Is there 'trustworthy' evidence for using manual therapy to treat patients with shoulder dysfunction?: A systematic review.

This review set out to answer a specific question: is there 'trustworthy' evidence for using manual therapy, things like joint mobilization or manipulation done by a clinician's hands, to treat shoulder problems? The researchers used a strict, pre-planned method that went beyond the usual systematic review checklist. They didn't just look at whether studies were randomized controlled trials, they also checked whether each trial had been registered publicly before it started, whether the results matched what the researchers originally set out to measure, and whether the risk of bias was low. Out of 26 randomized controlled trials that fit their population and treatment criteria, only 15 were registered at all, and just 3 were registered before the study began, which is the standard that helps prevent researchers from changing their story after seeing the results. Of those three, two had conclusions that didn't line up with what they'd originally said they were measuring.

The one trial left standing turned out to have a high risk of bias, meaning it couldn't be trusted enough to move forward in the analysis. Because of this, the review found there were zero trials that met their full bar for 'trustworthy' evidence. As a result, no meta-analysis (a pooled statistical summary) could even be attempted. The authors conclude that, right now, a trustworthy living systematic review on manual therapy for shoulder dysfunction simply isn't possible given the current evidence base.

If you've been dealing with shoulder pain or dysfunction and have had or considered manual therapy, like a physiotherapist or chiropractor moving or manipulating your shoulder joint, this study looked at whether the research behind that treatment holds up to close scrutiny. The researchers reviewed 26 clinical trials on manual therapy for shoulder problems published over about 13 years. What they found was that most of these trials had issues with how they were set up or reported, things like not registering the study details in advance, or having results that didn't match what they said they'd measure.

After applying their strict criteria, not one single trial met the bar the researchers had set for being genuinely reliable evidence. This doesn't mean manual therapy doesn't work or that people haven't felt better after receiving it. It means the existing research, at least by this particular set of strict standards, doesn't yet offer strong enough proof one way or another.

If you've had a positive experience with manual therapy for your shoulder, this study isn't saying that experience is invalid, it's pointing out that the science backing it up still needs stronger, more transparent studies before firm conclusions can be drawn.

This is a summary of research, not medical advice. Talk to your own healthcare provider about anything affecting your care.

This review applied a notably rigorous filter, going beyond standard PEDro scores and risk-of-bias ratings to also check prospective trial registration and whether study conclusions actually matched the pre-specified primary outcomes. Of 26 randomized controlled trials meeting the PICOTS criteria for shoulder dysfunction and manual therapy, only 15 were registered in any form, and only 3 were prospectively registered.

Two of those three had discussion and conclusion sections that didn't align with their stated primary outcomes, and the remaining trial was rated high risk of bias. The practical result is that no trials passed through to the meta-analysis stage, meaning no pooled effect size for manual therapy in shoulder dysfunction could be generated under this protocol. This isn't a statement that manual therapy lacks any effect, it's a statement about the structural quality and transparency of the trial base itself. For clinicians consuming this literature, this raises a question about how many other reviews in musculoskeletal manual therapy might reach different conclusions depending on how strictly registration and outcome alignment are enforced.

It also underscores that a high PEDro score or low traditional risk-of-bias rating doesn't necessarily capture issues like outcome switching or non-prospective registration, which this protocol specifically screened for.