Intercenter variation in initial management of children with Crohn's disease

Background: variation in care is a ubiquitous feature of medical practice and may lead to significant differences in health care costs, quality, and outcomes. We undertook this study to determine the extent of intercenter variation in the initial management of children newly diagnosed with Crohn's disease. Methods: We analyzed the utilization of 5 classes of medication (immunomodulators, prednisone, antibiotics, 5-aminosalicylates, and infliximab) among 311 children with newly diagnosed Crohn's disease followed at 10 North American pediatric gastroenterology centers. Multivariate logistic regression was used to compare the utilization rate of each class of medication at each of the 10 centers, adjusting for potential confounders including patient age, sex, race, disease severity, and anatomic location of disease. Results: Median utilization of each class of medication was: immunomodulators, 56% (range 29%-97%); prednisone, 78% (range 32%-88%); antibiotics, 29% (range 11%-68%); 5-aminosalicylates, 63.5% (range 18%-92%); and infliximab, 7.5% (range 3%-21%). Each of these treatments showed statistically significant intercenter variation in utilization (P < 0.001 for immunomodulators, prednisone, antibiotics, and 5-ASA; P = 0.02 for infliximab). After adjusting for the demographic and clinical factors listed above, intercenter variation remained significant; however, the low utilization of infliximab precluded multivariate analysis. Conclusions: Widespread intercenter variation in the medical management of newly diagnosed children with Crohn's disease was observed, even after adjusting for possible differences in case mix between institutions. This variation may lead to unintended differences in health care costs and outcomes.