Does the use of speech recognition influence the number of reports completed daily by the radiologists at King Saud Medical City (KSMC) in Riyadh

Abstract

Speech recognition (SR) software was introduced in the radiology department of the King Saud Medical City (KSMC) in 2016 to reduce the turnaround time and administrative costs of reports while improving their accuracy. The aim of the current study is to find and determine the effectiveness of SR software on the number of daily reports completed by radiologists. The study adopted a quantitative, retrospective, quasi-experimental design and used piece-wise linear regression to measure changes in the average number of reports over the time frame of the study (June 2015 – December 2015 before SR integration and June 2016 – December 2016 after SR integration) and an independent samples t-test to detect the significance of any relationship between the two variables. The data analysed were figures of the number of daily reports produced by 34 radiologists. The results showed that the mean number of reports produced before SR was 708.47 while the mean number produced after was 1128.76. A strong positive correlation of .834 was found between the use of SR software and the number of reports completed. The relationship was found to be highly significant p(0.013), so the use of speech recognition software is significantly related to the completion of daily reports by the radiologists. As technology advances, this potential must be matched with continuous research and investigation on how to improve SR going forward. Further research exploring the statistical relationships discovered in this research must be carried out to determine explanations for this phenomenon, and this would usefully be qualitative research that gathers rich, narrative data from radiologists about their use of SR.