Cohort Study of Structured Reporting Compared with Conventional Dictation

PURPOSE: To determine if radiology residents who used a structured reporting system (SRS) produced higher quality reports than residents who used conventional free-text dictation to report cranial magnetic resonance (MR) imaging in patients suspected of having a stroke. MATERIALS AND METHODS: The study was approved by an institutional review board and was HIPAA compliant; informed consent was obtained. This study included residents, with 16 in the control group and 18 in the intervention group. For phase 1, each subject reviewed the same set of 25 brain MR imaging cases and dictated the cases by using free-text conventional dictation. For phase 2, 4 months later, the control group repeated the same process, whereas the intervention group reread the same MR imaging cases by using SRS to create reports. Resident-generated reports were graded for accuracy and completeness by a neuroradiologist on the basis of consensus interpretations and criterion standard diagnoses as established with at least 6 months of clinical follow-up, imaging follow-up, and/or histologic examination where appropriate. Accuracy and completeness scores were analyzed by using a Wilcoxon signed rank test for paired data and a Mann-Whitney U test for nonpaired data. Intervention group residents were surveyed regarding their opinions of SRS. RESULTS: For phase 1 reports, no significant difference in accuracy or completeness scores between control and intervention groups was found. Decreases in accuracy (91.5 to 88.7) and completeness (68.7 to 54.3) scores for phase 2 compared with phase 1 for the intervention group were found; increases in accuracy (91.4 to 92.4) and completeness (67.8 to 71.7) scores for phase 2 compared with phase 1 for the control group were found (all P values < .001). The most common complaints were that the SRS was overly constraining with regard to report content and was time-consuming to use. CONCLUSION: While there are many potential benefits of structuring radiology reports, such changes cannot be assumed to improve report accuracy or completeness. Any SRS should be tested for effect on intrinsic report quality.