CI 0.62-0.97). The highest HRR quintile for critical care bed supply (4.23 ICU beds/ 10,000 adults) had adjusted mean (SD) rates of mechanical ventilation of 8.9% (95% CI 6-12%) of ICU admissions while 13.3% (95% CI 11-16%) of ICU patients were ventilated in the lowest HRR quintile of ICU bed supply (2.12 ICU beds / 10,000 population). Conclusion: Fewer ICU beds per capita correlated with a higher percentage of admissions to ICU with mechanical ventilation. This finding suggests that availability of ICU beds may help determine the regional case-mix.