Background: Although EGD is the mainstay of high-risk NVUGIH management, pre-EGD care is critical to optimize outcomes. Yet there is uncertainty about what defines “best practice” in pre-EGD management. Three areas, in particular, have been debated: (1) need for nasogastric lavage (NGL), (2) use of gastric motility agents, and (3) optimal pre-EGD disposition. We conducted a national survey to measure current practice in these 3 areas in a group of experts (Ex) and “non-expert” community providers (NEx). Methods: We developed an online survey that included a vignette of a high-risk NVUGIH patient. Respondents viewed a standardized presentation of a high-risk patient (summary: 42 yo on ASA with melena, epigastric pain, orthostatic BP, no stigmata of liver disease, Hgb = 11), and then received management questions guided by branching conditional logic. Upon EGD, respondents viewed an image of a “spurting” vessel in an ulcer. We measured appropriateness of pre-EGD NGL and motility agents using a standard 9-point RAND Appropriateness Scale (RAS) (1-3 = inappropriate, 4-6 = unsure, 7-9 = appropriate), and calculated the RAND Disagreement Index (DI) for each set of ratings. The DI is a validated measure of provider variation (DI≥1.0 = “extreme variation,” <1.0 = acceptable variation). We surveyed a random sample of 360 GIs from the AGA, and 40 internationally recognized NVUGIH experts. Results: 47% responded (N = 188; 25 experts). Both Ex and NEx groups were generally “unsure” about the appropriateness of NGL (Ex RAS = 5.8; NEx RAS = 5.7; p = 0.8). However, both groups were internally conflicted and polarized, and thus exhibited “extreme variation” regarding NGL (Ex & NEx DI = 4.7). Both groups rated metoclopramide as generally inappropriate (Ex RAS = 3.5, DI = 0.7; NEx RAS = 3.3, DI = 0.8), but rated erythromycin as more appropriate (Ex RAS = 5.4; NEx RAS = 4.1; p = 0.02). However, both groups exhibited extreme variation regarding use of erythromycin (Ex DI = 2.3; NEx DI = 1.6). The most common dispo was transfer to GI suite for immediate EGD (31%), followed by ICU (29%), remain in ER for EGD (15%), monitored floor bed (14%), and non-monitored bed (6%). There was no difference in dispo between Ex and NEx. Conclusions: Both experts and non-experts exhibit extreme variation in their opinion about NGL and promotility agents in the pre-EGD evaluation of high-risk NVUGIH. Although most providers maintain high-risk patients in a monitored environment, there is large variation in the site of monitoring. These variations within and between groups indicate that “best practices” in pre-EGD care remain uncertain and should be subjected to further research and guideline development.