The McKenzie method is the most structured answer the physiotherapy world has given to a real observation: that people with back pain often move better in one direction than another, and that the direction differs between people. Set against general stretching, it has the better theory and only slightly better trial results. This page sets out what the approach is, what the studies found, where it beats a general routine and where it does not, and why this site does not implement it.
The short answer
- Directional preference is a real, repeatable observation. Some people ease with extension, some with flexion.
- The approach built on it has randomised trials, and their effects are small.
- Matched exercise beat unmatched exercise in the best trial of the idea.
- Against other active care over months, the advantage largely disappears.
- Limber does not implement it, and nothing in the app assesses your direction.
Read this first
Get examined rather than working through any directional routine if you have loss of bladder or bowel control, difficulty passing urine, or numbness in the saddle area alongside back pain — that combination needs assessment the same day. Also: progressive leg weakness, back pain after significant trauma, back pain with fever or unexplained weight loss, a history of cancer, or pain that is constant and worse at night. Symptoms travelling below the knee, with or without numbness, deserve assessment before any routine.
The approach on this page is meant to be applied by a clinician who has examined you, and the difference between doing that and copying a movement off a page is the whole of its evidence base.
Nothing here diagnoses anything and nothing here is treatment.
What the McKenzie method is
Three parts, and only the first is what most people have heard of.
Assessment. The clinician takes you through repeated movements in each direction and watches what the symptoms do — not just whether they hurt more or less, but whether they move. Pain that retreats from the leg toward the back is called centralisation, and it is treated as the key signal.
Classification. The response sorts people into groups, of which the largest is the derangement group — the people whose symptoms change rapidly with repeated movement in one direction.
Treatment. The person is given repeated movements in their responsive direction, self-administered, several times a day, alongside advice about posture and load.
The reason it is often called "the extension exercises" is that extension is the commonest responsive direction, so the press-up became the movement everyone associates with the name. That is a caricature of the method. The direction is meant to be established by examination, and for some people it is flexion.
The whole approach turns on one claim: which direction helps is individual and discoverable. That is a much better starting point than a fixed list of back stretches, and it is the reason the approach deserves a serious hearing.
Directional preference
The strongest trial of the underlying idea is worth its own section.
Long, Donelson and Fung, Spine, 2004. Patients with low back pain were examined for a directional preference and then randomised to exercises matching their direction, exercises opposing it, or non-specific exercises. The matched group did better across pain, medication use and other outcomes. The mismatched group did worst.
That is a clean design and it establishes something specific and useful: within people who show a directional preference, giving them the wrong direction is worse than giving them nothing in particular. It does not establish that the method beats other care in general, because everybody in the trial was already selected for having a preference.
Grade: reasonably solid for the specific claim, and narrow. One well-designed randomised trial, a selected population, unblinded by necessity.
What the trials found
Now the broader question: does the whole approach beat the alternatives?
Machado and colleagues, Spine, 2006, reviewed the literature systematically. The approach came out better than passive comparators — the treatments where the patient does nothing — for acute low back pain. For chronic low back pain the evidence was insufficient to say.
Machado and colleagues again, BMC Medicine, 2010. A randomised trial adding the method to first-line care for acute low back pain. The effects on pain were small and short-term, and the other outcomes did not separate.
Set those beside the general picture. Hayden and colleagues' 2021 Cochrane review of exercise for chronic low back pain found modest benefits and small differences between exercise types; Saragiotto and colleagues' 2016 Cochrane review found motor control exercise similar to other forms rather than better. Every branded approach in this field runs into the same wall: it beats doing nothing, and it struggles to beat the other approaches that also beat doing nothing.
Grade: mixed, trending small. Better than passive care in acute presentations, unclear in chronic ones, and modest where measured against other active care.
General stretching, compared
Against that, what does an ordinary stretching routine have?
Sherman and colleagues, Archives of Internal Medicine, 2011 — 228 adults with chronic low back pain, randomised to twelve weeks of yoga classes, twelve weeks of stretching classes, or a self-care book. Both class groups improved on function more than the book. The two class groups matched each other.
So general stretching has one good trial showing a modest benefit in chronic pain, and directional work has one good trial showing that matching the direction matters within people who have a direction. Those are different claims and neither defeats the other.
The practical difference is what each asks of you. A stretching routine asks for time. A directional approach asks for an assessment — and without the assessment you are picking a direction at random, which the 2004 trial specifically showed can be the worst option available.
That is the honest comparison. If someone has examined you and told you which direction eases your symptoms, doing that direction repeatedly has evidence behind it. If nobody has, a general routine you will actually keep doing is a more sensible default than guessing at extension because a video said so. Stretches to avoid with back pain covers why we publish no universal list of forbidden directions either, for the same reason in reverse.
What we do not do
Limber does not implement any of this, and the reason is worth stating plainly.
Nothing in the app assesses your directional preference. There is no repeated-movement examination, no centralisation test, no classification, and no plan that changes based on which direction eases your symptoms. Those things require someone watching you move and interpreting what your symptoms do over several repetitions, and a phone cannot do it.
The app has no pain protocols at all. Not for the back, not for anywhere. The only pain-related mechanism in the engine is a contraindication filter that removes flagged movements from the pool before a session is built — a refusal, not a treatment.
And the app's own back track is not directional. The four-week Lower Back Care track is twelve days built from fifteen movements, most of them hips, hamstrings and mid-back, and its own description calls it mobility support rather than treatment. It does not ask which way you bend.
What the app can do is measure range and plan range work. The toe touch test reads the posterior chain in five bands, with the standing warning that it is a loaded forward fold and the wrong thing to attempt on a day when bending forward provokes symptoms. The Range Score composes what you have read and leaves the rest blank. The cluster's evidence in full is in lower back pain and the stretching evidence, and the rest sits under lower back.
Questions
Does the McKenzie method work for back pain?
Better than doing nothing, and less clearly better than other active care. A systematic review in Spine in 2006 found it superior to passive comparators for acute low back pain with insufficient evidence in chronic cases, and a 2010 randomised trial adding it to first-line care for acute pain found small, short-term effects on pain only. That pattern — beats nothing, struggles to beat the alternatives — is the same one every branded approach in this field runs into.
What is directional preference?
The observation that many people with back pain respond differently to repeated movement in different directions, and that the helpful direction varies between individuals. In a randomised trial published in Spine in 2004, people examined for a directional preference and then given matching exercises did better than those given non-specific exercises, and those given the opposing direction did worst of all.
Are McKenzie exercises just back extensions?
No, though that is what the name has come to mean. Extension is the commonest responsive direction, so the press-up became the movement associated with the approach, but the direction is meant to be established by examination and for some people it is flexion. Copying an extension routine because it is the famous one is exactly the situation the 2004 trial suggests can go wrong.
Should I do extensions or flexions for my back?
Nobody can tell you without watching your symptoms respond to repeated movement, which is the assessment the approach is built on. Guessing is not neutral — the trial that tested this found the mismatched group did worse than a non-specific routine. If the answer matters to your plan, that is a good reason to book an appointment with a physio rather than to pick a direction from a page.
Is McKenzie better than ordinary stretching?
They rest on different claims and neither defeats the other. General stretching has one good trial showing modest benefit in chronic low back pain, where twelve weeks of stretching classes matched twelve weeks of yoga. Directional work has one good trial showing that matching the direction matters among people who have one. Without an assessment you cannot access the second, so a general routine you will keep doing is a sensible default.
Does Limber do directional assessment?
No. There is no repeated-movement examination, no classification, and nothing in the plan that changes based on which direction eases your symptoms. The app measures range and plans range work, and its only pain-related mechanism is a filter that removes flagged movements before a session is built. Anything more than that needs a clinician who can watch you move.