SDOH Data: How Tech Helps Identify & Track Social Determinants of Health

Today’s healthcare organizations recognize the importance of addressing unmet social needs such as food, housing and transportation. Here’s how they can overcome common obstacles related to collecting, analyzing and using SDOH data.

Your browser doesn’t support HTML5 audio

In less than a decade, social determinants of health (SDOH) have evolved from a new concept into a core strategic imperative for healthcare providers, payers and government health agencies alike.

Few question the importance of addressing the nonclinical factors that impact health on an individual and population level. However, organizations often face obstacles in gathering relevant SDOH data, preparing it for analysis and putting it to use for clinical care and business planning.

“As an industry, we need to take a thoughtful and methodical approach,” says Jeremy Racine, healthcare director for Tableau. “After meaningful use, we had hundreds of electronic medical record systems but no standardization and a culture of information blocking. I don’t want to see the same thing happen again with SDOH data.”

Click the banner for access to exclusive HealthTech content and a customized experience.

SDOH: An Emerging Priority for Healthcare

The World Health Organization has defined social determinants of health as the range of nonmedical factors that influence health outcomes: “They are the conditions in which people are born, grow, work, live and age, and the wider set of forces and systems shaping the conditions of daily life.”

The U.S. Department of Health and Human Services groups SDOH into five domains: economic stability, education access and quality, healthcare access and quality, neighborhood and built environment, and social and community context. Specific social determinants include such things as safe housing and transportation, air and water quality, job opportunities and systemic discrimination.

A 2015 study in the American Journal of Preventive Medicine concluded that SDOH accounted for 84 percent of health outcomes, with medical factors accounting for just 16 percent. McKinsey estimated that individuals with unmet needs associated with SDOH are more likely to report higher healthcare utilization as well as poor physical and mental health.

DISCOVER: Why it's important to keep equity at the forefront in telehealth.

HHS has prioritized SDOH as one of the five goals of its Healthy People 2030 initiative. Providers also recognize the importance of addressing SDOH, whether the motivation is driven by a mission to improve outcomes or by a more pragmatic desire to meet regulatory guidelines such as readmission reduction.

“When healthcare organizations look at health equity, they find out there are disparities and then there’s an obligation to act,” says Andrea Green, director of healthcare strategy for SDOH at LexisNexis Risk Solutions.

However, taking action can pose a challenge for three related reasons: SDOH data is difficult to collect; analytics efforts need a clear focus; and organizations need to provide business insights to both clinical and operational leadership.

“Even if you’re able to pool the data, the challenge is discerning actionable intelligence,” says Catherine Robison, health innovation scientist with Oracle Health. “The question becomes, what action can I take that maximizes patient outcomes while controlling costs, and what’s within my ability to influence?”

Here’s what organizations can do to better use data and technology to position themselves for SDOH success.

1. Standardize SDOH Data Collection Across the Organization

Organizations cannot address SDOH at an individual or population level without knowing where inequity exists. This information often resides in several places: patient intake forms, clinical notes, discharge summaries and call center notes, to name a few.

Wherever possible, organizations should standardize the way data is collected and expressed. “You need standards and definitions, and they need to apply across the board,” Racine says. “Nearly everyone can do something with the data. Their use cases may be different, but the definitions need to be the same.”

Two existing sets of standards can help organizations do this. One is the use of Z codes, a set of International Classification of Diseases codes that document SDOH needs such as unstable housing or problems with a relationship with a partner. Z codes aren’t required for billing purposes, and they’re used infrequently — the Centers for Medicare & Medicaid Services (CMS) reported that Z codes have been captured for just 1.6 percent of Medicare beneficiaries — but they represent a codified way to document unmet needs.

LEARN MORE: What is personalized medicine and how can healthcare organizations achieve it?

The other is a set of new quality measures for health plans under the Healthcare Effectiveness Data and Information Set. One set is tied to screening of and intervention for members with unmet food, housing and transportation needs. Another is linked to race and ethnicity stratification for certain existing HEDIS measures.

Green says leveraging these standards and mandates can provide a “more complete picture” of socioeconomic barriers that exist and how they may impact care quality.

Racine agrees, adding that standards-based processes have the additional benefit of creating a “sound data culture” throughout an organization. “Think about existing processes that no one bats an eye about,” he says, whether it’s the ethical practice of medicine or computers automatically logging off a user due to inactivity. “That’s where we should be with SDOH data collection.”

2. Provide Robust Analysis of SDOH Data That Addresses a Clear Need

The goal of any analytics initiative is to link disparate pieces of information to tell a more complete story and determine a certain course of action. The best way to do that with SDOH is to start with the end in mind, Green says.

“You need to determine the needle you want to move and then work backwards — figure out the pathway and the types of data you might need,” she says. This is important, as organizations may need to augment internal data with external data sources (such as consumer financial data), and there’s a cost associated with acquiring it and harmonizing it so that it’s consistent with existing data sets. “No one has unlimited resources, so it’s best to identify the biggest impact and deploy the resources there.”

No one has unlimited resources, so it’s best to identify the biggest impact and deploy the resources there.”
Andrea Green

Director of Healthcare Strategy for SDOH, LexisNexis Risk Solutions

To get more bang for the buck, organizations should look for opportunities to make an impact at the individual and population level. Robison points to hospital readmission risk, which has financial implications for both the hospital (in the form of reduced reimbursements from CMS) and the patient (in cost of care and potential job loss).

Readmission is traditionally calculated based on a specific diagnosis (such as heart failure) without consideration for social needs.

“If machine learning were applied to address the readmission risk in a particular population, then the SDOH risk factors specific to a particular community would be present in a community-specific algorithm,” Robison says. “This would increase the likelihood that the data you have will drive effective decision-making and action.”

3. Be Intentional in Presenting Findings from SDOH Data

Data informaticists often refer to the “five rights” of data, Green says: the right information for the right person at the right time in the right way for the right reason. “Those who struggle to utilize data have typically missed one or more of those steps.”

Organizations that miss steps miss an opportunity to put “accurate, trustworthy and reliable” analytics in the hands of clinicians at the point of care, Racine says. This information must be presented as part of an “invisible footprint,” he adds — analyzed on the back end and integrated with the interface that clinicians are already using, without additional clicks or scrolling to find the relevant information.

MORE FROM HEALTHTECH: Find out how to implement patient portals to improve patient experience.

At the administrative level, this analysis needs to be integrated into the business intelligence and enterprise resource planning tools that drive decision-making, Robison says. This can result in insights such as ZIP codes linked to a disproportionate number of readmissions or external provider partners associated with poor health outcomes.

The size of the data set necessary for analysis and decision-making is often a topic of debate. Green and Racine caution that more isn’t necessarily better, as inaccurate, duplicative and unnecessary data sources can all lead an organization down the wrong path. Robison agrees, saying a data set that is “complete and representative” is necessary for garnering meaningful insight.

“Data needs to be of sufficient depth to enable statistical validation, which leads to it being visualized in ways that stakeholders outside health — such as state or national health departments, governing bodies and the public themselves — can understand,” she says.