A Randomized Controlled Trial to Improve Diabetes Outcomes Through Financial Incentives

Type II diabetes is associated with a host of adverse and costly complications, including heart attacks, strokes, blindness, kidney failure, and severe neuropathy that may result in amputations. For those with diabetes, intensive glycemic control is essential to minimize complications. Medication adherence, weight loss, increased exercise and improved diet have all been shown to significantly improve glycemic control, resulting in improved health outcomes and lower medical costs, including a reduction in emergency department visits and hospitalizations. Yet, despite the significant health benefits associated with adherence to diet and exercise regimes and taking diabetes medications as prescribed, non-adherence to all three is a significant problem. One strategy to improve adherence and thus long term health outcomes is to provide a clearer short term benefit. For example, those with consistent evidence of adherence to an exercise or medication regimen could receive subsidies or incentives.

Therefore, the investigators propose to test three theory-based intervention strategies aimed at improving diabetes outcomes amongst a population of uncontrolled patients (Haemoglobin A1c, HbA1c, levels of 8.0 or greater at baseline). The proposed 6-month study will randomise 240 participants, 60 in the control arm and 90 in each incentivized arm from the Geylang Polyclinic. The first strategy does not involve incentives but includes a Diabetes Educational Program (DEP) to help the patient manage their condition. Included in the program are text messages to encourage participants to take their medications as prescribed and prompt good dietary and exercise practices. Subsequent strategies incorporate incentives as core components. The incentives are tied either to processes aimed at improving blood sugar levels (glucose testing, physical activity and medication adherence) or directly to the intermediary outcome (blood glucose in the acceptable range). While process incentives are likely to provide more motivation for treatment adherence, as these goals may be comparably easier to meet, these incentives only reward intermediary outcomes and it might be more effective to reward successfully achieving a health outcome directly. The investigators see this as an important empirical question that will be answered by our proposed trial. Another advantage of outcome incentives is that they are likely to be more cost-effective than process incentives as these incentives are only spent on results.

Aims and hypotheses that will be tested:

• Aim 1A: To determine if adding financial incentives for diabetes management to a Diabetes Educational Program (DEP), which comprises text messaging and use of study devices to encourage patient medical adherence and prompt good dietary as well as exercise practices, is more effective at improving diabetes health outcomes compared to the DEP alone.
• Hypothesis 1A: The average reduction in HbA1c levels at 6 months will be greater for participants in the incentive arms compared to participants in the DEP arm.
• Aim 1B: To determine whether incentivising health outcome (self-monitored blood sugar within acceptable range) is more effective at improving diabetes health outcomes than incentivising intermediate processes (blood glucose testing, physical activity and medical adherence) aimed at improving the primary outcome.
• Hypothesis 1B: The average reduction in HbA1c levels at 6 months will be greater for participants in the health outcome incentive arm compared to participants in the processes incentive arm.
• Aim 2: To determine which intervention (i.e. incentivising processes or outcome) is more cost effective (incrementally) at achieving reductions in HbA1c levels at 6 months.
• Hypothesis 2: The Incremental Cost-Effectiveness Ratio (ICER) of the intervention incentivising health outcome will be greater than that of the intervention incentivising processes.
• Aim 3A: To determine whether adding financial incentives for diabetes management is more effective at improving treatment adherence (assessed based on whether self-monitored blood sugar falls within acceptable range) than incentivizing intermediate processes aimed at improving the primary outcome.
• Hypothesis 3A: The average increase in the proportion of medications and blood tests taken as prescribed and average number of daily steps at 6 months will be greater for participants in the incentive arms compared to participants in the DEP arm.
• Aim 3B: To determine whether incentivising treatment adherence (through assessing if self-monitored blood sugar falls within acceptable range) is more effective at improving treatment adherence than incentivising intermediate processes aimed at improving the primary outcome.
• Hypothesis 3B: The average increase in the proportion of medications and blood tests taken as prescribed and average number of daily steps at 6 months will be greater for participants in the outcome incentive arms compared to participants in the processes incentive arms.