Background: Germline genetic testing is now recommended for all pancreas cancer patients (pts); identification of a BRCA1/2, ATM or PALB2 mutation can impact upfront and/or second line therapy. These mutations have been found in approximately 4-11% of prior reports of pancreatic cancer populations. Previous reports have found that genetic testing is done in a minority of patients, especially in private practice. This project assesses how two interventions on physician behavior can impact the frequency of germline genetic testing in pancreas cancer patients in a multi-office private practice setting. Methods: Retrospectively, we used algorithms in Flatiron’s OncoEMR to identify baseline germline genetic testing frequency in all pancreas cancer patients. In phase 1, we used an EP to communicate with doctors immediately before upcoming appts for their pancreas cancer pts. In phase 2, advanced practice providers (APPs) educated pancreas cancer patients about germline genetic testing and ordered the test ahead of the patients’ visits. Our primary endpoint was to determine the percentage of pancreas cancer patients who had germline genetic testing; our secondary endpoint was to determine frequency of BRCA/PALB2/ATM mutations. Results: Retrospectively, 82 pancreas cancer pts were identified from 6/1/2020 - 1/1/2021. 32 pts (39%) had germline genetic testing. Prospectively, in phase 1, 35 pts with pancreas cancer and upcoming appointments were identified between 6/28/2021 - 8/6/2021. 21 (60%) had previously had germline testing. Of 14 patients who had not had testing, the EP prompted additional testing of 7 (50%) pts. Of the total 35 pts in phase 1, 2 (5.7%) were found to have mutations. In phase II, 47 of 74 pts (63%) had previously had germline testing, and APP intervention resulted in 19 of 27 patients (70%) receiving germline testing, with 5 (6.7%) found to have BRCA1/2, PALB2 or ATM mutations of the 74 patients. Conclusions: Our results reveal that while APP education and ordering of germline genetic testing is slightly more effective than involving the doctors in the process, further improvement is needed. Utilizing the EMR when a diagnosis code is entered and then automating genetic counseling and ordering for these patients will be needed. Lastly, a team based approach with navigators involved in rounding will also improve testing frequency even with more automated testing. While our % of patients with an actionable mutation is low, given the options it opens for these patients, universal testing is still warranted.