Skip to main content
Back

Oxford's QMT Turns a Smartphone Into a Chronic Pain Movement Lab

A new open-source pipeline turns a single smartphone video into 3D movement biomarkers for chronic pain patients, matching lab-grade motion capture at a fraction of the cost.

Paper, Full Access
  • What it is: QMT is an open-source computer vision pipeline that extracts 3D movement biomarkers from a single smartphone video, validated for chronic pain patients.
  • Key result: Lab validation against optical motion capture showed strong correlations (r > 0.85) and low mean absolute errors across three clinical movement tasks.
  • Clinical deployment: Tested in two real cohorts -- a fibromyalgia RCT and a 30-day at-home sciatica monitoring study -- with test-retest reliability above r=0.86.
  • Home monitoring works at group level: QMT distinguished sciatica patients from healthy controls using only remote home recordings, but individual-level classification remains noisy.
  • How it works: 2D keypoint detection + temporal 3D pose lifting + leave-one-out bias calibration + clinical kinematic feature extraction, all from one camera.
  • Availability: The pipeline and protocol are described as open-source; full paper on arXiv.

Measuring how much chronic pain actually limits a person's movement has always required a trip to a specialized lab. Optical motion capture systems -- the same technology used in film production -- are the clinical gold standard for biomechanical assessment, but they cost tens of thousands of dollars, require expert operators, and confine patients to an artificial environment. A new paper from Oxford researchers proposes a radically simpler alternative: a computer vision pipeline called Quantitative Movement Testing (QMT) that extracts clinically meaningful 3D movement data from a single standard smartphone video.

The results are striking. Validated against gold-standard optical motion capture, QMT achieved correlations above r=0.85 and demonstrated test-retest reliability above r=0.86 in real patient cohorts -- numbers that put it in the same ballpark as much more expensive clinical tools.

The measurement gap in chronic pain

Chronic pain diminishes quality of life by decreasing functional ability, yet objectively measuring this functional impact, rather than relying solely on subjective pain reporting, remains challenging in real-world settings. Clinicians have long known that pain changes the way people move -- reduced range of motion, protective guarding, altered gait -- but capturing those changes quantitatively outside a lab has been nearly impossible.

Currently, clinicians predominantly rely on subjective visual appraisal or static goniometry, neither of which captures the dynamic quality of movement required to monitor disease progression or treatment response in the real world. While highly precise, optical motion capture systems are prohibitively expensive, demand hours of expert labour, and lack ecological validity, as patients may use altered movement strategies when observed in laboratory-based environments. There is also a patient comfort issue: for chronic pain populations with sensory hypersensitivity or tactile allodynia, the standard protocol of attaching reflective markers to the skin can be physically uncomfortable, rendering the method impractical and inequitable.

Conditions driven by nociplastic or neuropathic mechanisms, such as fibromyalgia and sciatica, display distinct motor signatures often exacerbated by kinesiophobia (fear of movement) and central sensitisation. Kinesiophobia is the tendency to avoid movement out of fear of pain -- it is a core driver of disability in these conditions, yet there has been no scalable way to measure it objectively in the real world.