Research report Selecting key variables for depression prevention: results from a population-based prospective epidemiological study Filip Smit * , Aartjan Beekman, Pim Cuijpers, Ron de Graaf, Wilma Vollebergh Trimbos Institute, The Netherlands Institute of Mental Health and Addiction, P.O. Box 275, 3500 AS Utrecht, The Netherlands Received 16 January 2003; received in revised form 27 August 2003; accepted 29 August 2003 Abstract Background: Given the existing economic constraints, prevention of depression has to be carefully targeted, and applied where it is likely to yield the highest possible health benefits at the lowest possible cost. Aim: To identify those risk factors of depression that have the greatest use potential from the perspective of prevention. Method: Data were derived from a population-based prospective cohort of 4664 adults who had never experienced a depression. Their health status was re- examined after 1 year. Incidence rate ratios (IRR), population attributable risks (PAR) and numbers needed to be treated (NNT) were calculated to create a hierarchy of risk factors. Results: Selecting high-risk groups with migraine, abdominal and respiratory complaints and markers of vulnerability or childhood trauma appears to be an indicated strategy. Conclusions: This study illustrates how epidemiology can contribute towards setting a Research and Development agenda for primary prevention of depression. D 2003 Elsevier B.V. All rights reserved. Keywords: Depressive disorder; Prevention; Potential health benefits In public mental health, much effort is devoted to treating depression. The rationale is clear: depression is a serious condition (Murray et al., 1994), it carries a poor prognosis (Cuijpers and Smit, 2002) and affects many people (Bijl et al., 1998b). However, 40% of prevalent cases are in fact new, or incident, cases (see below in this paper). It is therefore important to reduce the influx of new cases. Here, depression prevention has a keyrole to play. Several criteria—such as feasi- bility, acceptability, availability, cost and effectiveness of preventive interventions—are used to select pre- ventive interventions that warrant interest from the perspective of public mental health. Testing interven- tions against these criteria is a necessary, but slow and expensive stage in the cycle of development, research and implementation of preventive interventions. To inform and rationalise this process at its earliest possible stage, we propose using three additional criteria to identify possibly interesting preventive interventions: the candidate intervention is directed at one (or more) amenable risk factors that (1) are strongly associated with the onset of depressive dis- order, (2) result in a substantial decrease in the incidence rate if their adverse effect can be success- fully prevented, and (3) are associated with a low number-needed-to-be-treated. The benefit of these criteria is that they can be assessed before embarking 0165-0327/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.jad.2003.08.007 * Corresponding author. Tel.: +31-30-2971100; fax: +31-30- 2971111. E-mail address: fsmit@trimbos.nl (F. Smit). www.elsevier.com/locate/jad Journal of Affective Disorders 81 (2004) 241 – 249