The social determinants of health
As we continue the transition toward value-based care, provider organizations are reexamining the concept of health care services and evaluating how social, economic, and environmental factors influence our health and quality of life. While estimates of impact vary according to different sources, our health status and overall wellbeing are driven by a number of factors:
- Traditional health care (10% to 25%);
- Genetics (up to 30%);
- Physical environment (5 to 10%);
- Social and economic factors (15 to 40%); and
- Individual behavior (30 to 40%).
According to the World Health Organization, the social determinants of health "are the conditions in which people are born, grow, live, work, and age." These factors (outlined in the figure below) profoundly influence access to and effectiveness of care, and they highlight the multifactorial nature of health outcomes and population health.
IT plays a critical role as providers work to manage the social determinants of health in their patient populations. Here's how other organizations have leveraged IT to address social determinants.
Examples of organizations addressing social determinants using IT
Mission Health Partners (MHP)—This Medicare Shared Savings Program (MSSP) accountable care organization based in North Carolina has been partnering with local agencies to identify and close gaps in care created by socioeconomic factors. For instance, MHP has partnered with the analytics firm Lumeris to identify high-risk patients and predict which of those individuals would benefit the most from intervention. Rather than rely on claims and clinical data alone, their predictive modeling tool also incorporates social determinants, allowing the system to tailor specific social service interventions for patients at risk of ending up in the hospital or ED – while also populating that data into the EHR. Using this expanded set of data over a six month period, MHP saw a 25% increase in the accuracy of its population risk predictions, greatly improving care coordination and reducing utilization.
Indiana University-Purdue University Indianapolis—Researchers at this public research university partnered with the Regenstrief Institute to develop algorithms that can predict the need for social service referrals at a safety-net health system in Indianapolis. Using data from 48 socioeconomic and public health indicators, the random forest decision models could help predict the need for mental health support, dietician counseling, social work services, and other related services. Data sources for the models included the health system's EHR, order entry data, clinician notes, the local health information exchange (HIE), and public health data pulled from the US Census Bureau, community health surveys, and county vital statistics. While still a work in progress, results of a study showed that the predictive algorithms yielded specificity measures ranging between 60% and 77%, depending on the type of referral.
Humana—Social determinants have played a part in Humana's Bold Goal initiative to improve the health of its members 20% by 2020. The Bold Goal program has multiple areas of focus, including food insecurity and social isolation. Humana researchers discovered that people who are food insecure are 50% more likely to be diabetic and 60% more likely to experience heart failure. In a study presented last year, Humana screened 369 members for food insecurity, and then offered Supplemental Nutrition Assistance Program (SNAP) benefits and emergency food boxes to those who qualified. As a result, the patients receiving food assistance saw the number of days they were either physically or mentally unhealthy cut in half. Additional Humana research can be found here showing how social determinants affect outcomes of the CDC's Health Days survey.
NYU Langone Health—Partnering with the Robert Wood Johnson Foundation, the Department of Population Health at NYU Langone Health built the City Health Dashboard, a public online resource that provides interactive clinical and social determinants of health data for the 500 largest U.S. cities. The dashboard currently covers 36 measures of health split across five categories: social and economic factors, health behaviors, health outcomes, physical environment, and clinical care. With this tool, users can see which neighborhoods fare worse across different health outcomes within their city, and see what is contributing to those differences—they can also benchmark their city to others within the database. In addition, the City Health Dashboard site provides resources linking users to evidence-based policies and programs to address any health and wellbeing challenges they identified in their community.