An examination of the availability and characteristics of social needs data in the electronic health records: A path to social data harmonization and standardization at Johns Hopkins medicine

OBJECTIVES: To analyze social needs documentation across structured electronic sources in Johns Hopkins Health System, describing practices, patient characteristics, and utility for care, research, and population health. 

MATERIALS AND METHODS: Retrospective study of electronic health records (EHR) data, 2016-2023. Social needs domains were extracted from flowsheets, ICD-10/SNOMED codes, the Social Needs Registry, and the Wellness Registry. A 6-step process refined and aggregated flowsheet entries; regression models assessed associations between documentation and demographic/clinical factors. 

RESULTS: Among 1 042 184 patients, overall, 63 595 unique patients (6.1%) had social needs documented in at least one data source. Demographically, patients with documented social needs were more likely to be Black or African American (40.2%) or Hispanic/Latino (10%) and had higher comorbidity (Charlson Comorbidity Index scores of 2.31 vs 0.84) and healthcare utilization (69.5 vs 6.1% hospitalized). Among patients with 3 or more domains of documented social needs, residential instability was the most prevalent, affecting 85.2% of this subgroup. Regression models showed African American race, Hispanic or Latino ethnicity, legal separation, and clinical severity were associated with greater documentation. 

DISCUSSION: Documentation of social needs in structured fields was sparse and inconsistent across sources, reflecting variable workflows and limited integration of structured data. These gaps hinder measurement, intervention, and equity goals. 

CONCLUSION: Given the inconsistencies we observed in social needs documentation across EHR components and need for properly harmonizing extracted values, we identified practices and workflows that enhance usability for clinical care, research, and population health, and to support equity. Standardized screening, staff training, and reporting are always needed to improve data interoperability and quality.