Background: Although barriers exist to secondary use of primary care electronic medical record (EMR) data, the Alliance for Healthier Communities (the Alliance) in Ontario, Canada has successfully created one of the largest structured primary care EMR datasets in Canada. In 2018, the Alliance and the Canadian Institute for Health Information (CIHI), an organization that provides comparable and actionable data to accelerate improvements in health across Canada, entered into a partnership to share EMR data. In this paper, we describe (i) the processes that enabled the collection of structured EMR data by the Alliance; (ii) how CIHI connected with the Alliance to share data and assess its quality; and, (iii) demonstrate the value of linking structured EMR data to administrative acute care data in illustrating the patient journey through the care continuum, using COPD as a case study. Methods: CIHI and the Alliance entered into a formal data sharing agreement that enabled the sharing of linkable structured EMR data by the Alliance's 75 community health centres (CHCs) with CIHI. By leveraging the Alliance's Business Intelligence Reporting Tool, 3 years of EMR data containing patient-level clinical data were shared with CIHI. CIHI assessed the EMR data quality using its Data Source Assessment Tool to ensure fitness for analytical use. By linking the patient level EMR records with hospital records (CIHI's discharge abstract database (DAD) and the national ambulatory care reporting system (NACRS)), we examined aspects of COPD patient management in primary care and followed their journey through the health care continuum, including follow-up in primary care after hospital discharge. Results: Alliance EMR data representing approximately 570,000 patients and 8.5 million primary care encounters between April 1, 2015 and March 31, 2018 were shared with CIHI. A data quality assessment, centered on completeness and concordance, confirmed that the data was fit for analytical purposes. Overall, 13,023 enrolled primary care patients were identified as having COPD, representing an overall crude prevalence of 8.7%. The average age of COPD patients was 64 years and equally affected males and females. Patients were most likely to have completed high school education or equivalent, speak English, live alone, and have a household income less than $15,000. They most commonly had between 10-19 primary care encounters a year with a range of providers where they most commonly sought services for health advice/ instructions, to discuss their treatment plans and for medication renewals. By linking the EMR data to CIHI's NACRS and DAD, we found that 74.1% of COPD patients had at least one ED visit and that 34.4% of COPD patients had at least one acute care hospitalization during the study period. Further, 16.2% of ED visits resulted in an acute care hospital admission. Of those hospitalized, the majority of COPD patients were discharged home (81.6%) and received timely follow-up in primary care (81.0% within 30 days). Conclusion: Structured and linkable EMR data provides opportunities to examine the patient journey through the care continuum in an innovative way. Using structured EMR data from the Alliance, linked with CIHI's NACRS and DAD databases, we were able to generate a cohort of patients with COPD, explore the complexities of their primary care encounters and follow them through the continuum of care, namely emergency department visits and hospitalizations. It is hoped that the partnership between the Alliance and CIHI will help drive future efforts on addressing the gap in comparable EMR data in Canada, and internationally.
