Assessing social and behavioral data in electronic health records: Availability, Accuracy, and Applicability
LASSER-DISSERTATION-2021.pdf (1.746Mb) (embargoed until: 2022-12-01)
Lasser, Elyse C
MetadataShow full item record
Problem statement: There is an increased appreciation of the importance of social and behavioral determinants of health (SBDH) on health outcomes, but no standards for collecting this information in various clinical data sources. The increased use of electronic health records (EHRs) provides a unique opportunity to understand SBDH and its impact on health. This dissertation aims to understand perspectives of, and assess trends in SBDH data collection, and compare rates of SBDH ICD10 code documentation in an EHR and insurance claims. Method: A qualitative study was undertaken to understand the facilitators and barriers to accessing SBDH information in an EHR. Using data from 2017, a cross-sectional retrospective data analysis was performed in an EHR’s social history table. Logistic regressions were used to calculate odds ratios to identify factors associated with completion rates. The documentation of behavior related ICD10 codes within a linked EHR, and insurance claims was compared with information in the social history section of the EHR. Results: Providers and researchers felt that SBDH data captured in the EHR was inconsistent and unreliable. Health systems should prioritize capturing some SBDH in a consistent manner, but it is unclear which variables to select. Individuals who are black, female, and between the ages of 30-65 are more likely to have their behavior documented in the social history section. With the move to ICD10, a wider range of SBDH information can be coded in a patient’s EHR and claims record. At this study site, the overlap of codes across these two data systems is limited and thus a fuller picture of the patient’s situation can be obtained by merging both sources. Conclusion: It appears that SBDH data collection is not consistent at this site. To improve this, clearer guidelines on how to capture SBDH risk factors are needed. Since there is no widely accepted “gold standard”, information in the EHR and insurance claims vary, which makes it more challenging to effectively understand SBDH factors in order to assess and enhance health outcomes. Improving data collection and data reliability will allow providers and researchers alike to utilize digital data for both patient care and population health.