Hospital characteristics have long been viewed as variables to be used in either risk stratification or risk adjustment measures. The long held views were that teaching hospitals and urban hospitals have disproportionate shares of sicker patients. Hospital size (by numbers of beds), hospital teaching status (Y/N as well as residents/bed count ratios), hospital location (within a Core Based Statistic Area (CBSA) for urban versus rural designation) and hospital ownership or control type (for private/profit, private/non-profit, public/government, etc) are all hospital characteristics that have been used in risk stratification or risk adjustment measures.
In this study, researchers examined the statistical relationship between hospital ownership and teaching status and hospital rates for potentially preventable adverse events measured using Patient Safety Indicators (PSI) developed by the Agency for Healthcare Research and Quality (AHRQ). Using a nationally representative sample of hospitals grouped into mutually exclusive combinations of control/ownership, teaching status, and rurality was defined using the Nationwide Inpatient Sample data set for the year 2000. Hospital rates for 5 categories of preventable adverse events were measured in 3 forms: unadjusted, risk-adjusted, and risk-adjusted ratios with smoothing. Multivariable regression analysis was used to measure the statistical significance of the relationship between hospital type and rates for potentially preventable adverse events, with adjustments for differences in hospital bed size and region.
This analysis found a lack of any relationship between categories of hospital type and quality care measured by alternative indicators of potentially preventable conditions. Hospital ownership and teaching status are not consistent predictors of differences in rates of potentially preventable adverse events as measured by the AHRQ’s PSI standard performance measures. These hospital characteristics explain little of the observed variation in the rates of these events across hospitals.