Adverse events in surgical inpatients: A comparative analysis of public hospitals in Victoria

We compare adverse event rates for surgical inpatients across 36 public hospitals in the state of Victoria, Australia, conditioning on differences in patient complexity across hospitals. We estimate separate models for elective and emergency patients which stay at least one night in hospitals, using fixed effects complementary log-log models to estimate AEs as a function of patient and episode characteristics, and hospital effects. We use 4 years of patient level administrative hospital data (2002/03 to 2005/06), and estimate separate models for each year. Averaged over four years, we find that adverse event rates are 12% for elective surgical inpatients, and 12.5% for emergency surgical inpatients. Most teaching hospitals have surprisingly low adverse event rates, at least after adjusting for the higher medical complexity of their patients. Some larger regional hospitals have high adverse events rates, in particular after adjusting for the below average complexity of their patients. Also, some suburban hospitals have high rates, especially the ones located in areas of low socioeconomic profile. We speculate that high rates may be due to factors beyond the control of the hospitals, such as staff shortages. We conclude that at present, care should be taken when using adverse event rates as indicators of hospital quality