Lifestyle Intervention from a Health Economics Perspective

Sammanfattning: Abstract The aim of this thesis was to perform health economic analyses of lifestyle intervention for people at a high risk of cardiovascular diseases (CVDs) and type 2 diabetes mellitus (T2DM). The results can help decision-makers to make informed decisions over implementing lifestyle intervention, taking several dilemmas into consideration. Paper I provides evidence on the cost-effectiveness of lifestyle intervention for the prevention of CVDs and T2DM. We found that 10 research groups in 9 countries demonstrated that 10 of 11 analyses of lifestyle intervention were cost-effective. Adherence to the lifestyle changes is one important predictor for the cost-effectiveness of an intervention. In Paper II, we performed a long-term cost-effectiveness analysis of a randomized controlled trial of lifestyle intervention (the Swedish Björknäs study) with a decision-analytic Markov model (DAM). A differences-in-differences approach was used to control for baseline differences between the two groups, and three-year follow-up data were extrapolated over a lifetime. The DAM predicted that the lifestyle intervention was cost-saving from a societal perspective (US$-7500; +0.46 QALY), but would no longer be cost-saving if the effectiveness lasted only for the intervention period. In Paper III, we revisited the cost-effectiveness of the Björknäs study with seven-year follow-up data, and compared a real-world control group to the within-trial control group to capture the “do nothing” scenario. An observational cohort, the Swedish MONICA study was used to identify a real-world general population, and matched with the Björknäs study participants by propensity score matching. The results showed the intervention to still be cost-saving with seven-year follow-up data (US$-6100; +0.45 QALY), and the benefit to be even higher when comparing a real-world general population (US$-16600; +0.35 QALY). In Paper IV, we used register data on the real-life healthcare resource utilization of the Björknäs participants over a 10-year period to estimate the long-term benefit with real-life data instead of DAM-based predictions. We used a generalized estimating equation and controlled for baseline characteristics including healthcare expenditures. The results showed that the intervention group had a downward trend of outpatient and total expenditures in the long term, while the control group had an upward trend. The real-life findings from registers confirmed and complemented the DAM-based findings. Lifestyle intervention has long-term benefits, and decision-makers need to implement lifestyle intervention in primary care for people at high risk of CVDs and T2DM; this will save valuable societal resources.