In partnership with
Hi!
Welcome to AIMedily.
This week I’m in Chicago for the American Congress of Rehabilitation Medicine (ACRM) conference — an incredible space to learn, connect, and see where our field is heading.
I’ll share more reflections on that (and on LLMs in medicine) this Friday.
Let’s dive into today’s issue.
🤖 AIBytes
Researchers evaluated whether an AI markerless gait analysis can accurately measure gait kinematics in post-stroke patients—without traditional motion-capture markers.
🔬 Methods
Design: Cross-sectional validation study comparing AI-based markerless gait analysis vs the gold-standard marker-based motion capture system.
The deep-learning model extracted kinematic data from video.
Participants: Adults with chronic hemiparetic stroke and healthy controls.
📊 Results
High agreement between markerless AI and traditional motion capture for hip, knee, and ankle kinematics (correlation coefficients > 0.90).
Errors were within clinically acceptable thresholds for gait evaluation.
The AI system required only a single camera and no reflective markers, reducing setup time and cost.
Minor accuracy differences found in the swing phase and rapid movements.
🔑 Key Takeaways
Markerless AI analysis can quantify post-stroke gait kinematics with similar laboratory accuracy.
Useful in cases when a low-cost and portable setup is necessary.
Further refinement is needed for dynamic or occluded movements, but the technology shows strong clinical promise.
🔗 Markerless analysis of gait kinematics after stroke using computer vision and artificial intelligence. Scientific Reports.2025;15:34577. doi:10.1038/s41598-025-18055-3.
This study tested how MedQuest, a mobile app could speed up and improve the accuracy of mapping the International Classification of Functioning, Disability and Health (ICF) during rehabilitation assessments.
🔬 Methods
Design: Randomized controlled trial at two rehabilitation facilities.
Participants: 185 inpatients with musculoskeletal disorders.
Groups:
Control: paper-based questionnaires.
Experimental: MedQuest app with AI-assisted ICF code generation.
Main outcomes: completion time and agreement between AI- and clinician-generated ICF codes.
📊 Results
App users finished in 18 minutes vs 28 minutes on paper with a 35.7 % reduction (p < 0.001).
AI and clinicians matched on 80 % of codes.
Mean System Usability Scale = 86.8.
Best accuracy: functional domains such as dressing and lifting objects.
Lower accuracy: subjective areas (e.g., pain, family support).
Sensitivity / Specificity: 0.806 / 0.952.
🔑 Key Takeaways
AI-assisted ICF mapping can save clinicians time while maintaining accuracy.
Combining AI + human review ensures reliable results in complex or subjective areas.
The approach could standardize functional evaluations and reduce documentation load in rehabilitation settings.
🔗 Kurban Z, Khassenov D, Burkitbaev Z, et al. Artificial intelligence–enhanced mapping of the International Classification of Functioning, Disability and Health via a mobile app: a randomized controlled trial. Front Public Health. 2025;13:1590401. doi:10.3389/fpubh.2025.1590401
🦾TechTool
Recall lets you save and edit summaries from many sources — videos, articles, PDFs, podcasts, etc. — all in your personal database.
Each summary is categorized and linked to related past content.
Recall uses active recall and spaced repetition to help you test your memory, and keep knowledge over time. Works in 10 languages.
Uses your phone’s camera to track movements and guide exercises after knee, or hip surgery.
The app analyzes motion and provides instant corrections on range and form.
Personalized recovery. Built by physiotherapists, it adapts programs to your stage of recovery and tracks progress automatically.
The Librarian connects your Gmail, Calendar, and Slack to give you a daily summary of tasks, meetings, and priorities.
It remembers links, files, and notes—making your digital workspace searchable and easy to navigate.
You can chat or use voice commands to manage emails, reminders, and documents.
🧬AIMedily Snaps
A Smartphone-based digital ruler to automatically measure strabismus (Link)
The most promising digital health startups of 2025 (Link).
Cleveland Clinic expands its remote patient monitoring program (Link).
Stanford will test an AI tools that provides evidence-based insights personalized to each patient within the Electronic Health Record (Link).
🧩TriviaRX
The American Association for Artificial Intelligence formed a subgroup focused on medical applications in what year?
A. 1965
B. 1975
C. 1980
D. 1985
Now, the answer to last week’s TriviaRX.
✅ B. Dr. Howard A. Rusk founded the first multidisciplinary rehabilitation center after World War II.
That’s it for today.
As always, thank you for taking the time to read.
You’re already ahead of the curve in medical AI— don’t keep it to yourself. Forward AIMedily to a colleague who’d appreciate the insights.
You can either:
💌Forward this email
📲Share this link on your phone.
Thank you!
Until next Wednesday.
Itzel Fer, MD PM&R
Join my Newsletter 👉 AIMedily.com
Forwarded this email? Sign up here
How did you like today's newsletter?
1
