To evaluate physician compliance with use of a diagnostic algorithm for appendicitis in children. Our secondary objective was to determine the impact of the algorithm on diagnostic accuracy and morbidity.

We conducted a clustered randomized trial in eight centers. A total of 866 patients were included and, depending on the period of randomization at particular centers, 543 patients were managed before the formal institution of the diagnostic algorithm; their diagnostic management was compared to that of the subsequent 323 patients.

There was a 29.1% mean increase in the use of imaging studies included in the algorithm after algorithm set-up, rising from 50.8 to 79.9% (P<0.02). When we used a composite endpoint of “poor results” (grouping patients with incorrect diagnoses and/or post-operative complications), no statistically significant difference was found between the two periods (85/543 (15.6%) before vs. 45/323 (13.9%) after set-up, P=0.5). But when the number of incorrect diagnoses of appendicitis made without the use of the algorithm was compared to that of patients who took advantage of the algorithm, the difference was highly significant (67/332 [20.2%] vs. 63/534 [11.8%], P<0.001), and the rate of unnecessary appendectomy decreased from 11.9 to 5.3% (P<0.01).

Our diagnostic algorithm improved the adherence to good practice for the diagnosis of appendicitis in children, reducing the rates of unnecessary appendectomy and morbidity. This strategy, combining laboratory tests and imaging, should permit pediatric surgeons to adapt their therapeutic approaches to specific cases.