Evaluating the Impact of Clinical Decision Tools in Pediatric Acute Gastroenteritis: A Population-based Cohort Study

Academic Emergency Medicine, Volume 23, Issue 5, pages 599–609, May 2016

Acute gastroenteritis (AGE) is a leading cause of pediatric emergency department (ED) visits. Despite evidence-based guidelines, variation in adherence exists. Clinical decision tools can enhance evidence-based care, but little is known about their use and effectiveness in pediatric AGE. This study sought to determine if the following tools—1) pathways/order sets, 2) medical directives for oral rehydration therapy (ORT) or ondansetron, and 3) printed discharge instructions—are associated with AGE admission and ED revisits.

This was a retrospective population-based cohort study of all children 3 months–18 years with an AGE ED visit in Ontario, Canada, from 2008 to 2010, using linked survey and health administrative databases. Logistic regression models associating clinical decision tools (CDTs) with hospitalizations and revisits controlling for hospital and patient characteristics were employed.

Of the 57,921 patient visits during the study period, there were 2,401 hospitalizations (4.2%). A total of 55,520 patients were discharged from the ED, with 2,378 (4.3%) experiencing a 72-­hour return visit. In adjusted models, none of the tools were significantly associated with admission. Medical directive for ORT was associated with lower return visit rates (adjusted odds ratio [aOR] = 0.86, 95% confidence interval [CI] = 0.79–0.94] and printed discharge instructions with higher return visits (aOR = 1.33, 95% CI = 1.08–1.65); pathways/order sets and medical directives for ondansetron had no association.


Admissions in children with AGE are not associated with the presence of CDTs. While ORT medical directives are associated with lower ED revisits, printed discharge instructions have the opposite effect. The simple presence/absence of decision support tools does not guarantee improved clinical outcomes.

Personal comment: Clearly, there is a need to review all of this data and find those factors that are related to a significant reduction in AGE admissions and return visits. This is a classic opportunity for machine learning algorithms to be tested.