Clinical care site data integration reveals heterogeneity in EHR phenotyping and healthcare utilization patterns

The study examined how data from different clinical care sites in electronic health records (EHR) vary and how this affects genetic research. Researchers developed two methods: ClinicScan identifies where diagnoses are documented, and ClinicWAS identifies specialties associated with risk factors. The analysis included 64,983,257 patient encounters across 2,544 clinical care sites mapped to 57 specialties. For depression, most diagnoses (30.3%) were recorded in primary care and 19.8% in psychiatry. When a psychiatry encounter was required, the association between genetic risk and depression diagnosis strengthened. For coronary heart disease, 19 care sites with higher occurrence of high-risk patients were identified, including 5 catheterization laboratories. Patients with high genetic risk underwent catheterization 3.1 to 4.6 years earlier than those with normal risk.