Brian H Willis1, Shyamaly D Sur. 1. Emergency Department, Horton Hospital, Oxford Radcliffe Hospitals NHS Trust, Oxon, UK. b.h.willis@doctors.org.uk
Abstract
OBJECTIVE: To assess the accuracy of Senior House Officers at interpreting plain X-rays following their triage by radiographers in an emergency department. METHOD: We collected 2593 patients' records by systematic sampling of all those seen by emergency physicians between January 2002 and April 2002 (ca 10 000 patients) in a UK emergency department. The variables recorded included evidence of X-ray investigations and, when present, the Senior House Officer's diagnosis, the presence (abnormal) or absence of a radiographers red dot and the reference standard diagnosis. A separate category of uncertain (inconclusive) was applied to the Senior House Officer and reference standard diagnosis where appropriate. Diagnostic performance was measured by likelihood ratios with associated pre-test and post-test probabilities. RESULTS: Including the uncertain category as abnormal gave the following results: there were 967 X-rays and those with a red dot had a probability of an abnormality of 80%. Although a further opinion of abnormal by a Senior House Officer increased this probability to 89% when they overrode the red dot opinion of the radiographer, it was incorrect in 26% of cases. CONCLUSION: Currently, the Senior House Officer contributes to the red dot system by improving on the radiographer in rates of diagnosis of both abnormal and normal X-rays. Further reductions in error rates, however, are unlikely to be achieved until there is a change to the existing system. This may ultimately involve removing some of the responsibility of X-ray interpretation from the Senior House Officer. Any future research should consider the methodological issues highlighted by this study.
OBJECTIVE: To assess the accuracy of Senior House Officers at interpreting plain X-rays following their triage by radiographers in an emergency department. METHOD: We collected 2593 patients' records by systematic sampling of all those seen by emergency physicians between January 2002 and April 2002 (ca 10 000 patients) in a UK emergency department. The variables recorded included evidence of X-ray investigations and, when present, the Senior House Officer's diagnosis, the presence (abnormal) or absence of a radiographers red dot and the reference standard diagnosis. A separate category of uncertain (inconclusive) was applied to the Senior House Officer and reference standard diagnosis where appropriate. Diagnostic performance was measured by likelihood ratios with associated pre-test and post-test probabilities. RESULTS: Including the uncertain category as abnormal gave the following results: there were 967 X-rays and those with a red dot had a probability of an abnormality of 80%. Although a further opinion of abnormal by a Senior House Officer increased this probability to 89% when they overrode the red dot opinion of the radiographer, it was incorrect in 26% of cases. CONCLUSION: Currently, the Senior House Officer contributes to the red dot system by improving on the radiographer in rates of diagnosis of both abnormal and normal X-rays. Further reductions in error rates, however, are unlikely to be achieved until there is a change to the existing system. This may ultimately involve removing some of the responsibility of X-ray interpretation from the Senior House Officer. Any future research should consider the methodological issues highlighted by this study.