Session: Posters
Room: TBA
Time: Fri 13:00-14:30
Presenter: Marie Kruse (University of Southern Denmark. National Institute of Public Health)
Background: Previous analyses have shown that victims of violence have significantly higher health care costs than a comparable group of non-victims, driven in particular by higher costs of psychiatric treatment among victims. We wanted to investigate whether victims of violence had higher health care costs a priori and whether characteristics other than violence determine the level of health care costs.
Aim: The aim of this study was to analyse how the higher health care cost associated with violence related to time from exposure, number of events, gender, co-morbidity and socio-economic characteristics. Also, if violence exposure per se had an impact on health care costs or if the high health care costs associated with violence were driven by risk factors for violence alone.
Data: The data for this analysis was Danish register data on victims of violence (registered by emergency departments and the police) and their health care costs, comprising hospital costs, primary health care costs and pharmaceuticals costs in the period 2002-07.
Methods: Firstly, in a panel data analysis, health care costs were arranged according to the time of first event (exposure to violence). We applied random effect models to investigate the development of health care costs before and after exposure. Secondly we selected a reference population of non-victims, by developing propensity scores for violence exposure based on selected risk factors for violence and subsequently matching victims with non-victims by this score. In the matched population, we used conditional logistic regression models to assess whether violence caused high health care costs. As a sensitivity analysis, we matched victims with non-victims by socio-economic and demographic parameters.
Authors:
Software © 2010 iHEA - International Health Economics Association