Rationale
Diabetes prevalence has been steadily increasing worldwide and is projected to cost the global economy up to US$ 2.5 trillion, or 2.2% of the global gross domestic product, in 2030 alone [
1,
2]. In Singapore, forecasts suggest that, without interventions, the lifetime risk of developing type 2 diabetes mellitus (T2DM) will be one in two by 2050 with a concomitant increase in total economic costs to US$ 1.9 billion in 2050 alone due to increased morbidity resulting from the condition [
3,
4].
Systematic reviews and meta-analyses strongly support the effectiveness of various lifestyle interventions in reducing blood sugar levels—as measured by glycated hemoglobin (HbA
1c)—and body weight, two key outcomes that have been associated with lowering diabetes-associated health risks [
5,
6], especially among patients with sub-optimal glycemic control [
7]. Successful interventions include diabetes self-management education [
8], physical activity targets [
9], weight management [
10], blood glucose self-monitoring [
11], medication adherence [
12], and personal health coaching [
13].
Although behavioral interventions are often impractical to administer in primary care and community settings due to limited resources and infrequent patient interaction, technological advances now enable us to deliver lifestyle interventions through mobile health (mHealth) apps and devices, providing low-cost, highly adaptable, and scalable solutions. We hypothesize that a comprehensive mHealth program incorporating key lifestyle intervention strategies, with or without financial incentives for healthy behaviors, could offer a scalable, cost-effective, and potentially cost-saving approach to address Singapore’s diabetes epidemic.
In a previous feasibility study, we evaluated a proprietary lifestyle management program, GlycoLeap (designed and produced by KKT Technology Pte. Ltd., Holmusk, Singapore), which was originally developed for adults with T2DM in Singapore [
14]. The 24-week GlycoLeap program consisted of a comprehensive T2DM educational curriculum delivered online and the Glyco smartphone app, through which users logged and monitored their blood glucose levels, weight, meals, and physical activity and received personal health coaching by accredited dietitians. Although we observed statistically significant improvements in HbA
1c (− 1.3 percentage points,
P < 0.001) and weight reduction (2.3% reduction from baseline weight,
P < 0.001), the proportion of participants meeting recommended weekly self-care process targets declined throughout the 24-week period for all evaluated components, potentially attenuating the longer term benefits of sustained lifestyle management [
14]. To address this concern, we have developed a rewards program based on behavioral economic theory to complement GlycoLeap. The rewards program leverages present bias, loss aversion, and habit formation and thus offers the potential for lasting benefits.
In this randomized controlled trial, we test whether adding a comprehensive diabetes management package (DMP), with or without a financial incentives program (M-POWER Rewards), can improve HbA
1c levels and other health outcomes of individuals with T2DM. One of the health outcomes we will assess is change in body weight, as weight reduction in overweight (body mass index, BMI ≥ 23 kg/m
2) individuals with T2DM is associated with decreases in HbA
1c levels [
15]. The DMP comprises the Glyco app and the M-POWER smartphone app and also includes the following features: diabetes self-management education, physical activity tracking, weight management, blood glucose self-monitoring, medication adherence tracking, and personal health coaching. The M-POWER app serves as a one-stop portal for participants to monitor their own diabetes self-management processes and also incorporates social norms as a behavioral tool by comparing individual self-care processes with those of others (descriptive norms) and displaying congratulatory or motivating messages for satisfactory or inadequate performance, respectively (injunctive norms) [
16]. Social norms have been tested in behavioral health interventions and shown to encourage healthier food consumption [
17‐
19]. Additionally, the app includes a gamified element to harness innate motivation by displaying an individual’s current and best streaks (number of consecutive weeks where the target has been met) for all components. The M-POWER Rewards program offers rewards in the form of M-Points for both short-term processes (weekly activity targets) and longer term health outcomes (biannual HbA
1c and weight reduction goals). Our incentive strategy leverages loss aversion by disbursing M-Points as rebates for approved medical expenditures, including expenses typically incurred at usual care visits [
20]. This rebate strategy could potentially encourage greater clinic appointment attendance and lower barriers to purchasing prescribed diabetes medication, glucometer consumables for sustained self-monitoring, and other diabetes-care consumables.
Because of the high costs involved in treating people with chronic conditions, employers, insurers, and governments all have a financial incentive to contain the chronic disease epidemic. Therefore, these third-party payers have shown a willingness to invest in some level of prevention and treatment efforts. If this study demonstrates cost-effectiveness, or even cost savings, we believe that third-party payers will be inclined to fund or subsidize the adoption of our intervention and incentive program in the primary care or community setting.
Objectives and hypotheses
Primary outcome and hypothesis
The objective of this study is to determine whether complementing usual care with the DMP, with or without financial rewards (M-POWER Rewards) can improve mean HbA1c levels at Month 12 (primary endpoint) of individuals with T2DM.
We hypothesize that between baseline (date of HbA1c test blood sample collection and randomization) and Month 12, mean improvements in HbA1c level will be greatest in the DMP plus M-POWER Rewards arm, followed by DMP alone, followed by usual care.
Secondary outcomes and hypotheses
We will test for differences between groups in mean change in HbA1c levels from baseline at Months 6, 18, and 24. We will also test mean differences between groups for changes in weight and blood pressure, proportion of participants who progress to insulin, self-reported physical activity, weight monitoring, blood glucose monitoring, medication adherence, diabetes self-management, sleep quality, work productivity, daily activity impairment, and health utility index at all four time points (Months 6, 12, 18, and 24). All outcomes will be tested with the hypothesis that improvements will be greatest in the DMP plus M-POWER arm, followed by DMP, followed by usual care.
We will also test potential effect modifiers, including duration of T2DM, baseline HbA1c level, previous or existing experience with mobile apps to manage health conditions, and number of diabetes medications. We hypothesize that those with a longer duration of T2DM, no experience with health management apps, and more diabetes medications will be less successful in improving the primary outcome.
Secondary objectives
The secondary objectives are to determine net costs and incremental cost-effectiveness ratios (ICERs) of each intervention arm. ICERs will be calculated based on costs per improvement in HbA
1c at Month 12 and converted to cost per quality-adjusted life year (QALY) gained. Using established benchmarks for cost-effectiveness [
21,
22], we hypothesize that DMP alone will be cost-effective compared to usual care and, despite higher implementation cost, DMP plus M-POWER will be incrementally cost-effective relative to DMP alone.