Data capture and coding for social determinants of health are works in progress, per reports

In the effort to bring social determinants of health (SDoH) more under the purview of healthcare providers, one tricky aspect is establishing a data and coding infrastructure.

Recent reports highlight this challenge. For example, a survey conducted for the American Health Information Management Association (AHIMA) found that although 78% of 2,637 respondents said their organizations are collecting SDoH data, several issues hamper the quality and accuracy of such data.

One is a lack of standardization. Of the respondents whose organizations collect SDoH data, 59% said they focus on capturing information on health insurance coverage and on health behaviors. But fewer than half collect data on housing security, economic stability, food security, transportation and education.

“These findings suggest that these variables might not have structured fields in the organization’s EHR that would facilitate collection,” states the report on the survey, which was conducted by NORC at the University of Chicago.

Another possible reason is a “lack of incentives for collecting [such data] compared to other elements tied to value-based care contract terms, like smoking status.”

Methods for capturing SDoH data

Of respondents who collect SDoH data, 71% use an EHR. The most frequently used screening tool is CMS’s Accountable Health Communities Core Domains, according to the survey, even though that tool was cited in only 19% of responses. Other tools, such as the federal Uniform Data System, came in at under 10%.

The scattered responses “indicate there is no singular preferred screening tool and an overall lack of awareness of available screening tools,” the report states.

The report recommends “a concerted effort to prioritize collection of certain high-priority data elements … to create meaningful alignment across healthcare sites and settings as well as raising awareness as to the type of screening tools currently available.”

There was much greater uniformity in the responses about which coding systems promote SDoH data collection. Respondents who gather SDoH data almost unanimously said they use ICD-10, with substantially smaller shares using CPT codes, SNOMED-CT and LOINC. The ICD-10 code set in 2016 introduced the Z55-65 codes, which can be used to identify SDoH such as education, employment, housing status, and occupational and environmental risks.

In a 2021 analysis of pre-pandemic data, CMS reported that of all Medicare fee-for-service claims in 2019, only 0.11% were Z code claims. That was 21.4% higher than in 2017, however.

One possible issue is that EHRs are not being deployed to optimal effect for SDoH data collection, the AHIMA report notes: “Integration into the EHR remains a prevailing challenge. The fact that the data may be incomplete, unstructured and/or buried in inconsistent fields — such as health concerns, goals, social history, etc. —may limit the integration of this information into the EHR.”

Policymakers can help by establishing “a set of standardized, clinically valid and actionable SDoH data elements for collection,” the report states.

How to spur more widespread reporting

Among additional policy steps to improve the collection and application of SDoH data, according to the AHIMA report, CMS should offer incentives to providers and insurers “to ensure that all stakeholders are working together to meet their community’s needs.” Likewise, the CMS report said the absence of a payment incentive is a barrier to increasing the documentation of Z codes.

A recent analysis published by the Commonwealth Fund examines federal payments as a lever for incorporating SDoH in clinical models. Based on takeaways from a 2022 workshop, the analysis states that increased funding would be required in Medicare and Medicaid.

“Providers know addressing social needs is important and ethical, but it can be a source of burnout if they do not have the needed resources, time or staffing to identify social needs and connect patients to services,” the authors wrote.

Payments should help providers hire care navigators or community health workers who can connect families with support services. Another funding need is for technology to make and track referrals.

Measures of neighborhood health should be used in risk adjustment, thereby ensuring determinants such as poverty and unemployment are factors in payments, the authors wrote. Among ongoing examples, the authors noted that the federal ACO REACH program features the Area Deprivation Index (ADI) in its health equity benchmark adjustment, while recent changes to the Medicare Shared Savings Program include using the ADI to determine advance payments to eligible accountable care organizations.

“Evidence shows addressing social needs can generate savings for providers and payers, particularly in the Medicaid program, which is already under-resourced,” the authors wrote. “However, [workshop] participants stressed that savings should not be the primary goal.”

Instead, the authors said, “Policy targets must include improved health outcomes and equity.”