Accuracy of clinicians in predicting site and type of lesion as well as outcome in horses with colic.

Abstract

OBJECTIVE: To assess the ability of clinicians to predict the site and type of lesion as well as outcome in horses with colic. DESIGN: Prospective case study. SAMPLE POPULATION: 139 horses admitted for evaluation of signs of colic. PROCEDURE: Six interns and residents examined horses with colic and predicted the segment of intestine that was affected, the type of lesion, and whether the horse would survive to discharge. Accuracy of prediction of site and type of lesion and survival prediction was compared between the first and second halves of the year, using chi 2 analysis and 95% confidence intervals on sensitivity and specificity. chi 2 Analysis was used to assess accuracy between predicted site and type of lesion and intraoperative or necropsy findings and to assess accuracy between predicted survival and actual outcome. RESULTS: Significant association existed between predicted segment of affected intestine or type of lesion and intraoperative findings (P < 0.05). There was a significant association between predicted survival and outcome (P < 0.001). Accuracy of survival prediction improved significantly (P = 0.002) during the year. CLINICAL IMPLICATIONS: Clinicians can accurately predict horses with colic that will survive surgery on the basis of clinical impressions. The ability to predict those horses that will survive improves with training.