CT in Adults: Systematic Review and Meta-Analysis of Interpretation Discrepancy Rates

Abstract

Potentially useful reference ranges were identified in the subgroup analyses on the basis of body region scanned at adult CT. However, considerable heterogeneity that is only partially explained by subgroup analysis signifies that further research is necessary--particularly regarding the question of blinding of the reference radiologist.

The accuracy of interpretation of emergency abdominal CT in adult patients who present with non-traumatic abdominal pain: results of a UK national audit

“…8,12,14,17,26e31 A meta-analysis looking at discrepancy rates in adult CT (all types and including elective and emergency) demonstrated overall no significant differences in rates of discrepancy between a registrar and more senior radiologists, with a pooled discrepancy rate for abdominopelvic CT of 2.6%. 32 A recent study looking at abdominal CT in surgical patients found a 14% rate (146/1071 reports) of clinically important management changes following double/expert reading of initial CT reports. 33 There is a relative paucity of published literature pertaining to discrepancy in outsourced, off-site radiology, a large series published in 2005 looking at a radiological group practice, reported a discrepancy rate of 2.1% for CT of the abdomen/ pelvis.…”

Interobserver Agreement in the Interpretation of Outpatient Head CT Scans in an Academic Neuroradiology Practice

“…[1][2][3] The content of these interventions may not be evident, however. In addition, the manner in which the error, discrepancy, and disagreement should be handled both in theory and in clinical practice is evolving.…”

“…However, there is no consensus on a standard method or protocol for evaluating errors and discrepancies in imaging reports, and rates published in the literature differ widely. [1][2][3][10][11][12][13][14] Multiple variations in study parameters, including sampling sources, methods, imaging modalities, specialties, categories, interpreter training levels, and degrees of blinding, may have contributed to this wide spectrum. 2,3,9 Recently, CT and MR imaging reports of the head, neck, and spine were re-read by staff neuroradiologists, and a 2% clinically significant discrepancy rate was found, an excellent result compared with the 3%-6% radiologic error rates published in general radiology practices.…”

“…[1][2][3][10][11][12][13][14] Multiple variations in study parameters, including sampling sources, methods, imaging modalities, specialties, categories, interpreter training levels, and degrees of blinding, may have contributed to this wide spectrum. 2,3,9 Recently, CT and MR imaging reports of the head, neck, and spine were re-read by staff neuroradiologists, and a 2% clinically significant discrepancy rate was found, an excellent result compared with the 3%-6% radiologic error rates published in general radiology practices. 3,15,16 To anyone who has studied reliability or precision of diagnostic imaging tests, such levels of disagreement between interpretations may appear unbelievably low.…”

“…Performance can be measured; deviations and outliers can be identified, and appropriate measures can then be taken to improve performance. [1][2][3] A different vocabulary is used when discrepancies are examined from a scientific point of view. In the typical absence of a criterion standard of "truth," the uncertainty is a reality that must be admitted and taken into account when using imaging reports for clinical decisions.…”