International comparability of patient safety indicators in 15 OECD member countries: A methodological approach of adjustment by secondary diagnoses

Objective. To improve the international comparability of patient safety indicators based on administrative hospital data, adjustment of country-specific rates by a proxy measure of diagnostic coding intensity was tested. Data Sources. Secondary data (numerator and denominator counts of patient safety indicators) based on adults discharged from acute care hospitals between 2006 and 2008 was used. Study Design. A retrospective cross-sectional study using hospital administrative data was performed. Data Collection. Belgium, Canada, Denmark, Germany, Italy, Ireland, New Zealand, Norway, Portugal, Singapore, Spain, Sweden, Switzerland, the United Kingdom, and the United States provided data according to detailed instructions. Principal Findings. Age- and sex-standardized rates varied across countries. An ordinary least squares regression model was estimated for each Patient Safety Indicator (PSI) using the mean number of secondary diagnoses among denominator cases as the predictor (R ²=23 percent to 56 percent). Estimated country-specific residuals were linearly transformed into adjusted PSI rates. Variation among age-sex standardized PSI rates decreased substantially after this adjustment. Conclusions. International comparisons of health system performance based on unadjusted patient safety indicators are problematic due to suspected coding or ascertainment bias. The model could be an interim approach to provide comparable information on hospital quality, with a long-term goal of improving international consistency in diagnostic reporting in administrative data.