Decoding Social Determinants of Health

Good health outcomes rest not only on quality clinical care, but also on nonclinical factors that influence a patient’s health. These factors, known as “social determinants of health,” are a frame around the conditions a clinician identifies as the picture of a patient’s health.

Throughout a decade, SDOH have been increasingly recognized as a factor in delivery of, and payment for, patient-centered care. During this time, clinicians and community stakeholders have worked to understand the scope and influence of SDOH. For example, Google searches of “social determinants of health” increased 163 percent between 2008 and 2018. However, as the nation’s health care system incorporated SDOH in care improvement strategies, clear definitions and a reliable measurement system with widely available data have been elusive. This is beginning to change.

Key Findings:

• The use of diagnostic coding for SDOH carries beneficial implications for hospitals, including potential future reimbursement, using SDOH in risk adjustment for incentive-based payments, identifying high-risk patients and informing community health needs assessments.
• Comparing rates of SDOH coding to poverty at county and ZIP code levels suggests inconsistent application of SDOH coding among Missouri hospitals — two safety net hospitals accounted for 9 percent of total discharges and 32.4 percent of SDOH-coded claims during the first 30 months of ICD-10.
• Despite suggestions of inconsistency, the frequency of hospital patients diagnosed with social complexity has steadily increased in Missouri since the conversion to ICD-10 in October 2015.
• Compared to all hospital patients, individuals diagnosed with social complexity in Missouri have significantly higher rates of hospital utilization and social, behavioral and clinical risk factors.
• The most common ICD-10 SDOH code used in Missouri is homelessness (Z590), with more than 34,000 diagnoses occurring for 17,068 unique patients between October 2015 and March 2018.
• Even with potential inconsistencies in the use of SDOH codes, bivariate and multivariate testing suggests they have significant predictive ability in health outcomes modeling, such as hospital super-utilization.

READ THE POLICY BRIEF