Hospital-level characteristics of the standardised mortality ratio for ischemic heart disease: a retrospective observational study using Japanese administrative claim data from 2012 to 2019

Ischemic heart disease (IHD) is one of the leading causes of mortality worldwide and imposes a heavy burden on patients. Previous studies have indicated that the optimal care for IHD during hospitalisation may reduce the risk of in-hospital mortality. The standardised mortality ratio (SMR) is an indicator for assessing the risk-adjusted in-hospital mortality ratio based on case-mix. This indicator can crucially identify hospitals that can be changed to improve patient safety and the quality of care. This study aimed to determine the hospital-level characteristics of the SMR for IHD in Japan. This study was designed as a retrospective observational study using the Japanese administrative claim data from 2012 to 2019. The data of all hospital admissions with a primary diagnosis of IHD (ICD-10, I20-I25) were used. Patients with complete variables data were included in this study. Hospitals with less than 200 IHD inpatients in each 2-year period were excluded. The SMR was defined as the ratio of the observed number of in-hospital deaths to the expected number of in-hospital deaths multiplied by 100.The observed number of in-hospital deaths was the sum of the actual number of in-hospital deaths at that hospital, and the expected number of in-hospital deaths was the sum of the probabilities of in-hospital deaths. Ratios of in-hospital mortality was risk-adjusted using multivariable logistic regression analyses. The c-statistic and Hosmer-Lemeshow test were used to evaluate the predictive accuracy of the logistic models. Changes in SMRs in each consecutive period were assessed using Spearman's correlation coefficient. A total of 64,831 were admitted patients with IHD in 27 hospitals as complete submission data. The SMRs showed wide variation among hospitals, ranging from 35.4 to 197.6, and analysis models indicated good predictive ability with a c-statistic of 0.93 (95% CI [0.92-0.94]) and Hosmer-Lemeshow test of 0.30. The results of chi-square tests and -tests for all variables to assess the association with in-hospital mortality were