Motyer R E, Liddy S, Torreggiani W C, Buckley O
Department of Radiology, The Adelaide and Meath Hospital, Dublin 24, Ireland.
Ir J Med Sci. 2016 Nov;185(4):921-927. doi: 10.1007/s11845-016-1507-6. Epub 2016 Oct 1.
Voice recognition (VR) dictation of radiology reports has become the mainstay of reporting in many institutions worldwide. Despite benefit, such software is not without limitations, and transcription errors have been widely reported.
Evaluate the frequency and nature of non-clinical transcription error using VR dictation software.
Retrospective audit of 378 finalised radiology reports. Errors were counted and categorised by significance, error type and sub-type. Data regarding imaging modality, report length and dictation time was collected.
67 (17.72 %) reports contained ≥1 errors, with 7 (1.85 %) containing 'significant' and 9 (2.38 %) containing 'very significant' errors. A total of 90 errors were identified from the 378 reports analysed, with 74 (82.22 %) classified as 'insignificant', 7 (7.78 %) as 'significant', 9 (10 %) as 'very significant'. 68 (75.56 %) errors were 'spelling and grammar', 20 (22.22 %) 'missense' and 2 (2.22 %) 'nonsense'. 'Punctuation' error was most common sub-type, accounting for 27 errors (30 %). Complex imaging modalities had higher error rates per report and sentence. Computed tomography contained 0.040 errors per sentence compared to plain film with 0.030. Longer reports had a higher error rate, with reports >25 sentences containing an average of 1.23 errors per report compared to 0-5 sentences containing 0.09.
These findings highlight the limitations of VR dictation software. While most error was deemed insignificant, there were occurrences of error with potential to alter report interpretation and patient management. Longer reports and reports on more complex imaging had higher error rates and this should be taken into account by the reporting radiologist.
放射学报告的语音识别（VR）听写已成为全球许多机构报告的主要方式。尽管有好处，但此类软件并非没有局限性，转录错误已被广泛报道。
使用VR听写软件评估非临床转录错误的频率和性质。
对378份最终放射学报告进行回顾性审核。根据错误的重要性、错误类型和子类型对错误进行计数和分类。收集有关成像方式、报告长度和听写时间的数据。
67份（17.72%）报告包含≥1个错误，其中7份（1.85%）包含“重大”错误，9份（2.38%）包含“非常重大”错误。在分析的378份报告中总共发现了90个错误，其中74个（82.22%）被归类为“不重大”，7个（7.78%）为“重大”，9个（10%）为“非常重大”。68个（75.56%）错误为“拼写和语法”错误，20个（22.22%）为“错义”错误，2个（2.22%）为“无义”错误。“标点”错误是最常见的子类型，占27个错误（30%）。复杂成像方式的每份报告和每句话的错误率更高。计算机断层扫描每句话包含0.040个错误，而平片为0.030个。报告越长错误率越高，超过25句话的报告每份平均包含1.23个错误，而0至5句话的报告每份包含0.09个错误。
这些发现突出了VR听写软件的局限性。虽然大多数错误被认为不重大，但仍有一些错误可能会改变报告解读和患者管理。较长的报告和关于更复杂成像的报告错误率更高，报告放射科医生应予以考虑。
