Linkage of the Australian stroke clinical registry to the national death index.

Introduction: Data collected in the Australian Stroke Clinical Registry (AuSCR) are used to monitor the quality of acute stroke care and its impact on patient outcomes. Survival is a main outcome and so these data must be reliable. We aimed to compare mortality data recorded in AuSCR with mortality data from the National Death Index (NDI). Method(s): Following ethics and other governance approvals, personal and clinical identifiers of all AuSCR registrants in 2009-2013 were securely submitted to the Australian Institute of Health and Welfare for linkage to the NDI. We calculated the sensitivity and specificity of clinician recorded in-hospital death data in AuSCR with the NDI. Identified deaths at 90-180 days post-discharge, recorded in AuSCR at patient follow-up, were compared to deaths during this period using the NDI. Result(s): Of the 16803 records in AuSCR submitted for linkage (mean age 73 years, 54% male), 1532 (11%) were identified as in-hospital deaths according to the NDI. Excellent agreement was demonstrated for in-hospital deaths between AuSCR and the NDI (98.8% sensitivity, 99.6% specificity). At 180 days after discharge from hospital 2913 deaths were identified from the NDI. In AuSCR, 2418 of the 2913 deaths (83%) were determined during follow-up at 90-180 days. Missed deaths were recorded as lost to follow-up. Conclusion(s): A good match between the AUSCR and the NDI for in-hospital deaths validates using the NDI to capture deaths outside hospital. Linkage with the NDI adds value to AuSCR and provides more comprehensive measure of post-discharge deaths for all registrants.