To compare management strategies for pediatric diabetic ketoacidosis (DKA) among physicians with different specialty training. We conducted a mail survey of 1000 randomly selected physicians, including 200 pediatric endocrinologists, 200 general emergency physicians, 200 pediatric emergency physicians, 200 pediatric intensivists, and 200 pediatric chief residents. We posed questions regarding a hypothetical 10-year-old patient with new onset of diabetes mellitus who is approximately 10% dehydrated but alert, with venous pH of 7.1 and serum glucose concentration of 34.7 mmol/L (625 mg/dL). Questions involved the rate of rehydration, content of intravenous fluids, insulin therapy, potassium replacement, use of sodium bicarbonate, and adjustments in therapy for decreasing serum glucose concentration. We compared responses of physicians in each specialty and used multiple regression analysis to adjust for potential confounding variables, including number of years in practice, number of children with DKA seen per month, and practice setting. Five hundred eighty-one physicians (58.1%) completed the survey, with responses demonstrating significant, consistent differences between specialties. Extremes of responses included the following: (1) 59% of endocrinologists vs 11% of general emergency physicians would give an initial fluid bolus of less than 20 mL/kg (odds ratio [OR], 11.7; 95% confidence interval [CI], 5.0-27.7) (P < .001); (2) 83.5% of general emergency physicians vs 42.5% of pediatric intensivists would administer an initial insulin bolus (OR, 4.1; 95% CI, 2.0-8.7) (P < .001); (3) 58.2% of pediatric intensivists vs 9% of general emergency physicians would replace fluids over a period of greater than 24 hours (OR, 14.1; 95% CI, 5.5-37.5) (P < .001); and (4) 54.3% of general emergency physicians vs 7.3% of pediatric intensivists would use potassium chloride alone for potassium replacement (OR, 10.8; 95% CI, 5.0-23.8) (P < .001). All of these differences persisted after adjusting for the potential confounding variables. Substantial differences exist in the management of pediatric DKA among physicians of different specialties, presumably due to differences in specialty training. These differences obscure our ability to evaluate the treatment of DKA and highlight the necessity for further studies comparing the outcomes of different treatment strategies.