PHILADELPHIA —Every dollar spent on patients receiving support from Penn Medicine’s community health worker (CHW) program resulted in an annual return on investment (ROI) of $2.47 for every dollar invested annually by Medicaid, according to a new study published online today in Health Affairs. The savings are generated by reducing hospitalizations.
Developed at Penn, IMPaCT (Individualized Management for Patient-Centered Targets) addresses unmet social needs such as housing and food insecurity and transportation needs in underserved populations with the goal of improving health. The evidence-based program hires and trains trusted neighborhood residents to become CHWs who carry out culturally appropriate outreach activities, social support, patient advocacy, and health system navigation.
“Within a Medicaid population, we now know that investing in our communities with this type of robust and personalized social support not only adds meaningful health value but also meaningful financial value,” said senior author Shreya Kangovi, MD, MSHP, founding executive director of the Penn Center for Community Health Workers and developer of the IMPaCT program. “These results paint a more realistic picture of the return on investment from a community health worker program that past studies haven’t accurately shown.”
The findings provide the latest round of evidence for state Medicaid programs or health systems looking to invest in proven community health worker programs to improve health and lower health care costs.
The study is the first economic analysis of a health system–based community health worker intervention for adults that used data from a randomized controlled trial. Past economic studies, have either not been rigorous enough or exaggerated savings because the reductions in spending were not attributable to an intervention but rather to random variation, or regression to the mean.
The researchers analyzed inpatient and outpatient costs on 302 patients (150 randomly assigned to the intervention arm and 152 to the control arm) who were treated within the University of Pennsylvania Health System between July 2013 and October 2014, along with the return on investment for a team of six IMPaCT CHWs. The patients were insured by Medicaid or uninsured, residents of high-poverty neighborhoods, and had been diagnosed with at least two chronic diseases (diabetes, obesity, tobacco dependence, or hypertension).
The intervention arm had both fewer and lower cost admissions, with a total inpatient cost of $2.3 million compared with $3.7 million in the control arm. Overall, the intervention arm had 149.6 hospital admissions per CHW team, while the control arm had 212.7 admissions per CHW team.
When outpatient costs were factored in, the total cost of care was $2.5 million for the intervention arm and $3.9 million for the control arm. Thus, the intervention resulted in a 38 percent reduction in cost. Overall, the team of community health workers saved Medicaid $1.4 million over one year. This savings divided by program expenses which fund salaries for CHWs, supervisors, transportation, rent, etc. (about $568,000) yielded a return of $2.47 for every dollar invested.
CHWs from IMPaCT meet with patients regularly to encourage healthier behaviors, and otherwise provide support for the patients’ own health goals by addressing social determinants of health. That can include helping them identify and purchase healthy food, connecting them with recreational activities such as urban gardening or simply being a shoulder to lean on.
Importantly, while some community health worker programs focus on patients more likely to incur high costs, IMPaCT casts a wider net to capture patients with not only frequent hospitalizations, for instance, but also other measures, such as patients with chronic diseases who never go to the hospital.
“This proves that even when applied to a broader population, IMPaCT still provided a significant return on investment,” Kangovi said. “What’s more, the return on investment underestimates the true social return.”
The randomized controlled trial not only demonstrated financial ROI, but also improved quality of care, and modest improvements in cigarette smoking, obesity, diabetes severity, and mental health. Two other randomized controlled trials of IMPaCT from Penn Medicine researchers over the past several years have similarly demonstrated its benefit. A 2014 study of hospitalized patients demonstrated improved mental health, quality of care and lower recurrent admissions. A recent multi-center randomized trial found that IMPaCT reduced hospital stays by 65 percent and doubled the rate of patient satisfaction with primary care.
IMPaCT has served more than 12,000 patients in the Philadelphia region, including hundreds of veterans at the Corporal Michael J. Crescenz VA Medical Center. In the last three years, IMPaCT has become the most widely disseminated community health worker program in the United States; it is being replicated by organizations across 18 different states including Veterans Health Administration, state Medicaid programs; integrated healthcare organizations and even retailers such as Walmart.
Penn co-authors of the study include Nandita Mitra, David Grande, Judith A. Long, and David A. Ash.