Speech recognition technology in radiology: impact on report preparation time

• Nikita D. Kudryavtsev
• Anton V. Vladzymyrskyy

Abstract

Speech recognition technology is becoming increasingly widespread in the Russian healthcare system. However, the medical community still has questions about the effectiveness of using voice input to complete medical documentation. Radiology was one of the first medical specialties where large-scale implementation of speech recognition technology was carried out. The aim of our study was to evaluate the effectiveness of speech recognition technology in the report preparation of different types of radiological studies.

Material and methods. A retrospective study was conducted at the Moscow Reference Center of the Centre of Diagnostics and Telemedicine. In the study 12 912 radiological reports prepared by 67 radiologists from May-November, 2022 were included by simple random sampling. The following study types were included: fluorography, diagnostic mammography, chest CT scan for suspected COVID-19, MRI of the brain with contrast, and CT of the chest, abdomen, and pelvis with contrast. Voice2Med software was used to fill the reports of radiological studies. Intergroup comparison was done by Mann-Whitney U-criterion with the statistical significance level of 0.05.

Results. The average duration of preparation of fluorographic study reports in the keyboard input group was 189.9 sec (0:03:09), in the voice input group 236.2 sec (0:03:56) (p<0.0001), for mammographic studies – 387.1 (0:06:27) and 444.8 sec (0:07:24) (p<0.0001), for radiographic studies – 247.8 (0:04:07) and 189.0 sec (0:03:09) (p<0.0001), for chest CT scan – 379.7 (0:06:19) and 382.7 sec (0:06:22) (p=0.12), for brain MRI – 709.9 (0:11:49) and 559.9 sec (0:09:19) (p<0.0001) and for chest, abdominal, and pelvic CT with contrast – 2714.6 sec (0:45:15) and 1778.4 sec (0:29:38), respectively.

Conclusion. The use of speech recognition technology in the preparation of radiological study reports demonstrated varying efficacy. The lowest efficiency was obtained when preparing reports of screening studies (fluorography and mammography). The highest efficiency was achieved in the preparation of reports of CT and MRI examinations.

Keywords:speech recognition technology; report preparation; timed study; radiology diagnostics department

References

1. Vechorko V.I. Distribution of working time at an outpatient appointment with a local general practitioner with a nurse in a Moscow polyclinic (photochronometric observation). Sotsial’nye aspekty zdorov’ya naseleniya [Social Aspects of Public Health]. 2016; 52 (6): 4. (in Russian)

2. Kaplieva O.V., et al. Timing of the working time of doctors of the children's consultative and diagnostic department. Dal’nevostochniy meditsinskyy zhurnal [Far Eastern Medical Journal]. 2018; (4): 72–6. (in Russian)

3. Ryabchikov I.V., et al. Distribution of working time at the outpatient appointment of a traumatologist-orthopedist. Moskovskiy khirurgicheskiy zhurnal [Moscow Surgical Journal]. 2018; (6): 38–43. (in Russian)

4. Kudryavtsev N.D., et al. Evaluation of the effectiveness of the implementation of speech recognition technology for the preparation of protocols for x-ray studies. Vrach i informatsionnye tekhnologii [Doctor and Information Technologies]. 2020; (51): 40–7. (in Russian)

5. Speech recognition technology helped doctors fill out more than 210,000 protocols for radiological examinations [Electronic resource]. In: Official website of the Mayor of Moscow. 2023. URL: https://www.mos.ru/news/item/118060073/ (date of access January 09, 2023). (in Russian)

6. Khramtsov A.I., Nasyrov R.A., Khramtsova G.F. Application of digital technology in the work of a pathologist: guidelines for learning how to use speech recognition systems. Pediatr. (Sankt Peterburg) [Pediatrician (Saint Petersburg]. 2021; 12 (3): 63–8.(in Russian)

7. Hammana I., et al. Speech recognition in the radiology department: a systematic review. Health Inf Manag. 2015; 44 (2): 4–10.

8. Poder T.G., Fisette J.-F., Déry V. Speech recognition for medical dictation: overview in Quebec and systematic review. J Med Syst. 2018; 42 (5): 89.

9. Team R.C. R: A language and environment for statistical computing [Electronic resource]. Vienna: R Foundation for Statistical Computing, 2013. URL: http://www.r-project.org/ (date of access January 09, 2023).

10. Sinitsyn V.E., Komarova M.A., Mershina E.A. Structured description protocols in radiation diagnostics. Vestnik rentgenoligii i radiologii [Bulletin of Radiology and Radiology]. 2014; (6): 47–52. (in Russian)

11. Ganeshan D., et al. Structured reporting in radiology. Acad Radiol. 2018; 25 (1): 66–73.

12. Morozov S.P., et al. Radiation diagnostics of coronavirus disease (COVID-19): organization, methodology, interpretation of results. Version 2 (17.04. 2020). 2020. (in Russian)

13. Andrianova M.G., Kudryavtsev N.D., Petryaykin A.V. Thesaurus of radiology terms for preparing reports using speech recognition technology. Digital Diagnostics. 2022; 3 (1S): 21–2. (in Russian)

14. Kudryavtsev N.D. Speech recognition technology: results of a survey of radiologists at the Moscow Reference Center for Radiation Diagnostics. ORGZDRAV: novosti, mneniya, obuchenie. Vestnik VShOUZ [ORGZDRAV: News, Opinions, Training. Bulletin of VShOUZ]. 2022; 8 (3): 95–104. (in Russian)

All articles in our journal are distributed under the Creative Commons Attribution 4.0 International License (CC BY 4.0 license)