Every few years our field, manual and movement therapy, finds a new thing to be certain about. And every few years, I fall for it. That's my honest truth after a long career. I've been sure of things that turned out to be wrong, and I've been sure in a new direction before the old certainty had even finished baking.
So let me walk you through it, because my own path is basically the argument.
I started as a mechanic. Muscles and joints. Something was out and I put it back. A bone, a rib, a facet, a stuck bit of tissue. It made sense to my hands and it made sense to the person on the table. You could feel the problem, you could feel the fix, and the story wrote itself. That story built entire professions, mine included.
Then I got the evidence I didn't want. When people look closely, joint position change is hard to show and harder to pin on the technique. You can't reliably move one segment on its own. Push posterior to anterior and the spine rotates instead of translating the neat way we were taught. And the skilled palpation, the thing we swear we can feel, turns out to be shaky from one clinician to the next. This isn't fringe talk anymore. It runs right through the recent reviews, and the biomechanical story sits on the thinnest evidence we've got.
But here's the part I want to be careful about, because this is where people hear me wrong. The evidence humbled the story, not the lever. Load still matters. Force still matters. Tissue still responds to what we do to it, right down to the neuro-immune chatter in bone and joint we're only starting to map. I never stopped being a mechanic. I just stopped believing the mechanic was the whole machine. I'm not throwing the baby out with the hyaluronic bath water.
So I did what the field does. I swung.
Fascia caught me next, and it felt like a promotion. I wasn't wrestling a stubborn joint anymore, I was working a living, body wide web, and the language was richer and the models were actually growing in evidence. And I'll say it plainly, I still believe fascia is one of the most exciting targets we have to understand and use well. But sit with it and you can see what I was really doing. I hadn't left the mechanic behind. I'd handed him a more interesting target. My mistake was never fascia. It was crowning it. I took a real and useful lever and, for a while, treated it like the answer.
Then came the nervous system, and this one felt like waking up. It's all context. It's all prediction. It's all the brain. I loved it, because it sounded like progress and it let me feel modern. And it's real. The nervous system is probably where the most of it happens, and descending modulation and prediction and context aren't decoration, they're central. But look closely at what I did with it. "I'm realigning your pelvis" quietly became "I'm down-regulating your nervous system," and if I couldn't measure the second claim any better than the first, all I'd done was move the certainty from the joint up to the brain. Same mistake, smarter outfit. Not the lever's fault. The swagger's.
Which brings me to where I actually live now, and it's the least sexy of them all. The person. The whole person attached to the joint and the fascia and the brain. Bio-psycho-social, if you want the textbook word, though I've grown to dislike how clean that word makes it sound. Because the person isn't a fifth thing to fix. The person is the console the other levers sit on.
And here's the thing I couldn't see while I was busy swinging. I never actually replaced anything. I kept the mechanic when I found fascia. I kept fascia when I found the brain. I kept all three when I found the person. They stacked. What changed wasn't the levers. It was how tightly I gripped each one.
So I've landed somewhere that sounds like a contradiction and isn't. It depends, and everything matters. Everything matters because every one of these levers is a real input into a system too complex to reduce to any single one of them. It depends because I can't tell you in advance which lever, at what weight, for the person in front of me, and I can't honestly tell you the exact reason it worked afterward either.
Hold everything. Grip nothing too hard.
Here's what the evidence actually hands us, and it's messier than any camp wants. Manual therapy helps. As part of a package of care it improves pain, movement, and function, and it does that at a fair cost. And yet, when you set it against a sham or another active treatment, the slice of the effect that belongs to the technique itself gets small. A recent review of dozens of reviews put it about as plainly as you can. The mechanisms are real and they run across many systems, the quality of that evidence runs from critically low to moderate, and the changes we measure are unlikely to be unique to the technique we picked. Put simply, a lot is happening, and almost none of it proves the why.
So the honest answer to "does manual therapy work" is, it depends. It depends on the person in front of you. Their pain. What they walked in expecting. The room you're both standing in. What they're really trying to get back to.
And it depends on your intent, which is the part we almost never say out loud. Are you putting hands on someone to chase a mechanical change, hunting a lesion to correct through a patho-anatomical lens. Or are you using touch as a sensory spotlight, a way to briefly turn down the threat and shift what the system is predicting, so a window opens. Same hands. Completely different job.
And now the part I've come to think matters most, the thread that runs underneath all four levers. Whatever tissue story I'm telling myself on any given day, my hands are doing something I wasn't trained to name. They're delivering affective touch.
There's a class of nerve fibers in the skin built for exactly this, the CT afferents, and they don't care about my model. They respond to slow, gentle (sometimes deep, but pleasant), skin temperature contact, and instead of reporting to the sensory map that tells you where and how hard, they feed the Insula, the part of the brain wrapped up in interoception, safety, and belonging. This is the touch of a hand on a shoulder, not the touch of a caliper. It signals, at a level below thought, that you are safe and not alone. Sit with what that means for our work. A good deal of what I used to credit to my clever technique may have been riding on the affective quality of the contact itself, present the whole time, no matter which lever I thought I was pulling. That's humbling, because it isn't my genius. And it's freeing, because it's real, it's teachable, and most of us are already doing it without permission to say so.
But affective touch is the doorway, not the room. It lowers the guard, quiets the alarm, and brings the person back into a body that felt, a moment ago, like enemy territory. What you do with that opening is where the change lives. So I couple it, always, with attentive movement. I walk the person from what they feel on the outside toward what they feel on the inside, from exteroception to interoception, and then I spend the window on movement they actually pay attention to. Not distracted reps, not braced guarding, but curious, exploratory, meaningful movement that rewrites the felt sense from the inside. Attention is the coupling agent. Attentive touch plus attentive movement compounds. Passive touch plus a distracted patient mostly evaporates.
Touch, in this version, is the doorway. What the person does once they walk through it, that's the treatment.
Now I have to hold myself to my own standard, because this is exactly the spot where our field trips over its own feet. The sensory and affective model is a better bet than the mechanical one. It fits the biology we can actually see, it fits the emotional and social side of touch we're finally taking seriously, and it changes how I talk to people in ways that help them. But it's a working model. I can't draw you a clean line from my hands to your outcome any more than the alignment crowd could draw theirs. And if I start selling affective touch and threat modulation with the same swagger the next guy sells a rib out of place, then I've learned nothing from any of this. The whole point was to stop needing the certainty, not to trade it in for a shinier one.
None of this changes overnight, and I don't expect it to. Most people walk in convinced something's out of place, because that's the story they've been handed for a hundred years. You don't talk someone out of that with a lecture on reliability coefficients. You change the experience and you change the words. You let the hands and the tools feel good, you pair that feeling with movement that means something, and then you point the credit where it belongs, at the person. Do that enough times and the story starts to shift on its own. Mine did, four times over.
It depends, and everything matters. Sit with both instead of rushing past them. In something as tangled as a human being in pain, that pair isn't weakness. It's the closest thing to the truth I've got.
References
Lewis KJ, Bent J, Brolinson PG, Keter D, Lumpkin E, Mundo J, Napadow V, Reed WR, Russell BS, Tang SY, Barbe MF. Mechano responsiveness of neuroimmune cross-talk in bone and joints. A narrative review with implications for force-based manipulations. Musculoskelet Sci Pract. 2026 Apr;82:103470.
Keter DL, Bialosky JE, Brochetti K, Courtney CA, Funabashi M, Karas S, Learman K, Cook CE. The mechanisms of manual therapy. A living review of systematic, narrative, and scoping reviews. PLoS One. 2025 Mar 18;20(3):e0319586.
Kerry R, Young KJ, Evans DW, Lee E, Georgopoulos V, Meakins A, McCarthy C, Cook C, Ridehalgh C, Vogel S, Banton A, Bergström C, Mazzieri AM, Mourad F, Hutting N. A modern way to teach and practice manual therapy. Chiropr Man Therap. 2024 May 21;32(1):17.
Narenthiran P, Granville Smith I, Williams FMK. Does the addition of manual therapy to exercise therapy improve pain and disability outcomes in chronic low back pain. A systematic review. J Bodyw Mov Ther. 2025 Jun;42:146-152.
Loghmani MT, Keter D, Bove GM, Winkelstein BA, Bulea TC, Olausson H, Pathak MN, Powell R, Cook CE. A Model to guide force-based manipulation research and practice. PLoS One. 2025 Sep 12;20(9):e0331606.