MIT Roybal Center for Translational Research to Improve Health Care for the Aging

Researchers: Marcella Alsan, Sarah Eichmeyer
Abstract: We experimentally vary signals and senders to identify which combination will increase vaccine demand among a disadvantaged
population in the United States – Black and White men without a college education. Our main finding is that laypeople (non-expert
concordant senders) are most effective at promoting vaccination, particularly among those least willing to become vaccinated. This
finding points to a trade-off between the higher qualifications of experts on the one hand, but lower social proximity to low socio-
economic status populations on the other hand, which may undermine credibility in settings of low trust.

Researchers: Matthew Notowidigdo, Craig Garthwaite
Abstract: In South Carolina, the Medicaid program is administered through Managed Care Organizations (MCOs), which offer different health care plans to Medicaid beneficiaries. These plans differ in their generosity, network coverage, and other attributes, and they are ranked by the state using a system of “star ratings.” The system of MCOs offers choices to health care consumers and allows plans to compete for consumers. In South Carolina, when consumers do not make an active plan choice, the state uses an algorithm to assign plans to consumers automatically. Starting earlier this year (in 2017), this auto-assignment is now being made using an explicitly random process. We propose to use this randomized assignment feature to study the effect of plan assignment on patient outcomes such as health care utilization and health care expenditures (both overall and by category). This prospective analysis will be complemented with a retrospective analysis that takes advantage of the state’s historical quasi-random round-robin assignment procedure to allocate households to plans. Additionally, we propose to combine the analysis of the randomly assigned population with the population that made active choices to try to distinguish between treatment and selection in accounting for which plans perform better.

Researchers: Esther Duflo, Abhijit Banerjee, Arun Chandrasekhar, Benjamin Olken, Emily Breza, Marcella Alsan, Paul Goldsmith-Pinkham
Abstract: Despite widespread availability of COVID-19 vaccination in the U.S., many people remain unvaccinated. Understanding the effectiveness (or lack thereof) of popular outreach campaign strategies for the vaccine is therefore critical. In this paper, we conduct a very large scale randomized study of the impact of two outreach strategies that were central to many vaccination campaigns.

Physicians are some of the most trusted figures in the US, and during the COVID-19 pandemic, direct appeals by physicians and nurses have been used as a strategy to promote protective behavior, particularly vaccination. Another cornerstone of the vaccination strategy, particularly in the United States, has been the use of local leaders, influential people, and simply friends and family, to promote immunization in their social networks. In this project, we tested the impact of both direct communication by doctors and efforts to motivate individuals to convince their friends to be vaccinated.  

Physicians and nurses recorded short videos to promote COVID-19 vaccination and address common doubts. In December 2021-January 2022, these messages were diffused as sponsored messages to Facebook users in randomly selected counties. In a second group of areas, Facebook users received a doctor-recorded video encouraging them to become an immunization ambassador for their friends. The message directed them to a site from where they could easily share any of the videos diffused in the first treatment. In a third group of areas, users were encouraged to think about their most influential friend, and to forward them the message encouraging them to become an immunization ambassador. Finally, there was a control group. In total 17,828,769 Facebook users received at least one video in the US.

Our primary outcome variable is the number of vaccinations delivered in the weeks following the intervention, obtained from administrative data (first doses, second doses and booster). We cannot reject the null of no effect on vaccination of any of the interventions and on any of the outcome variables. Early in 2022, near the peak of the Omicron wave, physicians messages were not sufficient to affect the decision to get immunized against SARS-Cov2, regardless of whether they were targeted to those not yet vaccinated or to those who were convinced of the benefits of vaccination and were tasked to convince others. 

Researchers: Jessica Cohen, Mark Clapp
Abstract: Chronic health conditions affect most older adults. Preventative medicine and risk management strategies, especially when applied earlier in life, are essential to altering the trajectory of a disease and ultimately improving health outcomes. Primary care providers (PCP) often provide most of these services, though younger adults are the least likely to receive primary care. This project leverages a period of high engagement and health activation during a woman’s life (pregnancy) to nudge her toward use of primary care after the pregnancy episode. With <33% of women currently receiving primary care services following a pregnancy, this randomized controlled trial will test the hypothesis that a behavioral science-informed intervention, incorporating defaults and salience, can increase the rates of PCP follow-up within 6 months following a delivery for women with hypertension, diabetes, obesity. If successful, this intervention could serve as a scalable solution to increase primary care use and preventative health services in a population that currently has low rates of engagement and utilization of these services.

Researchers: Rebecca Dizon-Ross, Ariel Zucker, Michael Fendrich, Mindy Waite, Richard Rosenthal
Abstract: Combating the epidemic of overdose deaths is a central challenge of U.S. health care policy, and substance use disorder (SUD) is increasingly harming older adults. One approach to treating SUD, incentivizing abstinence from substance use, has repeatedly proven effective but has limited evidence base in older populations. A uniquely scalable modality – of increasing interest to policymakers – for implementing this approach is with a mobile application (app) that delivers incentives for abstinence from opioids and stimulants. However, one concern about using an app for incentives is that it may hinder take-up and viability among older populations. Thus, testing the app specifically among older adults is of critical policy importance.
This project will conduct a second phase of a randomized controlled trial (RCT) to evaluate app-based incentives for abstinence from opioids and stimulants. The first phase of the RCT focused on evaluating app-based incentives among all adults and on comparing different incentive variations. The second phase will focus on the effectiveness of the app for older adults (aged 55 and over), and on understanding whether the acceptability and effectiveness of the app is decreasing with older age.

Specifically, our primary research aim is to assess whether app-based incentives are effective for older adults. Our secondary research aims are to understand whether older adults have lower take-up and engagement with the app than younger adults, and if app effectiveness is mediated by age, thus shedding light on the necessity of age-specific app design features. Effects on abstinence outcomes will be measured for all participants who engage with the app (N=59), and take-up and engagement outcomes will be measured for all participants who are eligible for the RCT even if they do not enroll (N=98). We project that 15 percent of participants will be older adults. In order to complete our research aims, we will then analyze data from the Roybal-funded study phase jointly with our previous study phase; thus, our power calculations refer to sample sizes of both study populations combined.