Watson Institute for International and Public Affairs

Bayesian diagnostic for etiological management of pediatric diarrhea

The Humanitarian Innovation Initiative at Brown University has received funding from the Gates Foundation to develop a predictive tool to guide the treatment of pediatric diarrhea. HIĀ² will be working with the University of Utah (project lead) on developing and testing a smartphone decision app that can provide healthcare workers with probability estimates of the etiology of individual cases of diarrhea in children. As part of an external validation of the prediction tool, Dr. Levine will lead and coordinate the pilot studies of the smartphone app at two sites in Africa and South Asia.

Treatment of diarrhea worldwide is commonly empiric, with antibiotic use mostly based on clinical suspicion. In lower- and middle-income countries (LMICs), due to cost constraints, etiological diagnosis is rarely made, and a large number (up to 70%) of patients with acute diarrhea are prescribed antibiotics. However, the majority of cases of diarrhea do not benefit from antibiotic use, and inappropriate use of antimicrobials may lead to toxicity, increased costs of care, and development of resistance. Thus, accurate determination of diarrhea etiology at point of care is important for proper case management. Current etiologic diagnosis of diarrhea relies on microbial detection, and testing for all potential pathogens is neither financially nor logistically feasible in the majority of clinical settings. Thus, methods to assess diarrhea etiology that reduce reliance on laboratory testing is needed. This project will use predictive analytics of clinical and epidemiological parameters, including Bayesian priors, to develop a prediction tool that can help healthcare workers use antibiotics more appropriately.