ILLUSTRATION Joe Magee/agoodson.com

SPH This Year 2021

SPH and Sharecare Release Community Well-Being Rankings

The School of Public Health, led by the Biostatistics and Epidemiology Data Analytics Center (BEDAC), and the digital health company Sharecare have released a 2019 US state rankings report using a newly derived metric, the Community Well-Being Index (CWBI).

The CWBI combines individual reports of well-being and community-level social determinants of health (SDOH) measures, and is the product of the first year of a five-year collaboration between the school and Sharecare.

Hawaii comes it at #1 and Massachusetts at #2 in the 2019 CWBI state rankings of all 50 states, which include results from over 300,000 respondents.

“Following a decade of work by Sharecare and Gallup, the CWBI provides the next generation of rankings incorporating SDOH context,” says Kimberly Ann Dukes, executive director of BEDAC, who led the SPH team on the project.

Previous rankings were based solely on individual reports of well-being, using an overall Well-Being Index (WBI) that included five inter-rated health domains: Purpose (liking what one does each day and being motivated to achieve goals), Social (having supportive and loving relationships), Financial (secure and minimally-stressful finances), Community (liking where one lives, feeling safe, and having pride in one’s community), and Physical (having good health and enough energy to get things done daily).

The Sharecare and SPH partnership made a pivotal shift in conceptualizing and measuring well-being by incorporating social determinants of health—a key population health concept that considers influences such as marginalization, pollution, and economic and educational opportunities.

The CWBI is the product of individual and lifestyle factors such as physical health and having supportive relationships on the one hand, and larger social determinants of health such as access to affordable housing and healthcare and transportation on the other.

“The rankings are intricate, and well-being is influenced by community context,” Dukes says. “Specifically, within a zip code, environmental context changes. Our goal is to capture well-being of neighborhoods (small area estimation) by obtaining comprehensive coverage of individuals residing within a neighborhood, to more accurately reflect health and well-being with a high level of precision.

“This effort will improve the health of all, by providing information to all,” says Dukes, who is a research associate professor of biostatistics. “We’re providing information to help people who are in vulnerable communities, or who are vulnerable personally, and providing policymakers the ability to make decisions based on data,” she says.

“It’s not only for people who are vulnerable—Will it help people who are marginalized? One hundred percent. But the idea is, you could also be wealthy and living in an environment that negatively impacts your health. So, it’s for all.”

In 2019, Sharecare and SPH began collecting over 600 items measuring domains of SDOH using various spatial and temporal scales, including the American Community Survey, United States Department of Agriculture, Area Health Resources, National Park Service, United States Geological Survey, Centers for Disease Control and Prevention (CDC) and National Aeronautics and Space Administration, and more.

The newly derived social determinants of health index (SDOHi) includes an overall index and five inter-related community-level domains: Healthcare access (the ratios of healthcare providers to population), Resource access (libraries and religious institutions, employment rates, and grocery stores), Food access (more specifically how far away grocery stores are from communities who often have less access to healthy food), Housing & transportation (home values, ratio of home value to income, and public transit use), and Economic security (rates of employment, labor force participation, health insurance coverage rate, and household income above the poverty level).

The newly developed CWBI combined county-level WBI scores and county-level SDOHi scores assigning equal weight to each, by aggregating county-level scores with weights proportional to county population sizes.  The CWBI now characterizes 3,140 counties in the United States (99.9 percent), representing the most comprehensive assessment of individual and community health and well-being in Sharecare’s history.

One of the main insights is that those residing in rural communities score, on average, one point lower on the CWBI than those residing urban communities. In fact, the report finds that, looking only at individual-level well-being, New Hampshire comes in at #1, but falls to #9 when taking into account community level social determinants of health factors that also shape health.

The SPH team also evaluated how strongly different determinants influence well-being, finding, for example, that housing and transportation have a bigger impact than personal relationships or even current physical health.

“I really love this project because I enjoy working with groups of people with diverse opinions, thoughts, and specialties and getting everyone to work cohesively and speak the same language,” Dukes says. “Typically, physicians, sociologists, biostatisticians, and business executives speak different languages—and now we all speak the same language, because we all have one objective: improving health for all,” she says.

“This work was the result of a true collaboration between Sharecare, led by Elizabeth Colyer, and BUSPH, anchored at the BEDAC,” says Dukes. “Not only did we have to apply and evolve methods used previously by Sharecare, we had to bring in geospatial capabilities (led by Kevin Lane) and SDOH content expertise (led by Nina Cesare),” she says.

“The most exciting methodologic work from my perspective were the advanced statistical solutions applied that included WBI mode adjustment and structural equation modeling to develop the SDOH indices (led by Pengsheng Ni), methods used to develop WBI and CWBI scores and subsequent rankings using small area estimation and multiple imputation techniques (led by Michael LaValley), and machine learning applications (led by Prasad Patil),” Dukes says. “This work would not have been possible without oversight with respect to statistical methods and reporting (led by Michael Winter), data warehouse and data science expertise (led by Joseph Palmisano), project management expertise (led by Cortney Miller) and the rest of the BEDAC team who directly supported our efforts: Keith Spangler, Biqi Wang, Leah Forman, Ahmad Yassine, Marisa Massaro, and Margaret Shea; as well as those providing administrative support: Susan Gomes, Brittany Parker, Ayesha Chowhan; and support and guidance to the team: Clara Chen, Jun Lu, Meirimar Raimundo, and Emily Sisson.” Dukes also notes her appreciation of Dean Sandro Galea’s insights and continued thought leadership in this area, and to Sharecare CEO Jeff Arnold and president Dawn Whaley “for this opportunity, encouraging our academic rigor, and trusting us as a partner.”

And this is far from the end. The collaboration with Sharecare will continue for three more years, with county special reports on COVID and structural racism, respectively, coming next.

After obtaining, cleaning, combing and analyzing all of these data that includes over 3 million individual assessments of well-being and SDOH, BEDAC also now has an unrivaled warehouse of data on wellness and the factors that shape it. That invaluable resource will be available to researchers across Boston University and at other institutions, as well as other partners in the private sector.

“I don’t believe there’s another resource like this that is available academically,” Dukes says. “Nobody’s done this before.”