How Many Days Of Monitoring are Necessary to Accurately Characterize Activity of Parkinson's Disease Patients?

Research ObjectivesTo determine a reliable sampling strategy to accurately characterize physical activity in individuals with Parkinson's disease (PD).DesignObservational study with individuals with PD wearing a wrist worn activity monitor for 30 days.SettingGeneral community.ParticipantsIndividuals with mild-Moderate Parkinson's disease (N=31).InterventionsNot applicable.Main Outcome MeasuresA step/day measurement period of various lengths (2, 3, 4, 5, 6, 7, and 14 days) were selected for consecutive and random days. Two-way intraclass correlation coefficients (ICC) were calculated for ten samples of both consecutive and random days for each participant. The ten samples were averaged for ICC for each measurement period.ResultsThe ICC averages for consecutive days ranged from 0.59-0.75, with stability of ICC starting at a 7-day measurement period (95% CI, 0.55- 0.80). The ICC averages for random days ranged from 0.59-0.71, with stability of ICC maintained starting with a random 7-day measurement period (95% CI, 0.51- 0.77).ConclusionsPD-related variability in motor behavior, and conventional methods of a four day measurement period may not accurately reflect physical activity behavior. Identifying a method to accurately characterize level of physical activity in individuals with PD over time provides those conducting short- and long-term intervention studies and trials with a robust physical activity outcome. The results of this study indicate that seven consecutive days of physical activity monitoring was the most stable measure of physical activity. For random sampling days, seven days also resulted in the greatest and most stable ICC of that sampling method. Therefore, in PD a seven-day consecutive measurement period may be the most representative of physical activity behavior in individuals with PD.Author(s) DisclosuresNo authors have disclosures or conflicts.