Association between primary care physicians' practice models and referral rates to specialists: A sex-based cross-sectional study

Abstract

Referrals from primary care physicians (PCPs) to specialists are a key function of the primary care system, enabling access to secondary and tertiary health care services. Since the early 2000s, Ontario has implemented substantial primary care practice reforms, however, PCP referral patterns have not been examined since reforms were implemented. We conducted a cross-sectional study in Ontario analyzing PCPs' referral patterns to specialists from January 1 to December 31, 2019. Data from physician administrative and Ontario Health Insurance Plan (OHIP) billing databases were linked for 9,301 PCPs practicing comprehensive primary care with 11.8 million patients. We calculated referral rates per physician and built a multivariable Poisson regression model stratified by physician sex, recognizing that female and male PCPs practice primary care differently, to examine the association between PCP's referral rates and their practice model. Subgroup analyses were conducted for medical, surgical, diagnostics and General Practitioner (GP) focused practice specialties. Overall, PCPs in fee-for-service practice models (females: 0.72, 95% CI 0.71-0.72, males 0.71 95% CI 0.71-0.72) and Family Health Groups (females: 0.90, 95% CI 0.90-0.91, males 0.85 95% CI 0.84-0.85) had lower adjusted relative referral rates compared to those in Family Health Teams (FHTs); a finding that was consistent across medical and surgical specialties. Younger, part-time PCPs, those practicing in urban areas, those with larger roster sizes and those affiliated with a large practice group showed higher adjusted referral rates. Female PCPs tended to be younger (average age 47.2 years vs. 54.1 years for males; SMD=0.56), work part-time (32.1% vs. 17.9% for males; SMD=0.33), had a smaller patient roster (average 1,097.8 rostered patients vs. 1,442.1 for males; SMD=0.44), and had higher unadjusted referral rates to specialists compared to male PCPs (32.9 vs. 29.9 per 100 rostered patients). PCPs' referral patterns in Ontario vary by practice model and PCP's sex. Future changes to primary care practices should account for their effects on referral volumes to specialists.

Fig 1. Inclusion and exclusion criteria.