A new digital dictation solution hit the market last week with the launch of AutoScribe, a fully-automated, AI-powered medical scribe tool from Ambience. The launch marks the developer’s debut in a dictation and speech recognition market yet to fully convince the healthcare industry of its true benefits.
If it is to succeed, AutoScribe must break down lingering barriers to software adoption, and prove it can simplify workflows, save costs and, vitally for an industry battling staff shortages and burnout, ease pressure on clinicians.
The Signify View
With AI-powered speech recognition at its core, AutoScribe potentially has an immediate advantage over less sophisticated solutions that require more human inputs, assuming it performs as claimed. However, even then it is not the only vendor using AI to address provider pain points in relation to recording encounters. It is a strategy several others, including market leader Nuance, are following.
However, providers largely remain unconvinced of the wider benefits of voice dictation. Or rather, they see the benefits, but these are outweighed by any number of barriers which have held back adoption. These range from resistance among staff and departments to changing workflows to accommodate new technology, to trust issues around the accuracy of the tool, and more.
As well as trying to alter providers’ mindsets (not easy), Ambience is also late in entering a highly competitive arena. It is up against heavyweights like Microsoft-backed Nuance, 3M and other developers that already enjoy tight-knit relationships with EHR vendors. Increasingly, big tech is also attempting to compete.
AutoScribe ticks many boxes for providers, but there are many moving parts here that will ultimately determine its success.
Barriers to Adoption
We have interviewed several health system buyers on this topic recently, and the majority confirm that they are interested in improved solutions and would allocate budget for one that could meet all their needs.
Tellingly, and surprisingly, our discussions also revealed that around 80% of clinicians still type up recorded encounters, and these are manually entered in the EHR.
This is symptomatic of the lukewarm reception voice dictation software has received in healthcare circles to date. Users often complain that software is slow, and cannot always be trusted – staff still feel compelled to manually check for errors and proof-read transcripts, or transcription is outsourced to third parties.
There are other, wider barriers to adoption. One is that most providers still just want speech and voice recognition as a ‘plug and play’ add-on to the EHR. But this can hamper integration with legacy EHR systems (more so where IDNs/ACOs and ICSs have multiple EHR systems in use) or proprietary software.
Then there is ‘cultural’ resistance to change. A clinician might be reluctant to stop using the services of a medical secretary, who do more than transcription and inputting, in favour of the narrow capabilities of technology. Others may be unwilling to adjust established working patterns, especially more senior doctors who may be less prepared than junior colleagues to embrace new technology.
There is also frustration with alternative solutions offered. One health system interviewee called delays between voice entry and order activation ‘unacceptable’ and said they would implement technology only when the time lag is significantly reduced. Other buyers expect the technology to be part of the EHR package they have already paid for.
As a result, voice dictation software will not be a top priority for many providers until the technology can match the performance promise. They see it as nice to have (if it works well), but do not see it as a gamechanger. And in the current economic environment, they would rather invest in other elements of healthcare IT.
So where does this leave Ambience?
It will take heart in the fact that IT procurement in IDNs and physician practices is becoming more centralised. Historically, different health system departments bought dictation solutions to meet their specific workflow needs, and it is not unusual for a health system to have five or six different solutions. It is no secret that some solutions work better with certain workflows – for example, Winscribe and BigHand are good for clinics and writing letters while Dragon Medical One is better for the direct entry of text into EHR (mainly due to its use of AI).
However, providers are often tied into long-term contracts with these existing vendors, and cannot easily switch to a new solution. This is another limitation for Ambience.
But with centralised procurement comes more strategic software purchasing across different settings, and Ambience will argue that AutoScribe aligns well with this trend.
It will also be encouraged by the fact that 80% of those organisations we interviewed would be prepared to invest in new real time speech recognition software (with most also stating that they have budget available) if the solution met performance expectations.
Although AI claims high accuracy when capturing clinical notes, coding support (e g ICD, SNOMED, Medication, Reimbursement, Quality Control etc) is high on any provider’s wish list when considering a voice solution, especially in the US reimbursement environment. Very few vendors provide a strong solution for automated coding alongside automated clinical notes/transcription. However, this is where real value can be added in terms of provider workflows. IT buyers we’ve interviewed indicate that this would be a game changer. If Ambience can use its AI technology to support this, it would be in a much stronger position to challenge the market leaders.
Evolution of Dictation Solutions Toward ‘Fully Ambient Solution’
Nuance DAX comes closest in terms of supporting automated coding, and is one solution that Ambience will now be pitting itself against. It is a highly competitive playing field dominated by Nuance, 3M (which bought M*Modal) and Dolby, and where relationships and EHR integrations are paramount. Voice technology/speech-to-text commands between most EHR systems and digital dictation solutions are still considered ‘clunky’. A vendor offering a dictation solution with a good interface with as many EHR systems as possible is at a distinct competitive advantage.
Nuance has this. Its Nuance DAX tool powers Epic’s Hello Epic ambient listening solution. By contrast, Ambience’s EHR ecosystem lacks the really big-name EHRs, although to be fair its integrations with primary and ambulatory care EHR/EMR vendors AdvancedMD, Dr Chrono, ElationEMR, athenahealth, adracare and eClinical Works are a solid combination.
Winning Hearts and Minds
It is not just the IT buyers within provider organisations that Ambience will need to convince of its credentials. Clinicians hold considerable sway with procurement teams, and their opinions carry a lot of weight during decision-making. Even as health systems move towards centralised IT procurement, clinicians will be influential in advocating single solutions across multiple hospital settings.
There is no doubt that AutoScribe, along with the latest, most sophisticated solutions from other vendors, address some major pain points identified by IT buyers and providers. As health systems seek to reduce costs and plug manpower gaps, AI solutions align well.
Some buyers we speak to say they are ‘excited’ about ambient clinical intelligence, and are realistic that it will take time for voice recognition to be reliable, add value and be worth the investment.
The biggest immediate challenge for Ambience will be persuading providers (and particularly clinicians) that speech and voice recognition tools are now a necessity rather than luxury. Prevailing economic conditions are already changing mentalities in this respect and, along with Nuance DAX, AutoScribe might just be a solution that changes the game.
Ambience nonetheless faces an uphill battle to shake Nuance’s foundations, especially in hospitals where the greater emphasis is on EHR integration. Perhaps it may find more fertile ground, initially, in primary care where providers have more flexibility to choose their real time dictation solution based on their clinical workflows. Costs may be prohibitive for some of these providers, however, but in a market that is only going to get more competitive, this might offer the path of least resistance for now.