Adults with type 2 diabetes expressed the strongest preferences for physical activity programs offering larger financial incentives, direct payments to the participant, and individual rather than competitive performance. In an online discrete choice experiment involving 277 adults receiving care through UC San Diego Health Family Medicine, incentive size had the largest effect on preference, accounting for 24%, followed by the incentive recipient at 18%. The type of payment had the smallest effect, at 5%. These results concern preferences among hypothetical programs and do not demonstrate that incentives increased physical activity or improved diabetes-related health.

The study enrolled adults aged 18 to 75 years who had a physician’s diagnosis of type 2 diabetes. Their mean age was 60.2 years, with a standard deviation of 11.51 years, and 58% were female. Each participant completed 15 choice sets, each presenting two hypothetical physical activity programs side by side. Experimental designs generated combinations of seven attributes: incentive frequency, collaboration through individual or team performance, gain- or loss-framed incentives, incentive size, competitiveness, payment as cash or a voucher, and payment to the participant or a charity. Participants selected the program they preferred and indicated whether they would enroll in it.

Researchers estimated part-worth utilities for the attribute levels using hierarchical Bayes methods. Larger incentives produced an estimated coefficient of 1.12, with a 95% confidence interval of 0.96 to 1.29, while payment directly to the individual was associated with a coefficient of 1.10 and a 95% confidence interval of 0.88 to 1.25. Individual performance was also preferred to the alternative examined, with a coefficient of 0.73 and a 95% confidence interval of 0.61 to 0.86. A noncompetitive design had a coefficient of 0.52, with a 95% confidence interval of 0.41 to 0.64. Payment upon completion of the physical activity program was favored over the other frequencies assessed, yielding a coefficient of 0.42 and a 95% confidence interval of 0.25 to 0.60.

The relative size of the reward and its recipient mattered more in the preference analysis than whether the incentive was delivered as cash or a voucher. The findings identify a combination of attributes that participants were modeled as favoring: a larger reward, paid to the individual when the program was completed, linked to individual performance, and structured without competition. This is a design signal rather than evidence of program effectiveness. The study did not test an incentive intervention in practice, and it did not measure whether such a program would produce greater physical activity, sustained attendance, or improved health outcomes. The population was also drawn from a single health system’s family medicine setting, limiting the basis for broad claims about all adults with type 2 diabetes.

A simulation model estimated that 84% of participants would enroll in a hypothetical program combining all the preferred attribute levels. In contrast, 15% were projected to enroll in a program combining the least preferred level of every attribute. These are modeled estimates of willingness to enroll, not observed recruitment, participation, or adherence rates. The researchers said that testing programs based on the preferred design in real-world settings would be necessary to determine actual uptake and adherence. Until such testing occurs, the experiment indicates how this group of adults chose among hypothetical financial incentive structures, rather than showing that any particular structure is effective at motivating exercise or managing type 2 diabetes.

Subject of Research: Preferences for financial incentive design in physical activity programs among adults with type 2 diabetes

Article Title: A discrete choice experiment for financial incentive-based physical activity programs among adults with type 2 diabetes

Article References: Castro, E., Carson, J., Bruhn, E., Fiorella, M. R., Sieber, W., Shi, Y., Mitchell, M. S., & Larsen, B. (2026). A discrete choice experiment for financial incentive-based physical activity programs among adults with type 2 diabetes. BMC Public Health. https://doi.org/10.1186/s12889-026-29802-y

Image Credits: AI Generated

DOI: 10.1186/s12889-026-29802-y

Keywords: type 2 diabetes, physical activity, financial incentives, discrete choice experiment, program design, enrollment preferences, hierarchical Bayes, patient preferences, exercise programs, discrete, choice, experiment