AN ANALYSIS OF INTEROBSERVER RELIABILITY AND REPRESENTATIVENESS OF DATA FROM THE VETERANS AFFAIRS COOPERATIVE STUDY ON PROCESSES, STRUCTURES, AND OUTCOMES IN CARDIAC-SURGERY

The authors made some preliminary judgments regarding the reliability and representativeness of the data in the early stages of the Veterans Affairs Cooperative Study entitled Processes, Structures, and Outcome s of Care in Cardiac Surgery (PSOCS). Preliminary PSOCS interobserver reliability and potential patient and site selection bias reported wer e based on comparisons with identical risk, procedure, and outcome dat a items collected independently in the Continuous Improvement in Cardi ac Surgery Study. PSOCS interobserver reliability for this limited set of variables was good to excellent. At the six pilot centers, there w ere few important differences between patients entered into PSOCS and those not entered. The 14 Veterans Affairs medical centers that will p articipate in the full-scale PSOCS study and the 29 nonparticipating c enters exhibited similar patient populations. A future study examining the process and structure variables of PSOCS will be valuable.