Accuracy of interim radiology resident CT reports given during on call hours in a competency based radiology training program

Date of Award


Document Type


Degree Name

Master of Medicine (MMed)


Imaging and Diagnostic Radiology (East Africa)


Background: In many academic hospitals globally, radiology residents provide preliminary interpretations of CT studies performed outside of regular working hours.

As part of a competence based education (CBE) system this is with a view to continue providing the necessary radiology consultation while at the same time exposing the radiology resident to much needed experience in making independent decisions and in developing professional skills.Realtime cases are the foundation of any competency based educational program.

Discrepancies however do occur between the residents’ interim reports generated after normal working hours and the final reports released to the clinician. This may lead to adverse clinical outcomes including worsened morbidity, repeated imaging or delayed diagnosis.

Objectives: The primary objective was to prospectively define the extent and factors contributing to discrepancy in out of hours CT reporting. Secondary objectives were to identify the most common discrepancies that occur during after-hours reporting, risk factors for such discrepancies and to establish ways of improving patient management in a competency based training institution.

Design: Prospective cohort study of patients undergoing CT scans at Aga Khan University Hospital (AKUH)-Nairobi, carried out over a consecutive duration of six months from October 2011 to end of March 2012.

Sampling method: Consecutive sampling of all interim CT scan reports by residents after normal working hours was done.

Data analysis: The data collected were analyzed using Stats Program Software, SPSS version 16.The proportion of major and minor discrepancies were determined within a 95% confidence interval.Categorical variables of interest including type of radiological examination, confidence level of the resident at the time of reporting, level of residency and time of reporting were plotted and displayed in tables, pie charts and stacked columns against the major and minor discrepancy rates. Potential associations between major discrepancy rates and numerical variables including adequacy of clinical information and level of confidence of the resident during reporting were tested using Chi square.

Results: The major discrepancy rates were 4.5% with an overall discrepancy rate (both major and minor discrepancies) of 11.7% i.e. 34 out of 291 reports. Majority of these of these were errors of observation (19 out of 34).There were five (5) false positives and one (1) error of interpretation or cognition.

These errors led to change in management in fifteen cases, increased patient morbidity in ten cases, required additional imaging in three cases, necessitated or resulted in extended hospital stay in four patients and did not potentially alter the patients’ clinical outcome in ten of the cases. No patient mortality was reported to have occurred as a direct cause of these errors.

There was no statistically significant association between discrepancy rates and the level of residency, time of reporting, type of examination, adequacy of history provided and level of confidence of the resident at the time of reporting.

Conclusion: Overall the study fully addressed its objectives. The point estimate of major discrepancy of 4.5% was within the estimated range used in the calculation of the sample size (7+/- 3%), and is comparable to rates described in literature. Lack of a statistically significant correlation between the discrepancy rate and the secondary outcomes including the level of residency, time of reporting, level of confidence of the resident at the time of reporting or the adequacy of the history given is a pointer to no evidence of effect rather than evidence of no effect of these variables on the discrepancy rate.

This document is available in the relevant AKU library