Social and behavioral determinants of health (SBDOH) data collection at time of breast cancer diagnosis: Results of a stepped-wedge, cluster randomized trial.

92 Background: Screening for unmet social needs can identify barriers to cancer care, but facilitating screen completion is challenging. We report findings from NCT06019988, which compared SBDOH data collection tools and modalities in breast cancer patients. Methods: Females ≥18y newly diagnosed with breast cancer at our institution were referred from 3/13/24-3/14/25. Patients with previously diagnosed disease, non-English primary language, or new patient visit (NPV) scheduled<48h after referral were excluded. Patients were randomized to 1 of 3 validated screening tools: Accountable Health Communities Health-Related Social Needs tool, Health Leads Screening ToolKit, or National Comprehensive Cancer Network Distress Thermometer and Problem List (NCCN DT). Enrolled patients first received their assigned tool via the MyPennMedicine (MPM) EHR patient portal. After 48h, non-completers were randomized to receive the same tool by text or phone. Patients with incomplete screens by time of NPV were given a final opportunity to complete the tool at NPV in person. Chi-square, ANOVA, and Kruskal-Wallis tests were used to assess differences across groups. Results: Of 559 referred patients, 187 were enrolled and randomized, and 6 were lost to follow up, yielding a final cohort of 181 (median age 59y). Tool assignment was balanced; 134 (74%) completed a screen prior to NPV. More than half (51.4%, n=93) completed a screen via MPM, with higher rates in White vs Black and privately vs government- and uninsured patients (Table). Graduate degree holders had higher rates of MPM completion (67.3%, n=37) vs those with ≤high school education (HSEd, 40.7%, n=11, p<0.001). Among non-MPM completers, more were randomized to phone (55.7%, n=49) than text (43.2%, n=38), but text had a higher completion rate (55.3%, n=21) than phone (42.9%, n=21, both p<0.001). Education was the only factor with statistically significant differences in completion by screening tool: NCCN DT had the highest rate among patients with ≤HSEd (n=13/25, 52%, p=0.03). Overall, 29.8% (n=54) of patients reported a social need or requested social work referral. Conclusions: In our pragmatic trial, patients who were Black, had non-private insurance, or ≤HSEd were least likely to complete SBDOH screens via EHR portal. NCCN DT had the highest completion among those with ≤HSEd. Nearly 1/3 of patients reported a social need, underscoring the importance of early, multi-modal SBDOH screening. Clinical trial information: NCT06019988 . N (%) MPM CompletedN (%) MPM Not CompletedN (%) p-value Race/ethnicity NH White, 105 (58) 65 (61.9) 40 (38.1) <0.001 NH Black, 55 (30.4) 16 (29.1) 39 (70.9) Hispanic, 9 (5) 7 (77.8) 2 (22.2) Other, 12, (6.6) 5 (41.7) 7 (58.3) Insurance Private, 122 (67.4) 72 (59.0) 50 (41.0) 0.02 Medicare, 46 (25.4) 18 (39.1) 28 (60.9) Medicaid, 11 (6.1) 3 (27.3) 8 (72.7) Government, 1 (0.6) 0 (0) 1 (100) Uninsured, 1 (0.6) 0 (0) 1 (100) NH=Non-Hispanic.