AI in Physical Therapy: Trends impacting the standard care model
Insights from APTA Magazine’s December 2023 article “What’s the Impact of AI on Physical Therapy”
Just a year ago, AI felt like a futuristic concept. But with ChatGPT passing the medical board exam and a massive influx of investment in 2023 (36 billion dollars!), it’s clear that AI, particularly language learning models, is here to stay, redefining healthcare.
This article focuses on the impacts of AI on physical therapists (PTs), including:
- positive impacts (increased efficiencies, remote monitoring, personalized treatments)
- and the concerns surrounding its implementation (ethics, privacy, misinformation).
Modernizing Physical Therapy with AI
As a PT with over a decade of experience, I’ve witnessed firsthand the inefficiencies plaguing healthcare.
AI could be the key to solving these issues, starting with reducing administrative burdens. Imagine EMR systems using AI to cut documentation time from hours to minutes!
This would free up clinicians to focus on patient care… which is better for everyone :)
Highlights of AI-driven workflows include:
- EMR systems integration of AI’s language learning models (LLM’s), which will likely cut down on documentation time. Luna has been a leader in this space since 2018 applying machine learning to cut down on text-to-speak documentation time inside a mobile friendly HIPAA compliant App. Hinge Health is now testing an AI-empowered documentation software that interprets unfiltered video and spoken word from the chart visit functioning as a real-time e-scribe. This could free up the therapist to provide a more personalized experience instead of dual tasking with chart notes during the visit.
- Application of video capture and analysis, allowing for real-time feedback to clinician and client. Companies like ExerHealth, ViFive and Hinge Health are already pioneers in computer vision and video movement capture for real-time clinically usable biofeedback. Imagine tracking movement quality without wearable sensors and providing real-time biofeedback, all without anyone needing to travel. ExerHealth is an App already in use in over 100 countries with video capture without the need for wearable devices, that tracks range of motion measurements and has chat capabilities which informs PTs exercise patient protocols. This real-time data enhances self-efficacy, improves compliance, productivity, patient confidence with the potential for cost-savings.
- Additional workflow features of patient self-scheduling and billing workflows which have the potential to reduce late cancellations, improve reimbursement and improve the overall bottom line.
Another exciting prospect is remote monitoring using AI-powered devices.
This could be a game-changer for patients facing barriers like time constraints, rural locations, or transportation challenges.
An added perk for the clinician is opportunities for hybrid work, enhancing work-life balance.
The Future of Clinical Care Practice
- Frees up the PT from lower level cognitive tasks and can reduce bias. One of the issues both new and seasoned clinicians can succumb to are errors in thinking, such as cognitive biases.
AI could reduce our blinders and offer diagnostic assistance when we need to reconsider another path, support interdisciplinary care or offer another treatment plan we did not readily consider.
- AI could offer quicker analysis.
- More precise diagnostics translating into more personalized treatment plans.
- AI could assist with research helping therapists stay up-to-date on the latest trends in practice.
Preparing for the AI Future
During a recent conversation with PT students, I discovered heavy use of the Clinical Practice Guidelines, while AI integration remains largely unexplored.
I asked them what is your school policy on use and integration of “AI” in your classes?
They both gave me blank stares. As AI is already influencing diagnostics and offering personalized treatment options, incorporating it into education, training and everyday practice will be crucial.
AI is still emerging into our professional landscape that many in academia do not yet have formal policies in place with regard to AI and student work.
This will inevitably change.
Our profession is an evolution from just a decade ago when the Clinical Practice Guidelines were being developed and now they are our standard of care.
AI could go hand in hand with our professional value of integrating current best practice protocols, suggested clinical tests to consider and inform the movement diagnosis as well as offering potential treatment options with corresponding billing codes.
Ethical Considerations and Safeguarding Our Profession
We all know that the real world application of new technologies has unforeseen consequences and concerns that need to be addressed. So many questions arise with the ethical practice of AI.
- How will a patient or clinician avoid incorrect misinterpretation of AI generated information?
- How would we ensure the quality of the data interpretation?
- How would we avoid generalization and stay true to the ICF model of functioning, the biopsychosocial model of care and provide high quality, personalized care in an AI workflow and environment?
Unfortunately, we do not have all the answers yet, but EMRs that integrate AI will continue to need to ensure HIPAA compliance and each practice will need to develop best practices around AI so the clinician can always uphold their patient care duty to “do no harm”.
Embracing the Electric Drill
One of the interviewed clinicians from the AI APTA article, CJ Morrow, had a great metaphor for AI application to the physical therapy world:
using a screw-driver and now having the opportunity to use an electric drill.
And figuratively needing to put on our safety goggles as a precaution to ensure the new tool is applied well to make our jobs better and easier.
I commend Keith Loria for posting such an insightful and timely article as the year comes to a close and we are encouraged to keep a pulse on what the future brings.