Economic evaluations of adult weight management interventions: a systematic literature review focusing on methods used for determining health impacts

BACKGROUND: One of the challenges when undertaking economic evaluations of weight management interventions is to adequately assess future health impacts. Clinical trials commonly measure impacts using surrogate outcomes, such as reductions in body mass index, and investigators need to decide how these can best be used to predict future health effects. Since obesity is associated with an increased risk of numerous chronic diseases occurring at different future time points, modelling is needed for predictions.

OBJECTIVE: To assess the methods used in economic evaluations to determine health impacts of weight management interventions and to investigate whether differences in methods affect the cost-effectiveness estimates.

METHODS: Eight databases were systematically searched. Included studies were categorized according to a decision analytic approach and effect measures incorporated.

RESULTS: A total of 44 articles were included; 21 evaluated behavioural interventions, 12 evaluated surgical procedures and 11 evaluated pharmacological compounds. Of the 27 papers that estimated future impacts, eleven used Markov modelling, seven used a decision tree, five used a mathematical application, two used patient-level simulation and the modelling method was unclear in two papers. The most common types of effects included were co-morbidity treatment costs, heath-related quality of life due to weight loss and gain in survival. Only 12 of the studies included heath-related quality of life gains due to reduced co-morbidities and only one study included productivity gains. Despite consensus that trial-based analysis on its own is inadequate in guiding resource allocation decisions, it was used in 39% of the studies. Several of the modelling papers used model structures not suitable for chronic diseases with changing health risks. Three studies concluded that the intervention dominated standard care; meaning that it generated more quality-adjusted life-years (QALYs) for less cost. The incremental costs per QALY gained varied from $US235 to $US56,836 in the remaining studies using this outcome measure. An implicit hypothesis of the review was that studies including long-term health effects would illustrate greater cost effectiveness compared with trial-based studies. This hypothesis is partly confirmed with three studies arriving at dominating results, as these reach their conclusion from modelling future co-morbidity treatment cost savings. However, for the remaining studies there is little indication that decision-analytic modelling disparities explain the differences.

CONCLUSIONS: This is the first literature review comparing methods used in economic evaluations of weight management interventions, and it is the first time that observed differences in study results are addressed with a view to methodological explanations. We conclude that many studies have methodological deficiencies and we urge analysts to follow recommended practices and use models capable of depicting long-term health consequences.