Welcome to Part 1 of our miniseries on telerehabilitation for musculoskeletal disorders, inspired by the uptick in demand for telehealth. Over the next several weeks, we’ll address this overarching question – How well does telerehab work, for whom, and under what circumstances?
Today, we launch with a discussion of a new article by Andrea Turolla and colleagues that provides an overview. If you don’t have the paper yet, you can download it for free.
Meanwhile, here’s the upshot –
Effectiveness of telerehab
Turolla et al. report on 10 systematic reviews or meta-analyses.
Evidence suggests that post-surgical outcomes with telerehab are similar or superior to face-to-face physical therapy. Conditions mentioned include total joint arthroplasty and surgeries involving the rotator cuff, carpel tunnel, and proximal humerus.
Data also suggest that telerehab may be an appropriate substitute for face-to-face care for chronic musculoskeletal conditions like neck and low back pain, lumbar stenosis, fibromyalgia, and arthritis.
Examination of pain, swelling, strength, balance, gait, and range of motion can be done accurately via telehealth. Exams may be less accurate when nerve function, scar tissue, or spinal posture is involved.
Turolla’s group cautions that the studies are limited in sample size, duration of follow-up, and blinding. More on this later.
Feasibility of telerehab
Three systematic reviews suggest that patients and therapists find telerehab acceptable; it may lead to cost savings.
One narrative review suggests that good access to communication technologies contributes to feasibility.
Data privacy, legal concerns, and threats to reimbursement may be limiting factors.
Now, let’s go back to the caveat
I want to talk about the importance of the limitations that Turolla lists – sample size, duration of follow-up, and blinding. We’ll take them one by one.
Sample size is important for 2 reasons – power and generalizability.
Power is the ability to detect differences between groups, given that they exist. A well-powered study will be more likely to detect differences than a poorly powered study. So, researches want a lot of power. How do they get it?
Three things affect power –
- The size of the difference between groups à Big difference, big power
- The amount of variation within groups à Small variation, big power
- Sample size à Large sample, big power
When a study is not adequately powered, it can fail to detect real differences and erroneously conclude that there is no differencebetween groups, when there really is.
Power is important in the telerehab literature because the claim is that there is no difference between telerehab and face-to-face care.
When we find no difference, there is always a chance that the study was underpowered and failed to detect real differences. And, one explanation for low power… You guessed it – low sample size.
Generalizability is the extent to which study results apply to the entire population of patients and clients.
Researchers and clinicians want highly generalizable results.
So, they can be confident applying them to many, many patients.
Generalizability is highest when a study includes all types of patients, e.g. acute, chronic, old, young, mild, severe, etc.
This sort of variety can only be achieved with large samples.
When Turolla and colleagues suggest that small sample sizes are a limiting factor in the telerehab literature, they are worried that
- we might conclude erroneously that there is no difference between remote and face-to-face rehab (low power) or
- that the results do not apply beyond the individuals enrolled in the studies (low generalizability).
As we examine papers in this miniseries, we’ll look at these potential limitations and work out for ourselves the extent to which they tarnish the telerehab literature.
Duration of Follow-Up
Researchers and clinicians value studies that follow patients for a long time.
They want to be sure that the episode of care had lasting effects.
How long is long enough?
Our colleagues who treat people with temporomandibular dysfunction value self-management. They hope that at 1, 2, or 3 years post, patients continue their home exercises and behavioral modifications to keep their pain at bay. That’s a real success story.
But, how feasible is it to follow participants for 3 years?
Sure, it can be done. There are several famous examples. The Harvard Study of Adult Development lasted 80 years (https://www.adultdevelopmentstudy.org/). But, this work takes time, costs tons of money, and delays publication. Despite all the best efforts, some participants are always lost to follow-up, which creates new threats to the believability of the result. E.g. What happened to those folks who were lost to follow-up? Did they find a new, better therapy? Or, did they stop returning our calls because their symptoms resolved?
There is no perfect duration of follow-up. Except to say that follow-up should be long enough to see whether the intervention did or did not have the desired effect. If long term pain management is the goal, then we need to follow long term. If immediate relief of an acute exacerbation is what we’re after, then we need not follow for very long.
As we examine papers on telerehab, we’ll look at length of follow-up to see whether it was adequate to see (or not see) the desired outcome and whether that outcome is important in our clinic.
Blinding is the process wherein patients, therapists, and/or assessors do not know which intervention is applied. It’s done to minimize treatment expectation and placebo effect.
If patients know (or even think) they are getting the latest, greatest wonder-treatment, this knowledge alone can help them feel better. So, researchers try to keep everyone in the dark (through blinding) to avoid this phenomenon.
Blinding is the bane of our existence in rehab. There are so many things we do that cannot and should not be blinded. Would you like your manual therapist to be blind to whether she is doing a high or low amplitude thrust? Is it possible to blind a patient to exercise? Of course not. And in the case of telerehab, we cannot hide the fact that care is provided through a computer screen.
Can lack of blinding effect the result?
Yes, it can. The patient might think he’s getting short shrift when his therapist cannot lay on hands.
But what are we to do about it?
To compare telerehab to face-to-face care, we must sacrifice blinding.
So, as we look at the papers in this series, we’ll ask more than just “was there blinding?” Recognizing that blinding is impossible in some situations, we’ll ask questions like these –
“Were the effects of treatment expectation minimized?” “If so, how? “Was expectation bias or placebo effect a likely contributor to the result?”
At the end of the series, we should be able to come to our own conclusions about the value of telerehab for the people we serve.