An Algorithm Using Administrative Data to Identify Patient Attachment to a Family Physician.

Background. Commonly self-reported questions in population health surveys, such as "do you have a family physician?", represent one of the best-known sources of information about patients' attachment to family physicians. Is it possible to find a proxy for this information in administrative data? Objective. To identify the type of patient attachment to a family physician using administrative data. Methods. Using physician fee-for-service database and patients enrolment registries (Quebec, Canada, 2008-2010), we developed a step-by-step algorithm including three dimensions of the physician-patient relationship: patient enrolment with a physician, complete annual medical examinations (CME), and concentration of visits to a physician. Results. 68.1% of users were attached to a family physician; for 34.4% of them, attachment was defined by enrolment with a physician, for 31.5%, by CME without enrolment, and, for 34.1%, by concentration of visits to a physician without enrolment or CME. Eight types of patient attachment were described. Conclusion. When compared to findings with survey data, our measure comes out as a solid conceptual framework to identify patient attachment to a family physician in administrative databases. This measure could be of great value for physician/patient-based cohort development and impact assessment of different types of patient attachment on health services utilization.