Implementation Challenges for Practice Use of EHR and Patient Reported Data for Clinical Decision-Making

CAT HALLIWELL
PRO-EHR Program with University of Colorado Department of Family Medicine

The Colorado Patient Reported Outcomes and Electronic Health Record Program (PRO-EHR) partnered with a rural primary care practice in Colorado to use EHR and Quality of Life data to help vulnerable, high-risk patients during the COVID-19 pandemic. We report on significant challenges encountered. Our intent is to build a replicable, automated clinical decision-making tool for high-risk individuals. Challenges included time, EHR extraction and work-flow issues. We offer suggestions to limit the challenges described.

Mid-Valley Family Practice (MVFP) is a rural private practice in Basalt, CO with robust IT support. Patient population is 22% Medicaid, 27% Medicare, 17% 65+, 30% Hispanic, 70% white. It has 1 physician, 1 nurse practitioner, 1 physician assistant, 3 nurses, 2 medical assistants, 3 front desk staff, 1 IT Manager and 2 administrative staff.

We implemented three phases: start up, QOL administration, and EHR data extraction. Patients completed the QGEN and QDIS (Ware 2019) QOL surveys on iPads, generating a convenience sample of 250 patients. Survey responses were immediately EHR-accessible during the visit. For each respondent, EHR data was extracted based on the C-19 Index (Closed Loop). Data analysis is ongoing.

The project took 16 months from initiation to completion of data extraction. Participation was not paid for. It included 58 meetings. Role specific time requirements: Clinic director = 105 hrs, Practice manager = 27 hrs, Front desk staff = 19 hrs, and IT manager = 204 hrs, for a total of 355 hrs. Front desk staff helped with Spanish translation, and administered an additional 28 surveys in Spanish after collecting 244 in English. IT manager time included 24 hours for QOL survey incorporation, access, extraction, and quality control, and 180 hours for VI query creation, testing, and data extraction. The practice contacted an estimated 500 patients to yield 272 surveys. 12 of 40 VI variables were available in the EHR with over 80% complete data. Challenges included the amount of time needed, technical requirements, and extra attention to ensure inclusion of patients with additional barriers (e.g. language). Upcoming efforts now move to usability in addition to further analysis.

Implementation of practice-based EHR and patient reported measures requires a significant investment in clinic, provider, and staff time, and is feasible for other practices if the time commitment is available. Thus far we have accommodated the challenges, including the demands on clinic time by the ongoing COVID pandemic, novel demands on the EHR, and new workflows for collection of QOL data. Factors to consider include flexible time, designated EHR personnel, and inclusion of patients with barriers to care. Next critical usability issues include using QOL and EHR data to create a risk score that informs point of care clinical decision-making pathways for additional primary care sites.