Publication
Article
Author(s):
What is possible now, what is next, and what it means for physicians
In the early days of electronic health record (EHR) adoption, digital voice applications were more or less relegated to transcribing the exact information spoken into them. Physicians were tied to desktop computers and limited by error-prone solutions that lacked the ability to learn from their own shortcomings.
Fast forward a decade or so, and smart voice applications like Siri or Alexa have become commonplace in the cars, homes and office settings of today’s consumers. We have truly entered the world of conversational artificial intelligence (AI), and it is showing great promise for health care delivery.
According to a recent study from Voicebot.ai, use cases for voice assistants in health care expanded from just under 20 million U.S. adults as of mid-2019 to a whopping 54.5 million in 2021. The most advanced voice assistants can now process information with 99% (or greater) accuracy, even if a physician’s native language is not English.
And the good news is that the industry has just scratched the surface in terms of voice potential. By all indications, the best is yet to come. Over the next few months and years, we can expect to see new voice capabilities roll out, which will give physicians and health systems once-unimaginable advantages as they navigate big challenges.
Gaining momentum
Today’s breed of advanced voice assistants are much closer to being actual “assistants” than were previous offerings. Physicians can talk to them directly, dispatch a slew of commands or simply speak their observations and recommendations, and voice applications are equipped with the functionality to respond in kind. As such, today’s physicians can much more easily obtain information such as detailed drug-interaction lists or a patient’s most recent health history and/or emergency department utilization — while on the go, at home or from any location.
These advances have translated into increased time savings for clinicians throughout the country: A 2021-22 pilot study by the American Academy of Family Physicians showed that the use of a voice-enabled AI assistant achieved a 72% decrease in median documentation time per month, for an average savings of 3.3 hours per week.
Considering that clinical burnout has accelerated over the past few years since the onset of the pandemic, this is important. Many physicians, such as Dr. Elizabeth Goff of Virginia, have said that the time savings generated by voice allows them to stay employed full time, instead of cutting back their hours. The ability to use a voice assistant on a mobile device to create notes that get synced to the EHR saves hours that would have otherwise been spent typing.
When physicians gain back significant time savings, it allows them to repurpose their time into patient care activities, or simply close their laptops an hour earlier at the end of their days.
The importance of these advantages is game changing. Yet these capabilities have also stoked physicians’ appetites for more.
What’s next — and why it matters
The next iteration of the more evolved, intelligent voice assistant is materializing as we speak to address the burdens of the future, starting with provider shortages in every care setting. An August 2022 poll of physicians by STAT and the Medical Group Management Association revealed 4 in 10 medical practices had a physician resign or retire early in the past year due to burnout.
Meanwhile, the pressure to do more with less has gone into overdrive. As the American Hospital Association noted in a recent report, the pandemic took “a significant toll on hospitals and health systems and placed enormous strain on the nation’s health care workforce,” and hospital labor expenses per patient through 2021 were 19.1% higher than pre-pandemic levels in 2019.
For all these reasons and more, health leaders are hoping that the next wave of voice solutions will be even smarter than today’s generation. To achieve this vision, voice technology must go beyond simple “command and response” functions and become more intuitive, with the ability to listen and decipher critical information. Many of these so-called ambient capabilities — where physicians can speak naturally during a conversation and have their “assistant” filter through the droves of data to find the “gold” in spoken narratives — are not possible yet without humans working behind the scenes to decipher information. However, the presence of behind-the-scenes humans raises privacy concerns. It also can significantly increase costs.
Fortunately, we are getting closer to having market-ready, ambient voice solutions that can exude a “real” humanlike presence and produce human-assistant-like deliverables, but don’t require back-end support.
Having a voice assistant that can function at this level means a physician won’t have to dispatch commands such as “Can you show me a list of allergies?” They could simply say, aloud, “Can you pay attention to what I’m telling you?” and know that the voice assistant will document a visit without over-documenting unnecessary, unimportant information (e.g., a patient’s references to their recent holiday trip to the Caribbean).
Preparing for the future
Health leaders are faced with difficult decisions every day, including where and how much money to allocate toward the adoption of new technologies. With each version of voice AI improving and getting better at distilling meaningful information from every interaction, voice remains one of the wisest health care investments.
Although we have not quite reached the holy grail of what smart voice assistants can do, the time savings they yield can help organizations address real operational pain points. Health care organizations that utilize the best of voice today also have the advantage of becoming acclimated to a solution that will rapidly evolve in the near future.
Belwadi Srikanth is vice president of product and design at Suki, where he oversees and executes Suki’s aggressive product roadmap for its AI-powered voice solutions in health care. Suki Assistant makes use of generative AI to listen to physician-patient interactions in real time and auto generate clinical notes — without requiring human intervention in the background.