AI in Rehab Therapy: Expert Perspectives & Peer Experiences FAQ

You’ve heard about AI, but do you really know what AI is, or how it might work in your practice? Those questions and more are what we sought to answer during our most recent webinar, AI in Rehab Therapy: Expert Perspectives & Peer Experiences. The event was hosted by WebPT Senior Manager of Product Marketing Elisabeth Brown with expert insights from PredictionHealth Co-founder and CEO Pedro Teixeira.  We were also fortunate enough to get hands-on perspectives provided by Owner and Hand Therapist at Coastal Hand and Occupational Therapy Jeff Arnold, and Owner and Physical Therapist at Choice Physical Therapy Andrew Wagner as they shared their own journey from learning about AI to adopting it into their practices.

The session sparked a lot of good conversation and a bunch of questions — not all of which we could get to during the allotted hour. So we’ve collected them here in hopes of shedding some more light on AI to attendees and non-attendees alike. Before we dive in, I should note that, for the purposes of this blog, we’ve left out the more technical questions of how AI works within the WebPT platform — which we’ll be answering during our upcoming deminar “Experience AI in WebPT: Unlock Faster Charting and Better Compliance” on February 5.  

How do you see AI in 10 years in PT?

The continued development of AI within PT, and healthcare more broadly, is an interesting question. As Pedro Teixeira noted during the webinar, it’s a balancing act to find the right role for AI within a practice; no matter how much it may advance, it’s never going to replace what a clinician has in terms of clinical insight, and it certainly can’t replace the essential relationship building that’s key to a patient’s recovery. In the immediate future, AI can continue to improve at the things it’s doing now, with optential advances into things like remote monitoring, reducing diagnostic errors, and tackling unwieldy challenges like predicting patient needs or structuring public health data.

AI has been cited as using a lot of energy. Can we comfortably move forward with using more AI daily with this in mind?

While it’s true that the growth of AI is tied to a potential increase in the energy required to power it, there’s also room for AI to get better—which is to say greener and more efficient. In “AI and energy: Will AI help reduce emissions or increase demand? Here's what to know”, the World Economic Forum highlights some of the steps that need to happen, and are slowly happening, to reduce the energy demands related to AI.  It’s also worth noting that predictive AI, which is what clinicians would be using in this instance, requires less energy than does generative AI.

Is there any restriction on using AI on Medicare notes?

I enjoy checking out CMS’s website for the latest rules and updates as much as the next nerd, and I’ve never come across any restrictions on the use of AI to help create notes. But, to be really sure, I checked with our resident billing anc compliance expert John Wallace, who similarly stated that if there is a rule forbidding it, it’s not one he’s aware of at present. He further noted that the requirement of the note, no matter how it’s produced, is to be patient-specific.   

Do you foresee insurance companies using AI to determine denials in the future, if they're not already doing so?

Well, we don’t have to look too far into the future, because insurance companies are already using AI for exactly that. Humana, Cigna, and UHC have all been sued for their use of AI to automatically deny claims, accelerating the increase of denials that was already skyrocketing under humans.

________________

As I mentioned above, if you're looking for answers on how exactly AI and WebPT work together to offer greater efficiency and improved compliance, be sure to sign up for our deminar on February 5 - and bring your best questions!