Reducing hospital readmissions using data science and a social twist

Ohio State University Wexner Medical Center uses patient's social setting to improve adherence

Written with Lise Worthen-Chaudhari of The Ohio State University Wexner Medical Center. They will be speaking at Strata Rx 2013 on the topic of this post.

Heart disease is the most common cause of death in the United States for both men and women. And yet it is mostly attributable to things we can change, like a sedentary lifestyle. The good news is that we know how to help people with heart disease improve their health and quality of life. The bad news is that many people aren’t taking advantage of these critically important resources to get better.

For patients with heart disease, attending medically supervised cardiac rehabilitation (CR) is an absolute must. CR programs aim to improve health through structured exercise, education and counseling–and their impact is unequivocal. For every additional CR class attended, the risk of future death and heart attack decreases and expected lifespan increases. For these reasons and many more, CR is considered standard of care in clinical practice guidelines for heart disease.

Despite the extensive evidence of the positive effects of CR (and the willingness of insurers to cover it), lack of attendance remains a critical barrier to healthy outcomes for heart disease patients. Up to two thirds of CR participants drop out before the end of the 36 exercise sessions reimbursed by Medicare. And many more patients never make it to the first session.

So what can we do about it? Farsite, an advanced analytics firm, has teamed up with researchers and clinicians at The Ohio State University Wexner Medical Center to take a data science, socially-integrated approach to decreasing re-hospitalizations by increasing adherence to CR. Our work has two components:

  • Identify the risk factors of not completing the CR regimen, including demographic, behavioral, locational, seasonal, and health factors. Some of these have been studied previously, but not all together, and the interaction between them has yet to be addressed.
  • Evaluate the value of a social text messaging intervention for CR adherence.

A variety of recent behavioral economics approaches to incentivizing health behavior change have shown promise in areas as varied as smoking cessation, physical exercise, and weight loss. Although monetary based behavioral economic incentive systems have been effective, there is reason to believe that non-financial incentives, especially socially-based ones, may be effective as well (and potentially far less expensive). In addition to anecdotal proof as evidenced by the rapid proliferation of consumer social networks like Facebook and Twitter, a number of recent academic studies have demonstrated that social connections can be leveraged to create powerful behavioral change, both positive and negative.

Since the majority of patients with heart disease aren’t members of the Facebook (or MySpace, or even Netscape) generation, we decided to use text messaging as our social media because many older adults, while not on Facebook, have a cell phone that can receive text messages. Existing text messaging interventions have also reported success in user adoption and behavior change. Text4baby, a program created by the National Healthy Mothers, Healthy Babies coalition and supported by the Centers for Disease Control, reported that 96 percent of users would recommend the text messaging service to a friend. In addition, text messaging designed to curb smoking has been reported as efficacious. These existing messaging programs change behavior through delivery of educational texts.

We seek to use the same technology platform, text messaging, to change behavior through socially relevant interventions that leverage a subject’s social network. Our intervention incorporates a social twist in that we are structuring the text messages to come from loved ones rather than from us, in the hopes that socially-connected messaging (e.g., “Get to that exercise class, Dad. Love, your daughter” or “You can do this Mom! Love, your son”) will be even easier to heed and more effective than educational text messaging coming from a medical institution (e.g., “Did you know that you will live longer if you exercise?”).

We’ll be creating a text messaging banking and delivery platform so we can collect messages from a patient’s loved ones at the beginning of their CR and send them to the patient over time. At the end of this initial study we’ll be able to evaluate whether socially relevant text messages are promising for promoting CR adherence. Ideally this data platform will enable smart, rapid analysis of the impact of other novel behavioral interventions on adherence and health outcomes. And over time we hope this work will result in a recommendation engine that personalizes adherence regimens for individuals based on the nature of their hospitalization, demographic profile, self-reported behaviors, and other key variables.

We all know what’s at stake here. The United States spends more on healthcare than any other nation, and yet our outcomes leave a lot to be desired. With our work we hope to (a) improve patient outcomes, (b) reduce re-hospitalizations, (c) decrease the cost of delivering health care. Today especially, hospitals have a strong financial motivation, the Hospital Readmissions Reduction Program, for improving long-term outcomes among patients. Medicare has begun levying financial penalties, in the form of payment reductions, to short-term acute care hospitals if patients treated there for heart attack, heart failure, or pneumonia experience an unplanned readmission for any reason within 30 days. In addition, hospitals with high overall readmission rates will see a maximum penalty of 1% reduction in all Medicare payments for all patient services. Medicare’s changing fee structures and imposition of financial penalties drive hospitals to seek products that provide the types of adherence solutions we are researching.

We believe the solution requires both data science and socially-relevant interventions. For instance, couldn’t health recommendations be as personalized and as easy to engage as LinkedIn’s “People You May Know” or Amazon’s “Recommendations for You”? We think that a recommendations engine, dedicated to health, could potentially reduce re-hospitalizations and increase engagement in healthy behavior. But we’re not stopping there. We recognize that our social networks provide critical support in getting us through a trauma, such as recovery from a heart attack. Why not give our loved ones a more structured way to help us live longer? We think that something as simple as texts from friends or family members might just transform the prescribed choice into the preferred choice.

An interview with Michael Gold and Lise Worthen-Chaudhari follows.



  • Become a Text4baby Partner. 2012; Available at: . Accessed 11/01, 2012.
  • Christakis, N. A., & Fowler, J. H. (2009). Connected: The surprising power of our social networks and how they shape our lives. Hachette Digital, Inc..
  • Haisley, E., Volpp, K. G., Pellathy, T., & Loewenstein, G. (2012). The impact of alternative incentive schemes on completion of health risk assessments. American Journal of Health Promotion, 26(3), 184-188.
  • Pitts P, Golick RM. United States: Hospital Readmissions Reduction Program May Impact Post-Acute Providers. Mondaq June 21, 2012; Food, Drugs, HealthCare, and Life Sciences Section.
  • Rau J. Hospitals Face Pressure to Avert Readmissions. The New York Times Nov 26, 2012;Health Section.
  • Royer, H., Stehr, M., & Sydnor, J. (2013). Incentives, Commitments and Habit Formation in Exercise: Evidence from a Field Experiment with Workers at a Fortune-500 Company. Working paper.
  • Volpp KG, John LK, Troxel AB, Norton L, Fassbender J, & Loewenstein, G (2008).Financial Incentive-based Approaches for Weight Loss: A Randomized Trial. JAMA, 300(22), 2631-2637.
  • Volpp KG, Troxel AB, Pauly MV, Glick HA, Puig A, Asch DA & Audrain-McGovern J (2009). A randomized, controlled trial of financial incentives for smoking cessation. New England Journal of Medicine, 360(7), 699-709.
  • Whittaker R, Borland R, Bullen C, Lin RB, McRobbie H, Rodgers A (2009). Mobile phone-based interventions for smoking cessation. Cochrane Database Syst Rev;4.