Making Sense Of Speech Recognition Technology

Making sense of speech recognition technology options we have today is challenging.

Speech recognition technology is finding its way into the mainstream HIM environment. This technology will have a positive effect on TAT as more and more healthcare facilities and MTSOs employ it. The next question on the HIM survey was “Are you using SRT?”. Among 87 responses, 56 percent of respondents stated that they do not use SRT; 23 percent stated that they use back-end speech recognition (BESR); and 21 percent stated that they use front-end speech recognition (FESR). 39% used SRT in radiology only and the combination of HIM and radiology.

Performance challenges in light of sweeping promises by several providers has shifted the conversation on understanding the true benefits and issues associated SRT adoption. This is exacerbated by the lack of objective performance measurement that has impinged the development of standards and metrics which could be of value to the whole industry.
While the technology is shaping medical workflow and the ecology of physician narrative capturing in a dramatic way, many risks and hurdles must be understood and managed. Some of the salient considerations include
– understanding the role of human intelligence (such as medical transcriptionists [MTs]);
– advantages and disadvantages back-end and front-end SRT;
the relationship between SRT and the electronic health record (EHR).
– documentation quality and turnaround time;
– training, staffing, support, and other operational functions;
– cost / value assessment for providers and patients
risk management; and
Best practices for data capture

While SRT will never serve as a panacea to poor dictation, it remains the linchpin of the medical workflow. From the moment of interaction, several considerations are at play from the timeliness of report availability, reimbursement, and downstream value and impact of the captured data: so much hinges on quality, accuracy and completeness.

Quality assurance practices understood in the realm of pure dictation remain relevant and must be well understood and implemented. New practices must be put in place such as quality scoring and audit reviews of accuracy where all contributors are involved. Transcription editors must also hone new skills which were not in play in traditional transcription assignments such as recognizing contextual errors and ensuring an efficacious hand off to EHRs. These new roles may well be suited for the seasoned transcriptionist, and the person eager to make full use of technological tools put at their disposal.

While speech recognition technology has patently changed the manner in which documentation and narrative is processed in healthcare, the role of medical transcriptionists may well be forever changed as the new environment calls for more critical thinking and greater contextual filtering. MTO must adjust rapidly to these new practices, and if so will entertain new professional opportunities in an ever changing health documentation landscape.