A patient walks in holding a printout from a wellness website. It tells them their forward head posture is compressing their spine, accelerating degeneration, and causing their headaches. They want it fixed. They have been told, in confident anatomical language, that their head is too far forward and their body is suffering for it.
You were probably taught something similar. At least implicitly. Somewhere in your training.
The evidence no longer supports it.
The relationship between postural alignment and pain has been one of the more quietly embarrassing stories in musculoskeletal research. For decades, clinical training embedded the idea that measurable deviations from an idealised posture — forward head position, increased lumbar lordosis, rounded shoulders — were meaningful predictors of pain, dysfunction, and future harm. Biomechanically logical. Visually compelling. Largely wrong.
When researchers started testing these assumptions systematically, the correlations that should have been there simply were not. Large observational studies found that forward head posture, measured radiographically, does not reliably predict neck pain. Lumbar lordosis angle correlates poorly with low back pain prevalence. People with significant scoliotic curves live without symptoms. People with apparently ideal sagittal alignment present with debilitating pain. The architecture turns out to be a poor guide to what the nervous system is actually doing.
This does not mean posture is irrelevant. Sustained static loading in any position for prolonged periods does create tissue stress, and movement variability matters clinically. But that is a different argument from the one that dominated training. The old model implied a specific shape was pathological and that correcting it would reduce pain. The research suggests the relationship is far more contingent than that — mediated by sensitisation, psychosocial context, sleep, load history, and individual variation in a way that a plumb line cannot capture.
What makes this particularly uncomfortable is not just that the old model was imprecise. It is that it caused harm that was not immediately visible.
Telling a patient their posture is abnormal and damaging gives them a structural explanation for their pain that is frightening and, in many cases, inaccurate. It directs their attention toward a fixed bodily feature they have limited ability to change. It frames normal human variation as pathology. Research into pain cognition is consistent on this point: nocebo effects from clinical language are real, measurable, and durable. A single consultation built around postural correction as the primary explanatory model can reinforce fear-avoidance in a patient who might otherwise have recovered well. One conversation. Real consequences.
The revision this demands is not dramatic in terms of technique. It is primarily cognitive and linguistic.
It means being precise about what postural assessment can and cannot tell you. It means separating the finding of a forward head position from the conclusion that it is causing the pain in front of you. It means recognising that movement capacity, load tolerance, and a patient's beliefs about their own body are often far more useful clinical targets than the angle of their cervical curve. None of this requires abandoning skilled hands-on assessment. It requires placing that assessment within an evidence base that has moved considerably in the last fifteen years.
The patient with the printout does not need to be told the internet is wrong. They need someone who can explain, clearly and without condescension, that the body is more adaptable and less fragile than that model assumes. That posture changes with mood, fatigue, and context. That variation is not damage. That the goal is not to achieve a particular shape, but to build the capacity to move, load, and recover.
Arriving at that explanation with confidence, and with the evidence to underpin it, is what separates a consultation that helps from one that quietly adds to the problem.
How many patients left your clinic last month with a structural story they didn't need?