Introduction: NVAF is estimated to affect between 6.4 and 7.4 million Americans in 2018, and increases the risk of stroke 5-fold. To mitigate this risk, guidelines recommend anticoagulating AF patients unless their stroke risk is very low. Despite these recommendations, 30.0-60.0% of NVAF patients do not receive indicated anticoagulation. To better understand why this may be, we surveyed PCPs and cardiologists nationwide on their attitudes, knowledge and practices toward managing NVAF with warfarin and direct-acting oral anticoagulants (DOACs). Methods: We surveyed 1,000 PCPs and 500 cardiologists selected randomly from a master list of the American Medical Association, using a paper based, anonymous, self-administered, mailed scannable survey. The survey contained questions on key demographics and data concerning attitudes, knowledge and practices related to prescribing DOACs. The surveys went out in the fall/winter of 2017-8 with a $10 incentive gift card. Survey responses were scanned into an Excel database and analyzed using SAS 9.3 (Cary, NC) for descriptive and inferential statistics. Results: Two hundred and forty-nine providers (167 PCPs, 82 cardiologists) participated in the study with a response rate of 18.8% (249/1320). Respondent mean years ±SD of experience since completing residency was 23.2 ± 13.8. Relative to cardiologists, less PCPs use CHADsVASC (36.8% vs. 74.4%) (p < 0.0001); more have never used HAS-BLED, HEMORR2HAGES, or ATRIA (38.5% vs. 9.8%) (p < .0001); more felt that their lack of knowledge/experience with DOACs was a barrier to prescribing the agents (p = 0.005); and more reported that they could use additional education on DOACs (87.0% vs. 47.0%) (p < 0.0001). Overall, cardiologists were more concerned about ischemic stroke outcomes, while PCPs were more concerned with GI bleeding. Cardiologists also felt that clinical trial data were most helpful in choosing the most appropriate DOAC for their patients, while PCPs felt that Real World Data was most useful. Conclusions: Cardiologists were more concerned with ischemic stroke while anticoagulating patients and utilized screening instruments like CHADsVASC in a majority of their patients. PCPs were concerned with GI bleeds when anticoagulating but nearly 40.0% utilized no screening tools to assess bleeding risk. Our findings show that future education about DOACs would be warranted especially with PCPs.