Start Your Search
Mervat Abdelhady
Author of
-
OS21 - Adoption & Engagement (ID 27)
- Event: e-Health 2017 Virtual Meeting
- Type: Oral Session
- Track: Clinical
- Presentations: 1
- Coordinates: 6/06/2017, 01:00 PM - 02:00 PM, Room 206B
-
OS21.04 - Speech Recognition Versus Traditional Transcription Model for Documentation in EMR (ID 166)
Mervat Abdelhady, Medical Informatics Department, Humber River Hospital; Toronto/CA
- Abstract
Purpose/Objectives: Speech recognition (SR) is a valid new technology that provides an alternative to back-end transcription services. SR facilitates dictation and editing of clinical notes in real time. Front-end dictation using SR was proposed as a quality initiative in our organization in which physician documentation is done mostly through back-end dictation. Manual transcription not only is expensive but also results in delays of hours and days that substantially affect the continuity and quality of care provided. Our objectives were to evaluate the quality of documentation and the cost effectiveness of speech recognition versus back-end dictation. In using speech recognition, we aim to align with our digital vision for better integration and interoperability of care via: Ensuring accurate, timely completion of reports by utilizing a dictation tool that has speech to text capability and integrates with our EHR. Reducing annual dictation costs. Optimizing physician productivity and satisfaction through providing physicians with a tool that easily integrates with their workflow.
Methodology/Approach: We proceeded with a 1-year pilot project which involved 115 physicians, implementing a cloud-based SR for documentation in EMR. Analysis of the quality of documentation was performed on a sample of physicians (n=38); in which 2 files per physician were reviewed pre, mid and post implementation of front-end SR using a physician documentation quality instrument (PDQI) tool to evaluate the quality of clinical notes. Physician education and one-on-one training was followed by post-live support throughout implementation. We addressed change management and adoption delivery challenges, and monitored the KPI of the software system by soliciting physician feedback and undergoing post live surveys of project success from the physician perspective. Furthermore, we addressed, documented and analyzed troubleshooting issues.
Finding/Results: Based on our initial findings there is an overall benefit in relation to documentation speed when using SR which resulted in elimination of time lag to report availability. The accuracy and completion of clinical notes improved. With the use of SR, the quality and timeliness of documentation were evident, however, the main drawback reported by physicians was the time spent on editing documents. The software performance in regard to accuracy, response time and author accent or speech impediment were favorable. Preliminary findings show a reduction in cost as a result of adoption of front-end dictation versus back-end dictation. Barriers to adoption included; prior experience with other dictation software version that had slower response time and higher error rate; fear of increase in the workload and lack of familiarity with the new process.
Conclusion/Implication/Recommendations: We conclude that SR is a potentially valuable and effective tool that improved the quality of clinical documentation and reduced the cost incurred through manual transcription. The implementation of SR in clinical practice significantly changed the clinical processes undertaken by our organization and these changes along with their follow-on effects may need to be further analyzed. The advantage to adoption of this technology must be weighted in regard to the efficiency, timeliness and the impact it has on patient safety as well as the added value it gives to the quality and continuity of care.
140 Character Summary: Speech recognition software versus back-end dictation improves quality, timeliness and the utilization of resources.
