پذیرش تشخیص گفتار توسط پزشکان: بررسی انتظارات، تجارب و نفوذ اجتماعی

The present study has surveyed physician views and attitudes before and after the introduction of speech technology as a front end to an electronic medical record. At the hospital where the survey was made, speech technology recently (2006–2007) replaced traditional dictation and subsequent secretarial transcription for all physicians in clinical departments. The aim of the survey was (i) to identify how attitudes and perceptions among physicians affected the acceptance and success of the speech-recognition system and the new work procedures associated with it; and (ii) to assess the degree to which physicians’ attitudes and expectations to the use of speech technology changed after actually using it. The survey was based on two questionnaires—one administered when the physicians were about to begin training with the speech-recognition system and another, asking similar questions, when they had had some experience with the system. The survey data were supplemented with performance data from the speech-recognition system. The results show that the surveyed physicians tended to report a more negative view of the system after having used it for some months than before. When judging the system retrospectively, physicians are approximately evenly divided between those who think it was a good idea to introduce speech recognition (33%), those who think it was not (31%) and those who are neutral (36%). In particular, the physicians felt that they spent much more time producing medical records than before, including time correcting the speech recognition, and that the overall quality of records had declined. Nevertheless, workflow improvements and the possibility to access the records immediately after dictation were almost unanimously appreciated. Physicians’ affinity with the system seems to be quite dependent on their perception of the associated new work procedures.

Speech recognition has been refined and become more robust in recent years (Lai et al., 2008). The gradual maturation of the technology has been accompanied by adoptions of the technology in the medical domain, where it is used to enter comments into the electronic medical record (EMR), thus replacing the standard way of entering notes by physician dictation and subsequent transcription by medical secretaries or a dedicated service (Zafar et al., 1999; Al-Aynati and Chorneyko, 2003). At the same time as the technology has matured, speech recognition has been developed and implemented for languages spoken by much “smaller” populations, such as Danish (5.4 million speakers). Vejle and Give Hospital, Denmark, has been one of the first hospitals to introduce speech recognition for all major specialties and departments. Having run a successful project on speech recognition in its radiology department since 2000, this regional hospital (349 beds, and 217 000 outpatients in 2006) began to implement plans for having all physicians in clinical departments use speech recognition to input physician notes and instructions into the EMR. The speech-recognition system—software based on Philips Speech Magic, adapted to Danish and deployed by Max Manus A/S—was rolled out in all clinical departments in 2005–2006, and has about 240 physician users as of 2007. The main purpose of introducing speech recognition across all departments was to ensure a quicker completion of medical record entry and to achieve a higher quality of patient records. The old transcription service was known to sometimes produce backlogs of dictation tapes waiting to be transcribed by medical secretaries, or transcriptions waiting to be checked and approved by physicians. Additionally, an expected consequence was to allow secretaries, who would no longer need to spend time on transcriptions, to take over other duties. It was hoped that the quality of medical records would be enhanced, since physicians would now be going to check and revise their written (speech recognized) record immediately while their intentions were still fresh in memory. While little is known so far about the impacts of speech recognition on the various stages of the writing process and on the quality of outcome (Honeycutt, 2003), the above-mentioned goals fully match criteria such as those reported by Mönnich and Wetter (2000). The present study had two related objectives: First, to identify physicians’ attitudes and expectations about speech recognition that might explain their subsequent level of satisfaction with actual use of the technology. Second, to assess possible changes between prior expectations to and subsequent experience with the technology as a replacement for the traditional mode of producing medical records. 2. Related work Work about the acceptance of speech recognition falls into two main areas: speech recognition and technology acceptance. Studies of speech recognition have predominantly been devoted to recognition of spoken English. However, recognition rates of systems that recognize English are not necessarily transferable to a speech-recognition system for Danish.