Observer Variation in Clinical Diagnosis - A Computer-Aided Assessment of its Magnitude and Importance in 552 Patients with Abdominal Pain

Abstract

This report describes an evaluation of »observer variation« in history taking and examination of patients with abdominal pain. After an initial survey in which the degree of observer variation amongst the present authors fully confirmed previous rather gloomy forecasts, a system of »agreed definitions« was produced, and further studies showed a rapid and considerable fall in the degree of observer variation between the data recorded by the same authors. Finally, experience with a computer-based diagnostic system using the same system of agreed definitions showed the maximum diagnostic error rate due to faulty acquisition of data to be low (4.7°/o in a series of 552 cases). It is suggested as a result of these studies that — at least in respect of abdominal pain — errors in data acquisition by the clinician need not be the prime cause of faulty diagnoses.