Observer Performance in Computed Tomography Head Reporting

Area of Science:

• Medical Imaging
• Radiography
• Diagnostic Accuracy

Background:

• Accredited academic programs are training radiographers in clinical reporting of computed tomography (CT) head examinations.
• Auditing the performance of these radiographers is crucial for ensuring high standards of diagnostic reporting.

Purpose of the Study:

• To audit the reporting accuracy of radiographers (n=6) who completed an accredited academic program for CT head examinations.
• To assess the effectiveness of academic and clinical workload case banks in evaluating radiographer performance.

Main Methods:

• Retrospective and prospective audit of academic and clinical case banks, including diverse normal and abnormal CT head cases.
• Analysis of true positive/negative and false positive/negative fractions against a reference standard.
• Interobserver variability assessed using Cohen's kappa and ANOVA with Tukey's post-hoc tests.

Main Results:

• Mean accuracy score of 90.7% (95% CI, 88.3%-93.0%) across 3,008 reports.
• Mean sensitivity of 86.9% (95% CI, 85.8%-88.2%) and specificity of 94% (95% CI, 89.6%-98.3%).
• Common errors included subtle findings like herniation, lacunar infarctions, subtle fractures (false negatives), and involutional changes (false positives).

Conclusions:

• Appropriately trained radiographers can report CT head examinations to a high standard.
• Utilizing diverse academic and clinical case banks provides an accurate measure of radiographer performance.
• This approach supports the development of robust competency assessments for radiographers in CT head reporting.