Longitudinal Quantification of Parkinsonian Gait Using Apple HealthKit: A Single-Subject Digital Phenotyping Study

Background Parkinsonian gait is traditionally assessed using subjective clinical scales such as the Hoehn–Yahr classification and MDS-UPDRS, which lack temporal granularity and sensitivity to daily fluctuations. Advances in smartphone-based sensing now allow continuous, real‑world gait quantification. Objective To determine whether Apple HealthKit gait metrics can detect year‑to‑year progression, daily variability, and gait‑component contributions to walking‑speed decline in a single individual with Parkinson’s disease (PD). Methods A 77-year-old male with PD was monitored continuously from January 2024 to December 2025. HealthKit metrics included walking speed, step length, step count, double‑support time, gait asymmetry, and walking steadiness. Weekly and daily patterns were compared across years. Results Walking speed declined by 14.3% from 2024 to 2025. Step length decreased by 31%. Step count also declined (69% relative contribution in decomposition analysis), although this metric reflects free-living walking activity and may be influenced by reduced walking opportunity rather than intrinsic gait alteration. The decline in walking speed may be partly attributable to reduced step length. Walking asymmetry showed variable data acquisition patterns that were not solely explained by walking volume. Conclusion Smartphone‑derived gait metrics provide objective, sensitive detection of PD gait progression and daily motor fluctuations. These findings demonstrate the potential of pervasive digital monitoring to complement traditional clinical assessments and support individualized disease management.