Studying medical end-of-life decisions via death certificates in Flanders and Brussels, Belgium: a research protocol

Background: Medical end-of-life decisions (ELDs) have become an important issue in medical and societal debate. However, studying these ELDs is not evident, especially because of methodological difficulties. In 2007 we will conduct a study in Flanders and in Brussels to establish the incidence and characteristics of ELDs. The methodology of this study is outlined in this report.

Method: Similar to two earlier conducted studies (in 1998 and 2001), and in close cooperation with the Flemish and Brussels authorities, we investigate ELDs by using death certificates. Questionnaires are sent to the physicians who signed a representative sample of death certificates (deceased of 1 year or older) in Flanders as well as in Brussels. In the Flemish survey, a stratified sample is used. In the Brussels survey, a random sample is taken. Anonymity of both patients and physicians is guaranteed through a rigorous mailing procedure, involving a lawyer as intermediary between responding physicians and researchers. To increase response we used the Total Design Method. The questionnaire used in this study is nearly identical to the questionnaires used in the previous studies, and is tested and validated thoroughly. Its translation from Dutch to French is done via a comprehensive forward-backward translation procedure.

Discussion: Death certificate studies are appropriate in nationwide studies on incidence and characteristics of ELDs. Earlier studies in Belgium and other countries have shown the reliability and validity of this methodology. High response levels show that the anonymity and mailing procedure is trustworthy, which leads to results that are representative for all deaths in the year studied.