Three quantitative approaches to the diagnosis of abdominal pain in children: Practical applications of decision theory

Background/Purpose: The authors compared 3 quantitative methods for assisti ng clinicians in the differential diagnosis of abdominal pain in children, where the most common important endpoint is whether the patient has appendi citis. Pretest probability in different age and sex groups were determined to perform Bayesian analysis, binary logistic regression was used to determ ine which variables were statistically significantly likely to contribute t o a diagnosis, and recursive partitioning was used to build decision trees with quantitative endpoints. Methods: The records of all children (1,208) seen at a large urban emergenc y department (ED) with a chief complaint of abdominal pain were immediately reviewed retrospectively (24 to 72 hours after the encounter). Attempts we re made to contact all the patients' families to determine an accurate fina l diagnosis. A total of 1,008 (83%) families were contacted. Data were anal yzed by calculation of the posttest probability, recursive partitioning, an d binary logistic regression. Results: In all groups the most common diagnosis was abdominal pain (ICD-9 Code 789). After this, however, the order of the most common final diagnose s for abdominal pain varied significantly, The entire group had a pretest p robability of appendicitis of 0.06. This varied with age and sex from 0.02 in boys 2 to 5 years old to 0.16 in boys cider than 12 years. In boys age 5 to 12, recursive partitioning and binary logistic regression agreed on gua rding and anorexia as important variables, Guarding and tenderness were imp ortant in girls age 5 to 12. In boys age greater than 12, both agreed on gu arding and anorexia. Using sensitivities and specificities from the literat ure, computed tomography improved the posttest probability for the group fr om .06 to .33; ultrasound improved it from .06 to .48; and barium enema imp roved it from .06 to .58. Conclusions: Knowing the pretest probabilities in a specific population all ows the physician to evaluate the likely diagnoses first. Other quantitativ e methods can help judge how much importance a certain criterion should hav e in the decision making and how much a particular test is likely to influe nce the probability of a correct diagnosis. It now should be possible to ma ke these sophisticated quantitative methods readily available to clinicians via the computer. J Pediatr Surg 36:1375-1380. Copyright (C) 2001 by W.B. Saunders Company.