Primary Health Care Reform: Who joins a Family Medicine Group?

Reorganization of primary health care is being actively pursued and new models of primary health care delivery are being developed in the U.S. and in several Canadian provinces. In Quebec, Family Medicine Groups (FMGs) were created in 2002 in order to provide enhanced access and better coordination of care through a team based approach to primary care. Previous research on new models of primary health care has often failed to evaluate their effects within a causal inference framework, and little attention has been paid to the type of physicians and patients that voluntarily join them. Understanding who is attracted to new models is not only important to adjust for selection bias, but it may affect future reforms by helping to elucidate what would happen if FMGs were implemented on a population level. This thesis attempts to understand the voluntary selection of patients and physicians into Family Medicine Groups in Quebec, Canada. A longitudinal administrative dataset of vulnerable patients (elderly or chronically ill) from the Régie de l'assurance maladie du Québec (RAMQ) has been divided between FMG and non-FMG users, and includes information on demographic characteristics, chronic illnesses and ambulatory and tertiary health service use before the advent of FMGs. Physicians of these patients are characterized by their FMG status, demographics, and practice and patient characteristics before FMGs are in place. Multivariate regression is used to identify key predictors of joining a FMG among both patients and physicians. Lastly, comparable physician and patient populations are created using propensity scores in order to set up the evaluation of health outcomes, utilization of services and costs in the years after joining a FMG. The distribution of propensity scores and their ability to balance key covariates after different matching and weighting techniques was investigated. Results of the analysis reveal that geographic location, socio-economic status, visits in an ambulatory setting, emergency room visits, hospitalizations and having a usual provider of care are all factors which affect the probability of a patient joining a FMG. Specifically, residents of remote regions, low socio-economic status and those who use emergency rooms and hospitals more often are more likely to be enrolled, whereas patients that use ambulatory services and have a usual provider of care are less likely to be enrolled. Similarly, it is shown that factors that affect a physician's likelihood of joining a FMG include time since graduation, geographic region and revenue from traditional fee-for-service vs. other sources. Younger physicians and those who practice in a local community centre (CLSC) and short term/acute inpatient hospital care (CHSCD) are more likely to participate. Propensity scores were able to balance the pre-treatment differences, and this finding is robust across different mechanisms of adjusting for the propensity score. Overall, it was shown that participation in a FMG is not a random process and any further research on the effect of FMGs, or any other type of primary health care reform, should consider this. Accounting for the type of patients that join different models, by using propensity score analysis for example, will be critical to forming evidence based policy recommendations. Particular consideration for geographic location, patients' morbidity, socio-economic status, health service use, as well as physicians' age and experience working in other settings is needed.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.