Implementation of clinic-based data capture in a real-world epilepsy clinic

Abstract

Capturing structured clinical data during routine care remains an elusive but potentially high-impact goal. Learning Health Systems (LHS) are intended to capture data and provide evidence-based guidance, improving the care clinicians provide. Our epilepsy center implemented software to enable patient survey delivery and discrete data entry by clinicians during clinic visits. Here we evaluate the initial 18 months of this process. There were 2779 visits by 1664 patients seeing 12 clinicians. Manual entry of any data by clinicians occurred in 59.5 % of visits, and entry of all data domains occurred in 48.9 %, dominated by 3 clinicians who completed 74 % of these visits. Clinician manual entry rate was slightly higher among White, non-Hispanic and male patient visits, those with Medicaid or Self Pay as primary insurance, and among follow up visits (vs new patient). Patient surveys were all completed in 16.2 % and at least one survey in 54.7 % of visits. Higher rate of patient survey completion was associated with English speakers, in person visits (vs virtual), new patient visits, and private insurance. There was no association of survey completion with race, ethnicity or diagnosis. Overall, clinical data capture is achievable when there is strong clinician engagement and optimization of clinic workflows to enable survey completion. Care should be taken to ensure implementation captures representative data and avoids systematically marginalizing people facing social barriers to healthcare.