Effect of computed tomography on patient management and costs in children with suspected appendicitis.

OBJECTIVE: Children evaluated in the emergency department for possible appendicitis are often admitted for observation, despite the widespread availability of accurate diagnostic studies, particularly computed tomography (CT). We sought to establish effective and efficient strategies for using CT to diagnose and manage children with possible appendicitis.
DESIGN: Retrospective chart review and decision analysis. Setting. Emergency department of a large, urban tertiary care pediatric teaching hospital. PATIENTS: All patients admitted from January 1996 to August 1997 for suspected appendicitis. METHOD OF ANALYSIS: Three modeled strategies were empirically applied to the retrospective cohort of patients admitted for observation. Outcomes and costs under the modeled strategies were compared with those under current practice. The three strategies were: 1) to obtain CT scans on all patients and discharge those with normal findings; 2) to obtain CT scans and admit all patients; 3) to selectively obtain CT scans on those patients with a peripheral white blood cell count >10 000/mm(3) (10 x 10(9)/L) and admit all. The sensitivity and specificity of CT for diagnosing appendicitis were determined empirically from the data. A sensitivity analysis was performed. MAIN OUTCOME MEASURES: The number of preoperative inpatient observation days, total hospital costs, and the rates of both missed appendicitis and negative laparotomies.
RESULTS: Of 609 patients hospitalized for possible appendicitis, 287 went directly to the operating room and 14 patients had known perforation and abscess. Three hundred eight children were observed and comprised the study cohort. Of the cohort, 112 (36.4%) underwent appendectomy and 26 (23.2%) of these had a normal appendix at pathology. Three patients were discharged from the hospital after observation and were subsequently readmitted with appendicitis (missed appendicitis). Among the 75 patients who had CT performed, the sensitivity and specificity of CT were both 97%. Under the current practice strategy, the cohort collectively accumulated 487 inpatient observation days and incurred a per patient cost of $5831. All three CT strategies would have reduced the total number of inpatient observation days, operations, negative laparotomies, as well as the per patient cost. The strategy of obtaining CT scans on all patients and then admitting them had the lowest rate of missed appendicitis. The additional cost of preventing each case of missed appendicitis under this strategy compared with the strategy of obtaining CT scans and sending home those with negative findings was $150,304. Even at the lowest reported sensitivity and specificity of CT in the literature, the ordering of the three strategies remained constant and continued to reduce total cost per patient.
CONCLUSION: Compared with current practice, diagnostic strategies using CT could reduce costs and improve diagnosis, management, and outcomes for children with appendicitis.