Addressing Social Determinants of Health: How Health Care Organizations Can Act

Years of research and data have shown that social determinants of health have a significant impact on the profitability and sustainability of the health care industry. In fact, when considered broadly across racial disparities, education, social support, transportation, healthy food and poverty, social determinants of health have been shown to account for more than a third of total deaths annually in the United States, and up to 60 percent of health care costs, eclipsing actual direct medical expense. This is most likely attributed to the imbalance of medical and social spending in the U.S. On average, nations that are members of the Organization for Economic Cooperation and Development (OECD) spend about $1.70 on social services for every $1 on health services; the U.S. spends just 56 cents.

To correct this imbalance, we need to shift a portion of our current health care expenditures to investments that address upstream social factors that heavily influence downstream outcomes. Evidence suggests that addressing social determinants of health is not only important for improving overall health, but also for reducing health disparities that are often rooted in social and economic disadvantages. For example, in addition to lower body mass index and fewer risk factors for chronic disease, early childhood education has been associated with higher levels of education attainment and income and lower rates of violent crime and incarceration.

Given the far-reaching impact of these efforts, the return on investments addressing social determinants accrues not only to the health system in the form of reduced health care expense, but also to the broader community. However, current financing structures make it challenging for public sectors to pool resources together and measure the “full” return of these investments, which consist of capital infusion, tools and community-level mechanisms to deliver services. Most provider organizations don’t have the means to make these investments on their own, and those that have the means will likely find it difficult to see a near-term monetary ROI, as downstream efforts take time to take effect and may be extended beyond the health care system.

Until broader community impact can be measured, such that other sectors are helping fund these services, health systems will need to be thoughtful and targeted on where and how they invest in social determinants of health to ensure a positive ROI. By developing approaches that work backward from the outcome they’re trying to change, health systems can take progressive steps toward targeting the underlying causes of these issues, rather than siloed steps that treat only symptoms.

Some of the cursory discussions of social determinants of health suggest that addressing single factors can have a large impact on outcomes. Analysis shows this not to be true, yet most organizations are still tackling these issues in a silo. For example, we’ve seen evidence that providing free transportation services to Medicaid patients does not decrease missed primary care visits, and that building grocery stores in food deserts does not alter dietary habits.

One big reason why these interventions are not showing impact is because they are not targeted at those who would benefit most, and another is that they frequently lack an agreed-upon point of accountability for integrating these social services into the broader health care planning for these individuals. For example, through our own analysis, we know that for a specific set of individuals, having a transportation barrier is associated with a 63 percent increase in risk of readmission. However, providing just a ride for those patients isn’t enough. This needs to be coordinated with a medical professional visiting the home and ensuring that the conditions are conducive to a successful recovery. This includes making sure the patient has a follow-up visit with his or her physician; conducting a comprehensive medication review; and ensuring the individual has the support they need to obtain and adhere to the prescribed regimen to avoid a readmission.

But without someone taking accountability for coordinating this transportation service with all the other services needed, the chances of avoiding that readmission are low. It’s the diffuse responsibility that’s led to symptom-focused and ineffective solutions, and that’s what needs to change to see widespread impact and an actual ROI on these types of investments.

When accountability is present, however, a chain of connections answering to one another can help identify overall goals that can be approached in a concerted way. The team can work backward from there to drive forward progressive steps toward bigger goals and address social determinants of health in ways that show marked impact on health outcomes. To help ensure that social determinants of health efforts are accountable and productive, health care organizations can use these three action steps as a guide:

1.       Define accountability. As a care team comes online, they’ll need a leader—one who is not necessarily responsible for addressing individual social determinants, but who is accountable to the patient for the results. Primary care physicians—already the “quarterbacks” for their patients’ care and accountable for total cost of care in new payment models—are perfectly positioned for this role. To succeed, though, these quarterbacks must have a strong team behind them, consisting of dedicated clinicians who are integrated into a care delivery team and who themselves are empowered to advocate for change, act on data-informed recommendations and coordinate or monitor interventions within and without the health care provider.

2.       Use AI and machine learning to create and follow a comprehensive map. To change a patient’s health status and trajectory, one needs a clear understanding of where the patient is headed, what’s pushing them in that direction and any roadblocks to better paths. Can they easily access a store that sells food appropriate to their recommended diet? If told to come in for a follow-up, can they make time during the day, or are they a sole caregiver to a disabled relative?

Disparate data sets can shed light on neighborhood food and public transit access, household type, education and financial history, clinical notes from the electronic medical record and other variables. When these data sets are aggregated, artificial intelligence and machine learning can flag variables that, when viewed together, can pinpoint both clinical and social risk factors and flag opportunities for either physician or community intervention. Such machine-learning resources can be designed to provide push-notifications and other interactive support tools that convert data sets into actionable insights while minimizing any additions to administrative time.

3.       Redefine your measurement strategy by collaborating across stakeholders on shared goals. Realizing an ROI is muddy business when the investments made affect patients from multiple touchpoints. Metric definition and metric measurement, like interventions themselves, need to extend beyond a care provider’s four walls. Work that has traditionally been done purely at the social level should now be married with health and outcomes data to more robustly predict areas of need and define success. Considerable barriers remain, as clinicians who answer to their own facility’s balance sheets must answer to financial overseers who may not be willing to count a community benefit as a realized return. We may need to see new public discussion on tax exemption and definitions of community benefit here, but there’s strong potential, if we get it right, to truly redefine managed care and community health if we can redefine the metrics of care outcomes.

Author: Andrew Snyder and Anita Cattrell

Andrew Snyder is the Chief Medical Officer and Anita Cattrell is the Chief Innovation Officer at Evolent Health. They can be reached at and