112 Background: A remote Genetic Testing Station (GTS) workflow was implemented at an academic medical institution to expand access to genetic testing for patients with prostate cancer. During a telephone appointment, a genetic counselor assistant collects family history and facilitates genetics education, research consent, and remote sample collection for multi-gene panel testing. We compared testing completion and patient loss from workflow based on race/ethnicity and preferred language to identify disparities. Methods: Metrics were collected prospectively and analyzed retrospectively for patients with metastatic or high-grade prostate cancer referred to genetics between 3/15/2020 – 6/30/2022. Self-reported race, ethnicity, and preferred language were collected by chart review. Testing completion was compared between groups using Fisher’s exact test, with White non-Hispanic (WNH) and Preferred Language English (PLE) cohorts as controls. Odds ratios and 95% confidence intervals were reported. Patient loss at workflow checkpoints (scheduling, consenting, sample collection, and results release) was summarized for each group. Results: 827 eligible patients were identified: 78 (9%) Asian /Pacific Islander (API), 51 (6%) Black non-Hispanic (BNH), 42 (5%) Hispanic, and 625 (76%) WNH. 31 patients reporting other non-Hispanic race were not included in the analysis. 30 patients (4%) self-reported Preferred Language non-English (PLNE) and 797 (96%) PLE. BNH patients were significantly less likely to complete testing compared to WNH patients (OR 0.320, 95%CI: 0.168, 0.632, p<0.001). There was no difference in testing completion in API (OR 0.918, 95%CI: 0.467, 1.944, p=0.797) or Hispanic (OR 0.743, 95%CI: 0.325, 1.918, p=0.466) compared to WNH patients. PLNE were significantly less likely to complete testing (OR 0.393, 95%CI: 0.171, 0.965, p=0.016) compared to PLE patients. Patient loss occurred primarily at consenting and sample collection. 14% of BNH, and 9% of Hispanic patients did not consent, compared to 4% of WNH. 17% of PLNE did not consent compared to 5% of PLE patients. 13% of BNH did not return a sample, compared to 3% of WNH patients. Conclusions: In remote GTS, BNH and PLNE patients were significantly less likely to complete germline testing than WNH and PLE patients respectively. Disparities in patient loss were most pronounced at consenting and sample collection. Measures to mitigate disparities include assisted consenting (with interpreter as needed) and video-assisted or in-clinic sample collection. [Table: see text]