Posture is a position.
Not a verdict.

A mechanical model may favour an alignment that reduces a particular external moment. A performance task may favour reach, speed or precision instead. Passive tissues, active muscle forces, anatomy and the base of support all contribute. “Efficient” therefore needs a stated outcome and task.

Clinical validity asks whether an observation supports its intended interpretation. An angle may describe geometry well while being poor at identifying who has pain. This distinction is fundamental to all three pages.

A classification needs operational rules: which landmarks, what stance, what threshold, and what happens near a boundary? Without those rules, two observers can use the same name for different shapes. A person may also change category between a relaxed stance and a deliberately posed photograph.

Woldendorp and colleagues reviewed 41 studies of 32 static-posture assessment tools. Reliability received much more attention than validity; responsiveness was not studied. That is not direct validation of these four profiles as distinct clinical disorders.

Repeatability is useful, but insufficient. A reproducible category must still demonstrate that it predicts something important or changes a decision. The drawings above should be read as descriptive teaching conventions, not validated levels of risk.

A review of 15 cross-sectional studies found that adults with neck pain tended to have a more forward head position. The adolescent comparison did not show the same pattern. Because posture and pain were measured together, the review cannot establish which came first.

A later cohort followed 686 adolescents into young adulthood. Neck-posture subgroup did not predict persistent neck pain in men; some more relaxed subgroups in women had lower odds than the upright group. This is not proof that slouching prevents pain. It challenges the idea that one upright profile is universally protective.

Swain and colleagues examined 41 systematic reviews and found no consensus that spinal posture or physical exposures cause low back pain. The underlying studies and exposures varied substantially. This does not prove that a position never affects an individual; it limits a universal explanation.

A radiographic meta-analysis found less lumbar lordosis, on average, in some groups with low back pain—especially in studies involving disc disorders. Different populations and methods produced different results. A group average is not an instruction to increase or decrease everyone’s curve.

A review of 10 studies did not support a consistent difference in thoracic kyphosis between people with and without shoulder pain. Sitting more erect could increase available shoulder elevation in single-session experiments. An immediate change in range is a different outcome from lasting pain relief.

A review of chronic primary headache included 12 studies. Some pooled comparisons found a more forward head position in people with chronic headache. These observational findings do not establish that posture caused the headache or that changing an angle treats every headache type. Migraine, tension-type headache and cervicogenic headache should not be collapsed into one postural explanation.

Work combines duration, force, repetition, visual demand, recovery, autonomy and stress. Changing support or position can be useful without proving that the earlier posture was harmful. Occupational interventions should be judged on discomfort, function and participation—not only on a more upright photograph.

A six-month cluster trial in 193 high-risk office workers found fewer new episodes of neck and low back pain with active-break or postural-shift interventions than in the control condition. It tested structured interventions in a selected setting, not a universal ideal angle or a single break schedule for everyone.

A 2025 review included 46 observational studies of people with and without low back pain. Pelvic tilt showed a small average difference, with considerable variation between studies; lumbar lordosis did not show a clear pooled difference. Measurements and populations differed. In particular, a radiographic pelvic-tilt measure is not interchangeable with a surface ASIS–PSIS angle. These comparisons do not establish cause or a treatment target.

A 2024 review of 23 studies in 969 healthy adults found that strengthening could change cervical and thoracic postural measures. Stretching did not show a clear overall postural effect; the lumbar/pelvic strengthening analysis was based on only two studies. This addresses alignment in healthy adults, not whether changing alignment relieves pain.

A programme may change several things at once: movement exposure, exercise, support, attention, expectations and social contact. If participants improve, the package may be useful. To establish a specific mechanism—such as “the pain improved because a muscle imbalance was corrected”—the study must measure and test that pathway. An improvement alone is not enough.

Nocebo refers to adverse effects influenced by negative expectations and context. It is a useful concept when considering threatening explanations, but not a label to apply automatically to someone who is concerned. Over-medicalisation occurs when ordinary variation is treated as disease without adequate evidence. Reassurance should remain specific: uncertainty about a posture-based explanation does not justify dismissing symptoms.