Project at a Glance
- There are nearly 73 million cases of diabetes among adults in India
- In Tamil Nadu, 10.4 percent of adults are diabetic and the state government spends Rs. 20 billion on state subsidies for diabetic care
- During this study, 3,000 diabetics and pre-diabetics were covered
- Participants receiving incentives took more than 1,200 steps in a day
Why this study
In India, the prevalence of non-communicable diseases (NCDs), especially diabetes and hypertension, has increased manifold in recent years. Tamil Nadu has been particularly hard hit, with an estimated 10.4 percent of adults are diabetic or pre-diabetic. There is broad consensus that increasing prevalence of NCDs will put an enormous economic burden on Tamil Nadu in the coming years.
The cost for diabetes alone accounts for 1.5 percent of India’s GDP, suggesting a significant cost burden for Tamil Nadu. Moreover, the public health system bears a large proportion of these costs. In 2004, the subsidy for supporting outpatient visits and inpatient stays of NCD patients was 0.4 percent of India’s GDP, and these costs are likely to have grown over time.
Lifestyle changes and better daily disease management are recognized as promising avenues for decreasing public and private economic burden of these diseases. The former can decrease the rate of complications and avert early mortality. Hence, the Government of Tamil Nadu’s (GoTN) efforts have, so far, focused on immediate preventive strategies.
The government conducts broad screening efforts and offers free care through government clinics. However, after diagnosis, patients often do not engage in health-seeking behavior or modify their lifestyle. Finding ways to help them do so could thus, have a large impact on long-run health costs (both public and private). Tamil Nadu’s population is relatively educated, but time is a major constraint. Patients don’t lack knowledge about making lifestyle changes, they just lack motivation to do so.
In this context, a randomized controlled trial (RCT) was conducted to evaluate effectiveness of three promising interventions in reducing the burden of diabetes. The goal of each intervention was to encourage patients to modify their lifestyles (e.g., take their medicine, exercise, eat better) over a 12-week period. The interventions have potential long-run effects if they can help participants form healthy habits, which will improve their health and decrease health care costs.
The first intervention, Monitoring and Allowances, is motivated by a large body of research suggesting that one reason people fail to manage their disease is impatience: the cost of compliance (e.g., exercising, forgoing dessert) must be paid today, but the primary benefits are seen only in the distant future.
Extensive research shows that people have short time horizons and, thus, providing them with short-run positive reinforcement can be critical to bring about behavior change. This intervention has proven effective for health in many contexts. As a result, this intervention combined allowances and monitoring to encourage participants to walk a healthy amount every day.
Participants were given a step target (10,000 steps) and a pedometer. They were asked to report their steps daily, and given allowances in the form of mobile phone recharges/top-ups (henceforth, “allowances”) for each day they hit the step target. Further, within the Monitoring and Allowances group, there was variation in the way the allowances were delivered: minimum number of days in a week on which the target had to be achieved for the participant to receive allowances, and frequency (daily, weekly, monthly) at which the allowances were given.
Since monitoring itself may have positive effects, a second intervention—Monitoring Only—was included. Participants received pedometers and reported their behavior, but no allowances were given.
The final intervention, SMS Reminders, was one that GoTN originally proposed to the research team because of its low cost. The approach is also motivated by research showing that simple reminders can be effective even when people already know what they are supposed to do. This intervention provides participants with frequent SMS texts reminding them to engage in healthy behaviors.
The RCT was implemented with a sample of 3,068 adults with a pre-existing (but potentially undiagnosed) condition of diabetes or pre-diabetes. The study design involved a series of rolling activities: screening, interest assessment, baseline, phase-in, time-preference survey, intervention, midline, and endline surveys. All of these activities were based in Coimbatore, which was chosen based on secondary data and inputs from GoTN.
The RCT methodology randomly assigns individuals to receive different interventions, enabling researchers to estimate causal impacts by simply comparing the outcomes between groups that would be otherwise similar in the absence of the interventions.
In addition to RCT, a second component of the research was to conduct a broad needs assessment survey to understand the patterns of diabetes and hypertension prevalence, lifestyle management, and service use related to diabetes management in both rural and urban areas. The goal was to help inform government on its policy efforts to combat NCDs going forward.
The study found that each of the three interventions has a detectable impact on participants, but the impacts of each intervention are quite different. The Monitoring and Allowances intervention had a substantial impact on exercise: the percentage of days when participants met the daily target of 10,000 steps increased from roughly 30 percent to 50 percent, and the average steps walked increased by roughly 1,200 steps, which translates to roughly 12 minutes of exercise.
These impacts are important as they led to better health outcomes, with HbA1c (blood sugar level) falling by 0.1-0.2 percent and RBS (Random blood sugar) falling by 6 mg/dL. These effects are modest in size but still meaningful, and are similar in size (relative to the amount of exercise induced) to the health impacts of other exercise interventions among diabetics that have been studied in the literature. This is impressive since the other exercise interventions were done in more intensive and controlled manners that would not be scalable to large samples. In contrast, the Monitoring and Allowances intervention has the potential to be scaled up widely.
The effects of Monitoring and Allowances on exercise also persist throughout the intervention period, which suggests that the effects on both health and exercise may persist after the intervention, creating greater health impacts in the future. Monitoring Only had modest impacts on exercise and no detectable impacts on health. SMS Reminders on their own did not have detectable effects on exercise, other daily health management activities (such as diet or consumption of vices) or health.
Participants in this study responded positively to interventions meant to change their lifestyle habits, and the Monitoring and Allowances program was particularly efficient. But what happens when formal intervention ends and participants are left on their own? Will those changes in lifestyle habits persist over time? Can health care providers be assured that cost savings will extend across a lifetime?
To address these questions, researchers tracked whether participants continued their healthy habits after the intervention officially ended and allowances were no longer paid. They tracked participation for 12 additional weeks and found that 35 to 50 percent of the allowance effect on exercise persisted for those three months.
It is now up to the Central and state governments to experiment with measured application of interventions that address people’s short-term impatience with long-term effects.