Study Shows Wide Variability among VA Hospitals Regarding ICU Admission Patterns
Hospital variation in critical care resource use has implications for the quality and cost of acute inpatient care. The underuse of critical care is associated with high mortality rates and greater resource use. On the other hand, unnecessary use of critical care wastes valuable resources. This retrospective cohort study sought to describe hospital ICU admitting patterns for Veterans, after accounting for their severity of illness upon admission. Severity of patient illness was measured using the VA-ICU score based on laboratory data and comorbidities around admission. Using VA data, investigators identified all adult non-surgical admissions (n=289,310 Veterans) to 118 VA acute care hospitals between 7/09 and 6/10 to address three questions: 1) What is the 30-day predicted mortality of Veterans admitted directly to the ICU from the ED or outpatient clinic? [Of the original cohort, 31,555 Veterans were admitted directly to the ICU.]; 2) For Veterans with the same 30-day predicted mortality, how much does direct admission to the ICU vary between hospitals?; and 3) Are comparisons of hospital admitting patterns dependent upon severity of patient illness? Investigators also examined the relationship between use of the ICU and 30- and 90-day mortality rates.
- About half of all Veterans in this study (53%) who were admitted directly to the ICU had a 30-day predicted mortality of 2% or less. In more than half of cases, Veterans with a predicted mortality greater than 30% were not admitted to the ICU.
- At all levels of patient risk, hospitals varied widely in the proportion of Veterans admitted to the ICU. For example, the rate of admission for Veterans in the low-risk group (predicted mortality <2%) varied from 1% to 39%, while the rate of admission for Veterans in the high-risk group (predicted mortality >30%) varied from 11% to 50%.
- Investigators also found that for a one standard deviation increase in predicted mortality, the adjusted odds of ICU admission varied substantially across hospitals, ranging from a 15% decrease to a 122% increase.
- For patients with a high predicted mortality who were not admitted, investigators were unable to determine whether patients had been determined to be "too sick to benefit", or the roles of patient preferences or goals of care in the decision. For patients with a low predicted mortality who were admitted, investigators were unable to determine whether patients were "too well to benefit", the role of patient preferences, or the availaibility of alternative environments (e.g., step down units, telemetry units).
- Study data lacked physician-level characteristics within and across hospitals, which might contribute to hospital admitting patterns.
- The proportion of low-risk and high-risk patients admitted to the ICU, variation in ICU admitting patterns between hospitals, and the sensitivity of hospital rankings to patient risk, all likely reflect a lack of consensus about which Veterans most benefit from ICU admission.
This study was supported by HSR&D. Drs. Chen, Sales, Wiitala, and Hofer, and Mr. Kennedy are part of HSR&D's Center for Clinical Management Research, Ann Arbor, MI.
Chen L, Render M, Sales A, Kennedy E, Wiitala W, and Hofer T. Intensive Care Unit Admitting Patterns in the Veterans Affairs Healthcare System. Archives of Internal Medicine 2012 Sep 10;172(16):1220-6.