How do you measure gait?

Gait is measured by recording how a person walks and reducing that walk to numbers. Three approaches account for most practice: timed walk tests using a stopwatch and a marked distance, instrumented walkways or body-worn sensors, and camera-based motion analysis. All three report spatiotemporal parameters such as walking speed, cadence, and stride length.

Last reviewed .

The three families of gait measurement

Gait measurement methods differ mainly in what they instrument: the clock, the floor, the body, or the image. That choice determines which parameters you can report and how much of the walk survives as data.

MethodWhat it instrumentsTypically reports
Timed walk test (4 m, 10 m, 6-minute walk)A stopwatch over a marked distanceWalking speed, or distance covered in a fixed time
Instrumented walkway or pressure matThe floor surfaceSpeed, cadence, step and stride length, stance and swing time, step width
Body-worn inertial sensorsThe body, using accelerometers and gyroscopesCadence, step timing, symmetry, variability, and gait outside the clinic
Marker-based optical motion captureReflective markers tracked by fixed camerasFull three-dimensional joint kinematics, plus spatiotemporal parameters
Markerless, camera-based motion analysisThe video image, using pose estimationSpatiotemporal parameters and estimated joint positions, with nothing attached to the body

What actually gets measured

Nearly every gait report is built from the same small set of spatiotemporal parameters, derived from the timing of foot contacts and the distance the body travels between them.

  • Walking speed — distance divided by time, reported in meters per second.
  • Cadence — steps per minute.
  • Step length and stride length — distance between consecutive foot contacts, and between two contacts of the same foot.
  • Stance time and swing time — how long each limb spends in contact with the ground, and how long it spends in the air.
  • Double support time — the portion of the cycle with both feet on the ground, which lengthens as walking becomes more cautious.
  • Step width and variability — the base of support, and how consistent the pattern is from step to step.
  • Symmetry — the difference between limbs on any of the parameters above.

Choosing a method

The method should follow the question. A timed walk test is sufficient if the question is whether a patient walks faster than they did before, and walking speed on its own carries substantial prognostic information.

A stopwatch stops being sufficient when the question involves how the walk was produced rather than how fast it was: which limb is loaded longer, whether double support has shortened, whether an asymmetry is resolving. Those require the floor, the body, or the image to be instrumented.

Marker-based laboratory motion capture remains the reference standard for joint kinematics, and remains impractical for routine visits — it needs dedicated space, calibration, marker placement, and a trained operator.

Common sources of error

  • Distance too short — a walkway that includes acceleration and deceleration in the timed segment reports a slower speed than steady-state walking.
  • Inconsistent pace instruction — "usual pace" and "as fast as you safely can" produce different numbers and are not interchangeable across visits.
  • Changing the protocol between visits — a different distance, surface, footwear, or assistive device makes a comparison unreliable.
  • Single trials — gait varies from trial to trial, so averaging repeated passes gives a more stable estimate.
  • Rater variability — manual timing and observational scoring introduce differences between staff and between sites.

Whatever the method, comparability over time depends more on holding the protocol fixed than on the precision of the instrument.

Where CurveAssure fits

CurveAssure sits in the markerless, camera-based category. A calibration marker placed on the floor gives the algorithm a spatial reference, a device camera records the walk through a browser-based portal, and computer vision estimates body key points across the recording. Nothing is attached to the patient and no laboratory is required.

Because the recording is processed the same way every time, the protocol stays fixed between visits and between sites without depending on who ran the assessment.

References