Long-term health outcomes and cost-effectiveness of a computer-tailored physical activity intervention among people aged over fifty: modelling the results of a randomized controlled trial

Background: Physical inactivity is a significant predictor of several chronic diseases, becoming more prevalent as people age. Since the aging population increases demands on healthcare budgets, effectively stimulating physical activity (PA) against acceptable costs is of major relevance. This study provides insight into long-term health outcomes and cost-effectiveness of a tailored PA intervention among adults aged over fifty.

Methods: Intervention participants (N= 1729) received tailored advice three times within four months, targeting the psychosocial determinants of PA. The intervention was delivered in different conditions (i.e. print delivered versus Web based, and with or without additional information on local PA opportunities). In a clustered RCT, the effects of the different intervention conditions were compared to each other and to a control group. Effects on weekly Metabolic Equivalents (MET) hours of PA obtained one year after the intervention started were extrapolated to long-term outcomes (5 year, 10 year and lifetime horizons) in terms of health effects and quality-adjusted life years (OALYs) and its effect on healthcare costs, using a computer simulation model. Combining the model outcomes with intervention cost estimates, this study provides insight into the long-term cost-effectiveness of the intervention. Incremental cost-effectiveness ratios (ICERs) were calculated.

Results: For all extrapolated time horizons, the printed and the Web based intervention resulted in decreased incidence numbers for diabetes, colon cancer, breast cancer, acute myocardial infarctions, and stroke and increased QALYs as a result of increased PA. Considering a societal Willingness-to-Pay of 20,000/QALY, on a lifetime horizon the printed (ICER =E7,500/QALY) as well as the Web based interventions (ICER = E10,100/QALY) were cost-eftective. On a 5-year time horizon, the Web based intervention was preferred over the printed intervention. On a 10 year and lifetime horizon, the printed intervention was the preferred intervention condition, since the monetary savings of the Web based intervention did no longer outweigh its lower effects. Adding environmental information resulted in a lower cost-effectiveness.

Conclusion: A tailored PA intervention in a printed delivery mode, without environmental information, has the most potential for being cost-effective in adults aged over 50.