Center-specific differences in mortality: preliminary analyses using the Risk Adjustment in Congenital Heart Surgery (RACHS-1) method.

Abstract

We sought to explore the usefulness of the Risk Adjustment in Congenital Heart Surgery method (designated RACHS-1) of adjusting for case-mix differences when comparing institutional mortality after surgery for congenital heart disease. By using 1996 hospital discharge data from 6 states, centers performing at least 100 operations for congenital heart disease (patient age <18 years) were identified. Using the RACHS-1 method, procedures were grouped into 6 risk categories, and institutions were ranked in order of increasing mortality rate. A graphic display of ranks by risk category identified patterns of performance. Incorporating age, prematurity, and presence of a major noncardiac structural anomaly into multivariate models allowed computation of an overall risk-adjusted rank for each institution on the basis of its standardized mortality ratio. Among 109 centers performing 7177 operations for congenital heart disease, 22 performed at least 100 cases (72.3% of total operations). Unadjusted mortality rates ranged from 2.5% to 11.4%. A total of 4318 cases could be placed into 1 of the 6 risk categories. Few deaths occurred in risk category 1, and few institutions performed procedures in risk categories 5 and 6, making institutional comparisons in these categories uninformative. Considering mortality rates in categories 2 through 4, institutions displayed either relatively consistent ranks, a threshold increase in mortality as higher-risk procedures were performed, or a threshold decrease in mortality. Standardized mortality ratios indicated which institutions performed better or worse than expected on the basis of their case mix. The RACHS-1 method can be used to judge relative institutional performance, either by evaluating within-risk-category differences or by comparisons of observed and expected mortality rates.