KLAS: Speech Recognition For Clinical Documentation On The Rise

Front-End Speech Recognition

Interviewed organizations said that ambient speech recognition technology has improved over the last few years in terms of accuracy and ability to detect accents to the point that it is on par with the precision of transcription services.

Another selling point for speech recognition technology is the variety of platform types. Often, providers can use their mobile phones to capture dictation and review notes.

The biggest challenge with front-end speech recognition may be its price. Respondents noted nickel-and-diming on top of what is already considered an expensive system.

Transcription Services

With speech recognition on the rise, provider organizations are beginning to move away from transcription services and in-person scribes. However, respondents said transcription would continue playing a limited but important role in EHR documentation.

Transcriptionists and scribes have been especially useful as an aid for physicians who are hesitant to adopt speech recognition technology and as a backfill for organizations with large documentation backlogs.

Clients are generally satisfied with their transcription vendor’s ability to meet or exceed contractual turnaround times while maintaining accuracy. When customers find errors or inconsistencies in transcriptions, they noted that their firms are responsive to correction requests.

The transcription market has recently gone through a lot of changes. AQuity Solutions spun off from MModal after being acquired by 3M, and DeliverHealth acquired Nuance’s transcription services arm. As a result, respondents said that their vendor relationships suffered somewhat. Client satisfaction with executive involvement, proactive service, and communication has declined slightly in this market.

Virtual Scribes

Most respondents (94 percent) feel that their virtual scribe positively or highly impacts their documentation time. Additionally, most customers mention that the virtual scribe completes documentation by the end of the day.

The report found that provider organizations are most likely to recommend using a virtual scribe to help achieve same-day closures when clinicians have difficulty closing their notes on time.

Respondents said that the main issue with virtual scribes is the inconsistent quality of transcribers. The best scribe-clinician relationships come through consistency. Respondents said it is difficult to maintain the same quality output with replacement scribes.

Computer-Assisted Physician Documentation (CAPD)

Almost all respondents said CAPD enabled more efficient code capture, boosted revenue, and reduced documentation time. Still, physician adoption is a challenge for several reasons.

For instance, some Nuance clients said the CAPD system does not fit within a physician’s typical workflow. Additionally, 3M customers often reported that recommendations fire too often, so they are disregarded or turned off.

Clinical Documentation Integrity (CDI)

Clients noted that clinical documentation integrity (CDI) software often provides a solid ROI through increased billing capture, more efficient workflows, and improved prioritization and guidance. There is wide adoption of CDI in the inpatient space and significant growth in the use of CDI in outpatient care.

However, provider organizations said they felt that the development of CDI software has stagnated.

An exception may be Iodine Software, which customers see as a highly innovative firm. Customers of 3M, Dolbey, and Optum said they have not seen any timely improvements. While firms have upgraded technology incrementally, they have not updated the core software components, clients said. Provider organizations did not anticipate many major changes in the future.

Ambient Speech Technology

While a few companies either employ or are developing ambient speech technology, it is still a newer market that KLAS has not yet fully measured. However, it is a rapidly growing area, and KLAS is looking to measure these solutions more fully in the future.

Ambient speech recognition technology uses automatic speech recognition (ASR) and natural language processing (NLP) to document patient-provider conversations in the EHR.

Early KLAS data from Nuance Dragon Ambient eXperience (DAX) customers revealed high satisfaction rates, noting the system’s ability to alleviate clinician burnout.

However, respondents noted that they sometimes feel like DAX is a work in progress. Early adopters often report that the cost is expensive and that the technology needs more work to enhance EHR integration and reduce note return time.