RESUMO
Background: The aim of this study is to assess if a Health Interview Survey (HIS) targeting the general population is an appropriate tool to collect valid data on domestic violence. Studying item non-response on the question on domestic violence and its association with socio-demographic and health characteristics compared with victims of domestic violence can contribute to this. Methods: Cross-sectional data from the Belgian HIS 2013 were analysed. A question whether the perpetrator of a violent event was a member of the respondents' household was embedded in a general topic on violence in the self-administered questionnaire. This study is limited to people aged 15+ that at least completed the first question of this topic. Socio-demographic characteristics of item non-respondents and of victims of domestic violence were explored and the association with health status was assessed through ORs calculated via logistic regression. Results: The year prevalence of domestic violence is 1.1%. Although the question on domestic violence yields a high level of non-response (62%), this does not hinder the further completion of the questionnaire. When compared with victims of domestic violence, those not responding on the question on the perpetrator have better (mental) health. When compared with those not being victim of domestic violence, victims report poorer physical and mental health. Conclusion: An HIS can be an appropriate tool to assess domestic violence in the general population and its association with health. However, a solution should be found for the high item non-response on the question on the perpetrator of the violent event.
