Validating population-based registers for ALS: how accurate is death certification?

The purpose of this study was to determine the accuracy of death certification for amyotrophic lateral sclerosis (ALS)/motor neuron disease (MND) as recorded by the Central Statistics Office in the Republic of Ireland, and to examine its utility in capture-recapture analysis. The database of the Irish Central Statistics Office (CSO) was searched for death certificates of individuals over 15 years of age for whom ALS/MND (ICD9) was listed as a primary, secondary or tertiary cause of death from 2002 to 2006. This dataset was compared with mortality data from the Irish Register for ALS/MND for the same period. Three hundred ninety-eight cases with a diagnosis of ALS were identified through the CSO. The cause of death as identified by the CSO was 318 out of 398 (80%) cases known to the Register during the study period. ALS/MND was listed as a cause of death by the CSO in 79 (20%) cases that were unknown to the ALS/MND Register. Of these, ALS/MND was listed as a secondary/tertiary cause of death in 76, and 75% of all such deaths occurred in rural areas. Capture-recapture analysis using the CSO and ALS Register as independent datasets generated a likely over-estimation of the crude incidence of ALS from 2.1/100,000 to 3.45/100,000 in Ireland. Population-based registers are most accurate when ascertaining from a wide network of inter-dependent sources. Death certification for ALS/MND lacks both sensitivity and specificity. Capture-recapture exercises using such data sources are likely to over-estimate the true incidence of disease.