Department of Radiology, University of North Carolina Hospitals, Chapel Hill, 27599, USA.
AJR Am J Roentgenol. 2010 Jul;195(1):194-7. doi: 10.2214/AJR.09.3169.
Previous studies have documented reductions in turnaround time after implementation of voice recognition software in the generation of radiology reports. Our preliminary observations suggested that improvement in report turnaround time varies among users. The purpose of this study was to analyze the effect of work habits and caseload on such variations.
Data were collected for 9 months before and after the implementation of voice recognition after a 6-month training period. Thirty faculty members were ranked according to their report turnaround time before and after implementation of voice recognition and according to their percentage reduction in report turnaround time. The report turnaround times before and after implementation of voice recognition for faculty were compared with the number of verified reports and work habit type.
The average report turnaround time for the department before implementation of voice recognition was 28 hours. After implementation of voice recognition, the average turnaround time was 12.7 hours, and the volume of verified reports increased 5% between the two study periods. The improvement in report turnaround time for individual faculty members ranged from -33% to +93%, and the rank order did not change significantly (Spearman coefficient, 0.58; p < 0.05). Faculty members' ranks in report turnaround time did not correlate significantly with volume rank before and after implementation of voice recognition (Spearman coefficients, 0.341 and 0.346; p > 0.05). Faculty members who had type 1 work habits, that is, reviewed, revised, and finalized reports at the time of image review, benefited the most from use of voice recognition.
Use of voice recognition software decreased report turnaround time for the department and for 28 of 30 individual faculty members. Improvement in report turnaround time does not correlate with workload but does correlate with work habits, suggesting human behavior may play a role in determining the outcome of adopting a productivity-enhancing technology.
先前的研究记录表明，在生成放射学报告时实施语音识别软件后，周转时间会减少。我们的初步观察表明，报告周转时间的改善因用户而异。本研究的目的是分析工作习惯和工作量对这种变化的影响。
在经过 6 个月的培训后，在实施语音识别后的 9 个月内收集数据。根据实施语音识别前后的报告周转时间以及报告周转时间的减少百分比，对 30 名教职员工进行排名。在实施语音识别前后，比较教职员工的报告周转时间与已验证报告的数量和工作习惯类型。
在实施语音识别之前，该部门的平均报告周转时间为 28 小时。实施语音识别后，平均周转时间为 12.7 小时，两个研究期间已验证报告的数量增加了 5%。个别教职员工的报告周转时间改善范围为-33%至+93%，排名没有明显变化（Spearman 系数，0.58；p<0.05）。教职员工在实施语音识别前后的报告周转时间排名与工作量排名没有显著相关性（Spearman 系数分别为 0.341 和 0.346；p>0.05）。具有 1 型工作习惯的教职员工，即在图像审查时审查、修改和最终确定报告，从语音识别的使用中受益最大。
使用语音识别软件减少了部门和 30 名教职员工中的 28 名的报告周转时间。报告周转时间的改善与工作量无关，但与工作习惯相关，这表明人类行为可能在确定采用提高生产力的技术的结果方面发挥作用。
