Physical therapy is moving into living rooms, guided by phone cameras that track joint angles and software that counts repetitions and flags poor form. The convenience is real, and so is the trade. What a patient gives up is the part of rehabilitation that requires a clinician's hands and eyes in the same room.
A Bayesian network meta-analysis published this summer in the Journal of Medical Internet Research sharpened that picture for one of the most common reasons Americans see a physical therapist. Pooling 20 randomized trials and 1,854 participants with chronic nonspecific low back pain, researchers ranked four delivery models against each other: telerehabilitation, in-person rehabilitation, usual care, and, as an exploratory category, AI-assisted telerehabilitation.
The result was not a clean win for either side. For disability, in-person rehabilitation ranked highest at 4 weeks and at 12 weeks, both with low-certainty evidence, while telerehabilitation ranked highest at 8 weeks with high-certainty evidence. On pain intensity, in-person care ranked highest at four weeks, but telerehabilitation ranked highest at both eight and 12 weeks on moderate-certainty evidence. In-person rehabilitation ranked highest for kinesiophobia, the fear of movement that keeps people from using a painful back.
The AI layer, the one being marketed most aggressively, was analyzed only descriptively because the available evidence was too limited to support firm conclusions. The authors described possible short-term benefits with low-to-very-low certainty. That is the detail consumers should hold onto.
The Parts of a Session a Camera Cannot Replace
Rehabilitation is not only supervised exercise. A significant portion of an initial evaluation involves palpation, passive range-of-motion testing, joint mobility assessment, and manual techniques applied by the therapist.
A camera can measure how far a knee bends. It cannot feel end-range resistance, detect guarding, distinguish a capsular restriction from muscular tightness, or apply a mobilization. It also cannot catch things a clinician notices peripherally, such as a limp on the way into the room, a wince during transfer, or swelling that a patient did not mention.
Screening for serious conditions is the more consequential gap. Physical therapists are trained to recognize signs that a musculoskeletal complaint is something else, including fracture, infection, inflammatory disease or nerve compression requiring urgent evaluation. Remote screening tools depend on what the patient reports and what the camera frames.
This is why most professional models describe hybrid care rather than full substitution, with initial assessments, key treatment phases, and manual therapy done in clinic, and exercise progression, monitoring, and check-ins handled remotely.
Where the Remote Evidence Is Genuinely Strong
The case for telerehabilitation is not weak, and dismissing it would misrepresent the literature.
A narrative review of 42 peer-reviewed open-access articles concluded that telerehabilitation is as effective as in-person care for many conditions, with high patient satisfaction and rare adverse events. A separate review of studies involving knee replacement patients described telerehabilitation as a safe and effective adjunct to knee surgery, with hybrid models identified as the most sustainable approach.
A network meta-analysis of telerehabilitation for musculoskeletal pain found that programs delivered through self-managed rehabilitation applications and through videoconferencing significantly reduced pain intensity compared with usual care or minimal intervention.
Adherence is where remote care sometimes outperforms. A systematic review found real-time video physiotherapy produced attendance and adherence comparable to in-person physiotherapy or slightly better, with similar satisfaction. That matters because the most common reason home exercise fails is that it isn't done.
Access is the strongest argument of all. For a patient without transportation, in a rural county with no local therapist, or managing work and caregiving around appointment slots, the realistic comparison is often not between remote and clinic therapy. It is remote therapy versus no therapy.
The Conditions That Sort into Each Category
The practical question for a patient is not whether telerehabilitation works in general. It is whether it fits their specific problem.
Conditions that tend to do well remotely share a profile: a clear diagnosis already established, a treatment plan built mostly around progressive exercise, and a patient able to self-monitor safely. Chronic low back pain, general deconditioning, ongoing maintenance programs and later-stage post-surgical protocols fall in this group.
Conditions that generally warrant in-person assessment include undiagnosed pain, neurological symptoms such as numbness, weakness, or bowel and bladder changes, significant balance impairment or fall risk, complex post-surgical cases, and any presentation where the diagnosis itself is uncertain. Vestibular rehabilitation, hand therapy and post-stroke motor retraining typically involve hands-on components that do not translate.
Fall risk deserves particular emphasis. A frail older adult exercising alone at home carries a different risk profile than the same patient in a clinic with a therapist within arm's reach, and the comparative evidence in that specific population remains thin.
Equipment is another quiet variable. Clinic programs draw on resistance machines, parallel bars, treadmills, and manual resistance from the therapist. Home programs use body weight, elastic bands, and available furniture, which constrains how much load can be progressed and how quickly.
Questions Worth Asking Before Enrolling in a Program
Anyone considering an app-based or AI-supported program can start by asking who is actually delivering the care. Some products connect patients to a licensed physical therapist. Others provide coaching, exercise libraries, and movement tracking without a licensed clinician directing the plan, which is a materially different service even when the marketing looks similar.
Reasonable questions include whether a licensed physical therapist evaluates the case and sets the plan, how often that clinician reviews progress, what happens if symptoms worsen, and whether an in-person evaluation is available if the remote assessment is inconclusive.
Cost and coverage vary widely. Some programs are covered as a medical benefit, some are offered through employers or insurers as a wellness product, and some are direct-to-consumer subscriptions with no insurance involvement. Patients should confirm which category applies before assuming a claim will be paid.
Anyone with new or worsening symptoms, particularly numbness, progressive weakness, unexplained weight loss, fever with back pain, or loss of bowel or bladder control, should seek in-person medical evaluation rather than continuing a home program. This article is general information and is not a substitute for clinical assessment.
Larger trials of AI-assisted rehabilitation are underway, and the exploratory status of that category in current analyses should improve as data accumulate. MedicalDaily will report on new comparative evidence as it is published.
Key Questions Answered
Does remote physical therapy work? For many conditions, reviews find outcomes comparable to in-person care, with high satisfaction and rare adverse events. Results vary by condition and by the requirements of the treatment plan.
What does the newest evidence show for low back pain? A network meta-analysis of 20 trials found that in-person rehabilitation ranked highest for disability at 4 and 12 weeks, and telerehabilitation ranked highest at 8 weeks. Telerehabilitation also ranked highest for pain intensity at 8 and 12 weeks.
How strong is the evidence for AI-assisted programs specifically? Limited. The same analysis treated AI-assisted telerehabilitation as an exploratory category, describing possible short-term benefits with low-to-very-low certainty.
What cannot be done remotely? Palpation, passive range-of-motion and joint mobility testing, manual therapy, and hands-on screening for conditions that require urgent evaluation.
Which patients are better suited to in-person care? Those with undiagnosed pain, neurological symptoms, significant balance impairment or fall risk, complex post-surgical needs, or conditions requiring vestibular, hand, or post-stroke therapy.
What should someone ask before signing up? Whether a licensed physical therapist evaluates the case and directs the plan, how often progress is reviewed, what happens if symptoms worsen, and whether an in-person evaluation is available.
When should someone stop a home program and seek care? With new or worsening numbness, progressive weakness, unexplained weight loss, fever with back pain, or loss of bowel or bladder control.
