June 19, 2020
The Key to AI-Assisted Documentation? It’s Not Keywords
By Justin Wilson, Vice President of Communications

A few years ago I visited my dermatologist for a routine check-up. Like many providers, he felt the weight of administrative responsibilities that were overtaking his time with patients. He had turned to a dictation program to tackle his documentation. Rather than try to recall the visit later and document after the visit, he chose to take moments during our encounter to dictate into the EHR. We would chat for a minute, and then he would turn to the computer and input commands. Several times the inadequate software would incorrectly transcribe a more complex term so he would stop, say “delete, delete, delete,” and once again try “seborrheic keratosis.” Again, the system wouldn’t understand, he would say, “delete, delete, delete,” and then “seborrheic keratosis.” It was excruciating. Would it be rude to pull out my phone and play Angry Birds?

Perhaps the dictation companies recommend documenting after the patient encounter. But either way, it’s a sequential process that puts a large burden on the provider to dictate at some point, taking away valuable time with patients. And while it may be (arguably) somewhat more efficient than completely unassisted documentation, it was still significant admin time that could have otherwise been spent with more patients. Worst of all, those “delete and repeat” moments were painfully inefficient.

Fast forward a bit, and today we have new technology on the scene – AI-assisted digital documentation assistants. A little over a year ago, Becker’s Health IT surveyed eight chief medical information officers about how they thought voice recognition could impact healthcare. While hopeful, Dr. Ryan Walsh, CMIO and Vice President the University of Texas Health Science Center, asked, “can my brain process dictate text and computer commands at the same time?”

Dr. Patrick Woodard, associate chief medical officer for IT at Renown Health, is optimistic that technology has the potential to improve provider satisfaction and improve provider-patient relationships, but colloquially notes, “a good 10 percent of the time, my bathroom lights turn on when I want my bedroom lights to.” While unreliable voice recognition is fine in his house, it’s not okay when it impacts patient care.

Perhaps Dr. Walsh summarized it best, noting the voice recognition software is exciting, but “it needs to be well thought out and done right or we’ll wind up with the next wonder-gadget my colleagues yell at me about.”

The next era of documentation support is here, and it shows great promise, but not all platforms are equal. Much like the dictation of years past, many ambient documentation platforms create a technological barrier for providers. They require providers learn new workflows, and require keyword triggers (i.e. “hey Device, document this…”).

This is the same bad holiday gift, just in new wrapping. Much like that poor dictation experience, I can’t imagine sitting in the exam room while my doctor repeats keyword commands.

At SoundLines, we believe technology should seamlessly fit in your workflow, and not be another barrier to meaningful patient care. Speke has been created with simplicity in mind. Speke does not require keywords or triggers. In just a few seconds, the provider selects the current patient, hits one button, sets their phone aside and carries on a normal conversation with their patient. The encounter is captured and transcribed, and a scribe on the backend supports the documentation directly into the EHR. The provider simply reviews and signs off.

As part of the HealthChannels family, this has been our mission for nearly 20 years. We’re committed to workforce and technology solutions that help care team reduce administrative burdens and put them back at the bedside so they can once again practice traditional, personalized medicine.

We’ve heard firsthand how this seamless, hassle-free approach can make a meaningful difference. Dr. Fred Williams, a gastroenterologist, previously used a dictation service tied to his EHR, and noted how the frequent transcription errors were very frustrating. With Speke, he doesn’t have to worry about the system understanding commands and prompts, but knows his documentation is complete, and “the information is put forth exactly how he [wants it] to be.”

“It’s a wonderful feeling,” says Dr. Williams, “to walk out of the office in the afternoon knowing that all of my notes are done and I won’t be spending the evening dictating.”

That’s the promise of tomorrow’s technology, today. Providers don’t need to wait decades for effective and efficient AI-assisted documentation. When the platform is built by knowledgeable engineers, with leadership grounded in clinical experience, it delivers on that promise.