On the influence of data source in aggregated data studies: a comparative study of suicide information based on death certificates and judicial files.

Abstract

The aim was to assess the differences in suicide data obtained from different database sources. Death certificate based data on suicidal deaths were compared with the information obtained from the non-natural death investigation files at the Public Prosecutor's Office. The study was confined to the time period of 1981-1984 and to the residents of the district of Leuven (Belgium) who died in the district. Differences were found in the number of suicidal deaths reported as well as in the demographic variables, the distribution of suicide methods, and the place of death. Sixty percent of the records in both databases could be matched for all variables studied (age, gender, civil state, place of death, and suicide method), and another 10% if only place of death was allowed to differ. For 4.5% of the cases in the judicial files and for 8% of the records in the official statistics database no corresponding record from the other data source could be found. Based on simple probability statistics [P(missing record)+P(different information[present)] the possible minimum global information bias could be estimated to range from 7 to 42%. Different epidemiological pictures of suicide mortality may result from studying different data sources.