Closing the SDOH Gap with AI, ML, and Data Analysis

According to the Centers for Disease Control and Prevention (CDC), social determinants of health (SDOH) are the non-medical factors that influence health outcomes. These factors include conditions in which people are born, grow, work, live, and age. There are a wider set of forces and systems that shape the conditions of daily life as well, such as economic policies and systems, development agendas, social norms, social policies, racism, climate change, and political systems.

In recent years, healthcare organizations have begun to examine social and behavioral health data’s influencing on health equity.  In fact, the CDC recently announced the release of new diagnosis codes for 2023, with a focus on SDOH, that will become effective April 1, 2023. Included is a diagnosis code option that further captures SDOH problems related to education and literacy.

The Impact of Social Determinants of Health

There’s an increasing body of evidence that supports how the socioeconomic factors and everyday life choices have a direct impact on an individual’s well-being. These social determinants of health include:

• ZIP code
• Education level
• Access to food
• Physical safety
• Quality childcare
• Nutrition choices
• Employment
• Housing stability
• Reliable transportation support

Healthcare organization’s policies are determined by regulators which largely influence their interpretation of SDOH. Understanding the impact of this data can help providers and health plans address institutional biases associated with the delivery of care. They can also prevent poor decisions made by the healthcare community based on limited information collected through traditional data gathering methods (i.e., EHR). The goal of collecting SDOH data should be to place an additional lens and perspective on the various environmental factors effecting a patient’s overall health. These factors offer a more complete picture on where institutions should invest resources, services, education, support, and overall accountability for keeping all the population’s served healthier and happier.

SDOH are unique to each patient, however, patients are largely treated the same by their providers. Providers, payers, and employers have historically relied on data they’ve collected about a patient’s or employee’s health history from within their institutional systems, rather than SDOH data outside their walls. If a patient’s specific SDOH was well known by their team of healthcare providers, their treatment and care would be significantly improved. Depending only on the institution’s clinical and financial data often creates an incomplete profile of the patient or employee. It’s not uncommon for organizations to have patient data that’s not 100% accurate or has gaps in patient history.

It’s not uncommon for individuals to use the emergency department as their primary source of health care. In some areas of the country, patients are known to visit an emergency department dozens of times a year.  “Super-utilizers,” as they are referred to, typically have complex health problems, and often go to emergency departments for health issues that could be handled by primary care physicians and social workers. The U.S. Dept. of Health & Human Services defines a super-utilizer as a “patient who accumulates large numbers of emergency department visits and hospital admissions which might have been prevented by relatively inexpensive early intervention and primary care”.  

Interestingly enough, mental health visits to the emergency department are one of the most common reasons for visits in this class of patients. In 2020, 672,727 – or 1% – of Medicaid Beneficiaries consumed $174 billion of healthcare services. The top 1% Medicaid super-utilizers accounted for 25% of the 2020 Medicaid budget, averaging $260,000 annually per person. Leveraging multiple sources of patient population data that focus on the super-utilizers’ SDOH issues can be one of the most effective and economical ways of addressing inequities in the US healthcare delivery system.

Ensuring Equitable Access to Quality Healthcare

Access to quality care is another challenge many patients confront and requires a deeper assessment of the extenuating factors related to the SDOH. For example, many uninsured or underinsured individuals can’t afford the cost of quality health plan costs, or they live in communities that don’t offer qualified health care providers. Many of these individuals end up visiting the hospital emergency departments for care that could have been prevented in less costly settings or completely avoided had they had affordable health insurance. Much of this cost has to be absorbed by the health system due to the fact that the patient doesn’t have the resources needed to settle their medical debts.

According to The Wellesley Institute, if the goal is to ensure equitable access to high quality healthcare, regardless of social position, a two-pronged strategy is required:

1. Institutions must build health equity into all health planning and delivery. This doesn’t mean all programs are all about equity. However, it does mean they should take equity into account in planning their services and outreach.
   
   
2. Institutions must also invest in targeted resources and/or programs that specifically address disadvantaged populations or key access barriers. Specifically, they must pursue opportunities for investments and interventions that will have the highest impact on reducing health disparities or enhancing the opportunities for good health of the most vulnerable. To pursue any of these course of actions, holistic, timely, and accurate sources of data will be required.  

Closing the SDOH Gap with AI, ML, and Data Analysis

Data gaps can lead to inadequacies in information about patient’s health status, well-being, and social challenges. This could result in limiting the organization’s ability to implement improvements across the entire care continuum. By incorporating artificial intelligence (AI) and machine learning (ML) technologies as part of a data analysis strategy, healthcare organizations can assist in better analyzing a patient’s needs, offer the best care management options, and improve their coordination of SDOH resources.

How can a robust data analytics and AI strategy provide much richer insights into how SDOH factor impact a patient overall health? A careful assessment of food deficiencies, transportation, housing, and other socioeconomic needs make it possible to segment patients into key groups and subsequently determine the differences between patient needs at a microeconomic level. It also gives organizations the tools needed to implement value-based care driven initiatives and gather a wealth of insights around the variables that can predict a patient’s health. Some additional benefits include:

• Payers can better forecast patient healthcare costs
• Providers can better forecast healthcare demands, hospital beds, nurses, etc.
• Facilities can create alerts for anticipated patient overflow
• If known exactly how much health care a patient needs each year, payers can anticipate payer contributions, set optimal premiums, and set co-pays optimal for payer and patient
• And many, many more…

In conclusion, SDOH factors are clearly correlated with patient health, payer and patient costs, and provider demand. When SDOH data isn’t integrated into your overall data analytics strategy, your options and alignment of your business needs may be influenced by an incomplete understanding of the patient population you serve. Knowing how social determinants and health equity barriers affect these groups can positively or negatively impact time and quality of care they deliver to all patients.

There’s no secret to it. The more your organization is committed to utilizing SDOH data in a meaningful way, the greater the opportunity to improve patient outcomes for the people and communities they serve.