Back to 2026 Abstracts
Primary Care Physician Specialty and Prostate Cancer Screening: A Propensity Score-Matched Cohort Study
Eric J. Macdonald, MD1, Rumel Mahmood, PhD2, Thaise Uzeda, MD1, Benjamin Green, MD1, Andrew Williams, MD1, Michelle Shabo, MD1, Benjamin Allar, MD1, Stephanie Hastings, MD3, Boris Gershman, MD1, Gezer Ortega, MD4, Heidi Rayala, MD, PhD1.
1Beth Israel Deaconess Medical Center, Boston, MA, USA, 2OHSU, Portland, OR, USA, 3Cambridge Health Alliance, Cambridge, MA, USA, 4Mass General Brigham, Boston, MA, USA.
Background: In the United States, prostate cancer screening (PCS) guidelines recommend shared decision-making, yet practice patterns vary. Prior research comparing screening rates between Family Medicine (FM) and Internal Medicine (IM)-trained primary care physicians (PCP) is limited to survey-based studies. This study compared PCS rates by PCP specialty training using electronic medical record data.
Methods: This retrospective cohort study examined PCS among men aged 45-70 years at Cambridge Health Alliance, an urban academic-affiliated healthcare system serving a socioeconomically and racially diverse population in the greater Boston, Massachusetts area from December 2022 through November 2024. The primary outcome was PSA testing during the study period. Multivariable logistic regression assessed the association between PCP specialty training and PCS, adjusting for patient demographics, clinical risk factors, and healthcare utilization. Given baseline differences in patient characteristics in both PCP groups, propensity score matching was performed as a sensitivity analysis.
Results: Among 12,900 eligible patients (51.2% FM, 48.8% IM), the overall PCS rate was 42.1%. Patients of IM-trained physicians had higher unadjusted screening rates than those of FM-trained physicians (45.8% vs 38.7%; p<0.001). After multivariable adjustment, IM-trained PCPs remained associated with significantly higher odds of PCS (OR 1.28; 95% CI 1.19-1.38). In the propensity score matched cohort, the difference was driven by the interaction between race and IM PCPs (OR: 1.32, 95% CI 1.03-1.70, p = 0.028).
Conclusions: PCS rates differed significantly by PCP specialty training, independent of patient characteristics. These findings suggest that training background may contribute to variation in shared decision-making practices, with implications for understanding variation in preventive care delivery.
Back to 2026 Abstracts
