Walking Bouts as a Digital Biomarker

A walking bout is a continuous episode of walking, bounded by periods of non-walking. Describing a day in bouts rather than in totals changes what the data can answer. Two participants can record the same daily step count while one accumulates it in a few sustained walks and the other in many short shuffles between rooms, and those two patterns reflect very different functional states. Bout-based measures therefore include the number of bouts per day, the duration and length distribution of bouts, and the walking speed achieved within them. This family of measures was formalised by the Mobilise-D consortium, which defined a standard set of digital mobility outcomes and built a shared processing pipeline for extracting them from a single lower-back sensor, specifically so that real-world walking could be described consistently across studies and conditions.

Bout detection starts from continuous accelerometer, and often gyroscope, data recorded at the lower back, wrist or thigh. An algorithm first classifies which parts of the recording contain walking, then segments those into bouts using a minimum duration and a maximum allowed pause between steps. Both parameters are choices, not constants, and they determine the output: a short permitted pause fragments one walk into several bouts, a long one merges distinct walks. Within each detected bout the same pipeline derives cadence, stride length and walking speed. Mobilise-D has validated its algorithms against laboratory reference systems and across chronic conditions including COPD, Parkinson's disease, multiple sclerosis and proximal femoral fracture, which is the most systematic technical validation this measure family has received.

Walking bout measures are used where the clinical question concerns walking capacity in daily life rather than in a corridor. In COPD they distinguish participants who take many brief walks from those still able to sustain longer ones, a distinction that supervised walk tests miss entirely. In Parkinson's disease bout structure captures fragmentation that emerges before average speed declines, and after stroke it charts whether rehabilitation gains transfer into the community. In the public DiMe endpoint library, ambulation and walking bout endpoints are registered across neurological and respiratory trials, including as primary outcomes. Because the measure describes behaviour rather than capacity, it is usually reported alongside a supervised test such as the six minute walk so that capacity and actual use can be compared.

Bout definitions are not standardised outside the Mobilise-D framework, so bout counts from different studies are frequently incomparable even when derived from the same device. Very short bouts, which dominate in impaired populations, are the hardest to detect reliably and are exactly the ones that matter most in those groups. Lower-back placement gives the best accuracy but is less tolerable for long wear than a wrist device, and wrist-derived bout structure is considerably noisier. Environmental context is also invisible: the measure cannot distinguish a walk taken by choice from one required by circumstance.

• Micó-Amigo ME, et al. Assessing real-world gait with digital technology? Validation, insights and recommendations from the Mobilise-D consortium. J Neuroeng Rehabil. 2023. pubmed.ncbi.nlm.nih.gov
• Rochester L, et al. A roadmap to inform development, validation and approval of digital mobility outcomes: the Mobilise-D approach. Digit Biomark. 2020. pubmed.ncbi.nlm.nih.gov
• Del Din S, et al. Analysis of free-living gait in older adults with and without Parkinson's disease and with and without a history of falls. J Gerontol A Biol Sci Med Sci. 2019. pubmed.ncbi.nlm.nih.gov
• Megaritis D, et al. The construct validity of real-world digital mobility outcomes in people with COPD. ERJ Open Res. 2026. pubmed.ncbi.nlm.nih.gov