Hester van Eeren

General discussion | 6 145 | in youth care, to support this field in evaluating interventions on cost-effectiveness. The Netherlands organization for health research and development (ZonMw) recently funded research into the cost-effectiveness of youth care interventions (ZonMw, 2016). Part of this project was finding out whether economic evaluations in the youth sector could be standardized and which methodological issues and practical challenges would appear when applying economic evaluations in youth care (Dirksen & Evers, 2016). Some of the issues found were concerned with the perspective of the economic evaluation: If a societal perspective is taken, how are costs and effects distributed over different stakeholders? What time horizon should be used? This would preferably be a long-term estimation to model cost-effectiveness into adulthood, especially because childhood risks can predict economic burden into adulthood (Caspi, Houts, Belsky, Harrington, Hogan, Ramrakha, et al., 2016). However, long-term follow-up data is often lacking and if modelling over long-term, what effects should be taken into account? Another issue of concern is the identification, measurement, and valuation of costs and outcomes: Should we measure costs of the child or adolescent alone, or also costs of the parents, and how are all service types valued if these are not mentioned in the Dutch costing manual (Hakkaart van Roijen, van der Linden, Bouwmans, Kanters, & Tan, 2015)? Another question is whether we should use a generic outcome measure such as Quality Adjusted Life Years (QALY) in health economics, or not? Some of these issues also became apparent while conducting the research in this thesis, since we applied economic evaluation methods in youth care. For example, we modelled long-term cost- effectiveness estimates because we made assumptions on how these effects would last over time. Measuring these effects over time would be preferred, but highly depends on the study design and needs including adolescents and parents for a longer period. Although we measured the effect of the interventions with criminal activity free years, this measure cannot broadly be applied to compare interventions not aimed at reducing criminal activity in youth care. Therefore, a generic outcome which is clinically relevant as well is highly recommended. Though this thesis showed that it is possible to conduct economic evaluation studies in youth care, in light of the issues raised by Dirksen and Evers (2016) it is important to develop standards to conduct or at least report on these issues in an economic evaluation study to be able to compare studies and the results of the interventions evaluated. The methodological and practical challenges raised by Dirksen and Evers (2016) thus provide a starting point to conduct further research and develop future guidelines. Moreover, there are prominent and important steps made in conducting economic evaluation studies in youth care in the Netherlands, of which Kremer and colleagues (2016) gave a first overview. They recommend to validate the model used, to use a generic outcome measure such as the QALY, to include all relevant costs and to report carefully and in detail about the analyses to be able to compare the results with other studies in youth care. Fourth, an additional recommendation for future studies pertains to the judicial context of some interventions. Adolescents referred to FFT and MST, for instance, often have a court order (Baglivio, Jackowski, Greenwald, & Wolff, 2014). When interventions are used in a judicial context, criminological theories and studies should be considered in evaluating their effectiveness and cost-effectiveness (i.e., Velthoven, 2008). For