The Critical Role of Predictive Analytics in Addressing SDoH and Health Inequities

For comprehensive, high-quality care, proactive strategies to identify and address SDoH are key

By Linda T. Hand

While access to timely and affordable healthcare for all is a founding principle of the American ideal, disparities in health outcomes by population paint a starkly different picture. These data points highlight inequalities in care availability, quality, and outcomes. They exist across the spectrum of healthcare, from access to care and perceived trust in providers, to chronic conditions such as rates of diabetes and hypertension, maternal mortality, and even overall life expectancy. One study put it starkly: The most privileged members of society enjoy an additional 15 years of life expectancy when compared to individuals from the most disadvantaged backgrounds.

The good news is that finally, the industry is beginning to address this crucial issue as an urgent priority. From the highest offices to the most renowned medical organizations, a bolded, underlined, and italicized emphasis on health equity—and reduction of outcome disparities—is a primary strategic imperative.

Data driving the way forward 

As our industry confronts health inequities with unprecedented vigor, I’m encouraged by the innovative approaches many hospitals and health plans are exploring. Many are homing in on social determinants of health (SDoH) and are not only working to better understand the significant role they play in health outcomes, but also pioneering new approaches to address and overcome them.

One approach gaining attention is the application of predictive analytics to forecast future health risks. This risk prediction, which is based on robust algorithms that include a diverse blend of clinical, claims, and SDoH data, can help equip plans and providers with the early insights needed to proactively intervene and help patients experience better health journeys and healthier outcomes overall.

Precise member-level alerts catch behavioral, attitudinal, and other SDoH risks, and trigger flags to the relevant health plan, care coordinator, or physician so that early intervention is possible. For example, should a high-utilization diabetic member lose their housing, health plan administrators would be alerted to the change and could then connect that member with community resources to secure housing with stable refrigeration for medications. This can help ensure that the member remains stable, does not require emergency attention, and can continue adhering to protocol management for their chronic condition.

Identifying the barriers to care that impact outcomes, utilization, and costs

A recent report revealed just how outsized a role SDoH play in overall outcomes—and how using predictive analytics to more proactively address and overcome them can help patients live longer, healthier lives.

The report, based on an analysis of 154 nonclinical risk factors, revealed that these factors impact health and well-being across insured populations and medical conditions, including cardiovascular health, mental health, and maternal health. Along with the high-level findings, the report underscored the ability of predictive analytics not only to address these gaps in care, but to identify them earlier, engage members within their care journeys, and design proactive community-based interventions that can help minimize their impact on long-term health and well-being.

Cardiovascular health: The analysis found that nonclinical risk factors—such as limited access to care, poor physician relationships, and other socially contextual risks—are heavily associated with acute or emergency care needs for all insurance populations. While regular check-ups are critical for managing chronic conditions, SDoH have been shown to greatly impact the likelihood individuals will show up for these monitoring visits. For instance, unreliable access to transportation has been associated with a 15- to 20-fold decrease in likelihood for individuals with cardiovascular disease to attend these routine primary care visits; this was found to be true across both members covered by Medicare and members covered by commercial insurance.

Mental health: As one of the nation’s fastest-growing areas of care need, the overwhelming demand for services is inspiring health plans to update their approaches to mental and behavioral healthcare. By acknowledging the role SDoH play in exacerbating these conditions, stakeholders can work proactively to address them before they drive higher acuity or utilization. Notably, of all nonclinical risk factors included within the analysis, food security emerged as the most outsized influence on behavioral health utilization. For all non-commercially insured populations, limited access to healthy, nourishing, and affordable foods was strongly associated with increased utilization for behavioral healthcare services (associated with a 38- to 206-fold increase, across populations).

Maternal health: As the maternal health crisis intensifies, health plans are working to develop comprehensive care journeys that proactively identify risk in order to trigger early interventions. The report’s analysis reviewed the impact of nonclinical drivers on maternal complications, including pregnancy with complications, extreme prematurity at birth, and significant neonatal complications. Notably, the Medicaid population felt the consequences of SDoH risks most acutely, with the most impactful factors being hospital and emergency care access (seven-fold increase in utilization), financial stressors (six-fold increase), primary and public healthcare risk (6.6- and 10-fold increases), and transportation access (nine-fold increase).

From idea to action: Leveraging predictive analytics to drive lasting change

As healthcare organizations mobilize to address widening health disparities, stakeholders will need to adopt all the tools at their disposal. Knowing that health needs originate early and outside the clinical care setting, comprehensive predictive algorithms enhanced with a library of clinical, claims, and SDoH data must become part of the care delivery equation. Equipped with the insights to deploy full care plans, from preventive health literacy outreach to recurring chronic care management, plans and providers can leverage these advanced solutions to identify future health risks early and take steps to address them. Not only will this help transform healthcare delivery, but it will make meaningful strides toward a more equitable standard of care for all—one that ensures an outsized improvement in the overall health and well-being of individuals everywhere.

Linda Hand is the CEO of Prealize Health, a company that uses machine learning to transform care from reactive to proactive. She holds a BA in computer science from the University of California, Berkeley, and completed the Haas School of Business Executive Program.