Validation of linked administrative data on end-stage renal failure: application of record linkage to a 'clinical base population'

Objective: To evaluate the use of record linkage to monitor the occurrence of endstage renal failure in Western Australia in 1980-94. Methods: A clinical base population of 1,046 patients was identified from the Western Australian (WA) Health Services Research Linked Database. To exclude acute renal failure, patients were selected if they received in-hospital renal dialysis on more than 10 occasions over more than 28 days in 1980-94. Estimates of annual incident and prevalent cases were validated against the ANZDATA dialysis and transplant register. Reasons for discrepancy were investigated by an ad hoc linkage between the two data sources. Results:The WA Linked Database counted slightly fewer incident cases (-7%) and slightly more prevalent cases (+7%) than the ANZDATA Register. The Linked Database identified 97% of cases on the ANZDATA Register, but this fell to 83% post case definition, probably due to patients receiving home-based dialysis failing to meet our case definition. ANZDATA correctly identified 90% of cases in the linked file. Conclusion: Trends in end-stage renal failure from 1986 to 1994, based on the Linked Database, were the same as those reported from purpose-designed disease registers. Implications: Linked administrative data provide a valid and efficient means to plan and evaluate many of the routine aspects of renal dialysis and transplant services.