Engaging Primary Care in Cancer Survivorship (EPICS) trial: A prospective trial of primary care-led survivorship care versus oncology-led care on receipt of recommended and non-recommended services in breast cancer patients.

1 Background: There is a growing interest in initiating primary care-led survivorship prior to the traditional 5- or 10-year transition from oncology. However, the impact on receipt of guideline-recommended care is largely unknown, particularly in community-based settings. We tested a primary care physician (PCP)-led survivorship program compared to usual oncologist-led care for patients diagnosed with early-stage breast cancer in an integrated healthcare system. We hypothesized that patients in the PCP arm would have superior receipt of recommended care. Methods: Initially designed as a cluster randomized trial, we transitioned to a non-randomized control group design due to pressures on primary care during the COVID pandemic. 14 clinical sites participated, 4 intervention and 10 control. We identified eligible participants via an EMR-based algorithm. Stage 0-IIb breast cancer patients diagnosed with first primary cancer, last treatment date 5/1/2019-10/30/2022, and within 6-36 months post-treatment were eligible; those who received neoadjuvant therapy or ongoing ovarian suppression were excluded. At intervention sites PCPs received a 4-month survivorship training. Study outcomes were NCCN and USPSTF guideline-recommended cancer surveillance and preventive services extracted from the EMR up to 36 months after enrollment, as well as secondary outcomes of cancer-related hospitalizations and recurrence. A mean prevention index (PI) score was calculated to identify the proportion of service-eligible patients who received the service. Proportions were generated to evaluate secondary outcomes. Matching weights were used to address imbalances. Results: We enrolled 490 intervention patients and 4559 controls. Arms were similar on unweighted demographic and cancer characteristics, other than race: a higher proportion of Black participants enrolled at intervention sites, 29% vs. 7%. Weighted characteristics were not significantly different between arms. Intervention patients had significantly higher mean PI for flu vaccine (74.1 vs. 69.4, p=0.04) and diabetes screening (96.5 vs. 89.6, p <0.001), and similar mean PI for breast cancer surveillance screening (83.6 vs. 83.3, p=0.44) and colorectal cancer screening (85.0 vs. 84.8, p=0.77). Fewer intervention patients received non-recommended tumor marker tests (9.5% vs. 22.1%, p<0.001) and there were similar proportions of non-recommended PET scans (1.6% vs. 1.8%, p=0.66). Hospitalizations (0.8% vs. 0.7%, p=0.50) and recurrence (2.5% vs. 2.2%, p=0.54) were similar between arms. Conclusions: Patients in the PCP-led intervention arm received similar or higher proportions of recommended care and less non-recommended care, and it was feasible to transition patients as early as 6 months post-treatment. Clinical trial information: NCT04745754 .