[Health promotion effectiveness: testing the German statutory health insurance agencies evaluation system in health promotion, and preliminary findings from 212 health training courses].

The purpose of this study was to develop a measurement system for the evaluation of health promotion training courses offered by German statutory health insurance companies. In a field test, N=1 671 participants from 212 youth and adult courses for the promotion of either physical activity, coping with stress or nutritional improvement were included. 80% were female. Participants were questioned in a pre-post-design with a three month follow-up. The questionnaires covered health behaviour and health status. Participants' compliance and psychometric quality of the measurement instruments were good. On average, the health insurance companies assigned participants to different interventions adequately according to the participant's individual health problems. The health promotion courses triggered improvements of high effect sizes for health behaviour patterns, of moderate effect sizes for physical complaints, subjective health ratings, and health-related quality of life. Effects decreased after the end of the intervention but remained significantly above the initial values. BMI values continued their improvement after the end of the training courses. Thus, health promotion training courses generated stable health improvements of practically relevant effect sizes. The interventions provided good support and health improvements for all subgroups of participants, regardless of age, gender and educational background. Thus, the health promotion curricula of the health insurance companies offer a ubiquitous and easily accessible but effective intervention for health promotion in Germany, although men are clearly underrepresented among the participants. The trainings may be recommended and used by other health-care suppliers. The evaluation toolkit provides practical and valid instruments for a routine evaluation of health promotion trainings. It should be applied within random sampling designs. © Georg Thieme Verlag KG Stuttgart · New York.