Several states have implemented Healthy Behavior Incentive Programs (HBIPs) in Medicaid through Section 1115 demonstration waivers. These programs use financial incentives to encourage positive behavior changes, such as greater use of preventive services, smoking cessation, and weight loss.
To test for an association between the introduction of HBIPs and the rates of smoking cessation, weight loss, and annual preventive health visits in states that have adopted behavior-specific programs compared with states that have not.
Design, Setting, and Participants
A cohort study using a difference-in-differences analysis of the 2011-2016 Behavioral Risk Factor Surveillance Survey Interview Results data, adjusting for demographic conditions, state unemployment rates, state Medicaid expansion, national secular trends, and time invariant state-specific factors, was conducted. Two sets of participants were considered: adults aged 18 to 64 years who had a reported annual household income of less than $25,000 (n = 442 089) or adults aged 18 to 64 years who had completed high school education or less (n = 676 883).
Changes in health behavior outcomes in 4 states (Florida, Indiana, Iowa, and Michigan) that implemented behavior-specific HBIPs targeting smoking, obesity, and annual health checkups through a Section 1115 waiver, against changes in control states, including Washington, DC, that did not introduce an HBIP (n = 44).
Main Outcomes and Measures
Rate of smoking, obesity, and attendance at annual preventive health visits.
Of Behavioral Risk Factor Surveillance Service respondents used for the less than $25,000 annual household income cohort (n = 442 089), the mean (SD) age was 43.1 (0.8) years, and the mean (SD) percentage of women was 58.4% (2.5%). For the cohort of high school education or less (n = 676 883) population, the mean (SD) age was 41.6 (1.1) years, and the mean (SD) percentage of women was 46.6% (0.9%). During a 2-year period after implementation, there were no improvements in smoking and obesity in individuals with a household income of less than $25 000 (2.49 percentage points, 95% CI, 1.75-3.23 percentage points; P < .001 and −1.94 percentage points, 95% CI, −4.42 to 0.55 percentage points; P = .12, respectively) as well as in the population holding a high school education or less (1.74 percentage points, 95% CI, 0.64-2.85 percentage points; P = .003 and −0.73 percentage points, 95% CI, −1.84 to 0.38 percentage points; P = .19). An association was noted between an increase in preventive health visit rates among states adopting behavior-specific HBIPs relative to control states in the less than $25,000 household income population (3.89 percentage points, 95% CI, 2.64-5.14 percentage points; P < .001). However, these associations were substantively small and not robust across the high school education or less population (1.8 percentage points, 95% CI, −0.12 to 3.71 percentage points; P = .07).
Conclusions and Relevance
Early post implementation assessment may indicate that HBIPs were not associated with substantive improvements in incentivized healthy behaviors among populations likely to be Medicaid beneficiaries. The value, format, and timing of the incentive, complexity in delivery, and lack of awareness of incentives among target beneficiaries and clinicians may limit the usefulness of programs even over a longer follow-up period.