Abstract
We determined the number of days required, and whether to include weekdays and/or weekends, to obtain reliable measures of ambulatory physical activity in people with Parkinson’s disease (PD). Ninety-two persons with PD wore a step activity monitor for seven days. The number of days required to obtain a reliable estimate of daily activity was determined from the mean intraclass correlation (ICC2,1) for all possible combinations of 1–6 consecutive days of monitoring. Two days of monitoring were sufficient to obtain reliable daily activity estimates (ICC2,1 > 0.9). Amount (p = 0.03) but not intensity (p = 0.13) of ambulatory activity was greater on weekdays than weekends. Activity prescription based on amount rather than intensity may be more appropriate for people with PD.
INTRODUCTION
Routine daily physical activity, including exercise, is critically important for people with Parkinson’s disease (PD). However, clinicians seeking to help their patients with PD to set realistic, individualized physical activity goals face a significant measurement challenge: physical activity patterns of adults in general are highly variable [1, 2]. The degenerative nature of PD can impact patterns of physical activity over time [3] and contribute to the reduced activity levels commonly encountered in people with PD [4]. The extent to which activity patterns in people with PD are highly variable is not well understood.
Interventions which promote physical activity have been shown to be effective in people with PD when measured objectively with an activity monitor [5]. Use of an activity monitor to assess the amount and intensity of daily physical activity is preferable to self-report measures, given that cognitive impairment may limit accuracy of recall in people with PD [6]. A potential barrier to implementing the use of activity monitors in the clinic is that multiple days of user wear are required to obtain a reliable measurement of daily activity [7]. How many days of monitoring are required for patients with PD has not been reported.
The primary aim of this study was to determine how many days of monitoring were required to obtain a reliable measure of the daily amount and intensity of ambulatory physical activity (step counts) in a representative week of people with early to mid-stage PD. A secondary aim was to determine whether monitoring periods should include weekdays, weekend days, or both. We hypothesized that one day of data gathering would be insufficient to reliably estimate daily ambulatory activity, but that six to seven days of data gathering would constitute oversampling. We also hypothesized that physical activity levels would be higher on weekdays compared to weekends.
MATERIALS AND METHODS
Participants
The sample comprised two subsets of participants with PD: 58 participants from a prospective longitudinal study investigating the natural history of functional decline and quality of life in PD and 34 participants from a mobile-health intervention aimed at increasing physical activity. Individuals with idiopathic PD according to the UK Brain Bank Criteria, aged ≥40 years, and mobile with or without a walking aid were recruited from three universities within the USA. Participants were excluded if they were diagnosed with atypical parkinsonism, had previous surgical management for PD, or had cognitive or integumentary impairments which precluded them from wearing an activity monitor. Participants without seven days of valid physical activity monitor data were also excluded from the analysis. All data reported here were baseline data. Both studies’ protocols were approved by the respective Institutional Review Boards. All participants provided institution-specific informed consent prior to data collection.
Physical activity monitoring using step counts
The StepWatch 3 Step Activity Monitor (SAM; Modus Health, Washington DC) was used to capture ambulatory physical activity. The SAM is a triaxial accelerometer worn at the ankle which counts complete gait cycles (strides) taken by the leg to which it is attached. Stride counts are recorded in 1-minute intervals and register 0 during periods of inactivity. The validity of the SAM has been demonstrated in people with PD [8].
Participants were given a SAM on a non-specific weekday and instructed to wear the monitor during waking hours for seven consecutive days, except when bathing, showering or swimming. Using the manufacturer’s software, each SAM was set to the participant’s height, typical walking speed (i.e. slow, normal, fast), and leg motion (i.e. dynamic/fidgety, normal, gentle/geriatric). Participants wore the SAM on the ankle of the leg with the least severe motor impairment, as determined by the sum of the lower extremity items in the motor section of the Movement Disorders Society-sponsored Unified Parkinson Disease Rating Scale. Oral and written instructions were provided regarding proper SAM placement and wearing schedule. Optimal accuracy was verified during the first minutes of recording by comparing monitor step counts, identified via a flashing indicator light, with visual observation. Manufacturer software was used to transform recorded stride counts into step counts (i.e. step count = stride count×2) and to calculate the outcomes of interest: amount of step activity (i.e. number of steps taken per day averaged over seven days) and the number of minutes per day of moderate-vigorous activity (i.e. minutes where the participant recorded >100 steps/minute) [9].
Statistical analysis
The number of measurement days required to obtain a reliable estimate of daily ambulatory physical activity (intraclass correlation coefficient [ICC2,1] >0.90) was determined from the mean ICC2,1 for all possible combinations of 1-6 consecutive days of monitoring. Mean steps/day and moderate-vigorous activity minutes/day over weekdays were compared to weekend means using paired T-tests and Mann-Whitney-U tests, respectively, with α= 0.05. All data were analyzed using SPSS v19.0 (IBM Corp, Armonk NY).
RESULTS
Ninety-two community-dwelling people with PD wore a SAM for seven consecutive days. The mean amount of daily steps was generally high for the cohort but most participants (72%) did not achieve the recommended 21.5 daily minutes of moderate-vigorous intensity activity [9, 10] (Table 1).
Two days of activity monitoring were sufficient to reach ICCs >0.90 for measures of the mean number of steps/day and for moderate-vigorous activity minutes/day compared to the weekly average (Table 2). Average steps/day was significantly greater during weekdays compared to weekends (p = 0.03) but there was no difference in moderate-vigorous activity minutes/day between weekdays or weekends (p = 0.13).
DISCUSSION
This study determined that two consecutive days of objective monitoring were sufficient to establish a reliable daily estimate of ambulatory physical activity in people with PD. Our sample of people with PD took a greater number of steps/day during the week compared to weekend, but there were no weekday versus weekend differences for intensity of walking.
Our finding that reliable estimates of weekly physical activity may be obtained from two days of objective monitoring was similar to some prior findings in older adults [2, 11]. This may have been partly due to heterogeneity in our sample which favorably influenced the calculation of reliability coefficients [12]. In contrast, our finding that people with PD in the USA were more active on weekdays compared to weekends contrasted with recent findings from people with PD in Sweden [13] and in healthy adults [1]. The differences may have been due in part to those studies examining activity counts with graded levels of intensity, whereas we were only able to identify time participants spent walking with moderate or greater intensity. The lack of consistency in a weekday versus weekend effect across both outcomes concurred with variable findings in other samples [1, 11]. Our results suggested that monitoring of amount of steps should encompass both weekdays and weekends, whereas monitoring intensity of activity may encompass any days of the week.
Our results revealed that this representative sample of people with mild-moderate PD took relatively large numbers of steps per day, particularly during weekdays, but overall did not engage in recommended levels of moderate-vigorous activity [9, 10]. This finding suggests that a normative moderate-vigorous activity level target of at least 100 steps/minute may be unrealistic for some individuals with PD, consistent with evidence showing that some older adults may be unwilling or unable to engage in moderate-vigorous activity [10]. Emerging evidence demonstrates the benefits of light intensity activity, overall amount of daily activity and minimizing sedentary time [14]. A focus on increasing the amount rather than intensity of activity may therefore be more appropriate and feasible in promoting increased physical activity and reducing sedentarism for people with PD [5].
There were some limitations to this pilot study. The sample was restricted to participants who were independently mobile and community-dwelling, with relatively good cognition necessitated by the ability to wear and return the SAM. The composition of the sample was biased slightly toward participants who were considered at least somewhat active in terms of their daily step counts [9], suggesting that our estimates of variability in physical activity behavior may not have been entirely representative of people who are more sedentary. Differences in physical activity behavior between sites may have affected the results. There were also device and gait specific limitations. The SAM is an accurate device for tracking steps, not other forms of activity such as cycling; it is also designed for research purposes rather than consumer (i.e. participant) use. The cadence used in this study to define moderate-vigorous activity was determined from healthy young adults, therefore further work is required to determine cadences which reflect moderate-vigorous activity in people with PD. Lastly, health promotion interventions which included use of an activity monitor have led to long term (up to two years) increases in physical activity levels in people with PD [5], suggesting the benefit of such devices as a motivating tool. The recent availability of many new commercially available activity monitors which provide instant feedback on physical activity levels requires further research to determine which devices are able to accurately ascertain physical activity levels in people with PD [15].
In conclusion, two days of monitoring were sufficient to obtain reliable estimates for the daily amount and intensity of ambulatory physical activity during a representative week in people with PD. Nevertheless, it would be prudent from a clinical perspective to have patients wear an activity monitor for an additional day or two to account for the potential of forgetting to wear the device, incorrect placement of the device, or transient problems with the devices. If a patient’s goal is to improve amount of daily steps, monitoring periods should include both weekdays and weekends. If the goal is to improve intensity of activity, any monitoring period can be used.
CONFLICT OF INTEREST
The authors have no conflict of interest to report.
Footnotes
ACKNOWLEDGMENTS
This study was supported by the Davis Phinney Foundation; the Parkinson’s Disease Foundation; the American Parkinson Disease Association and the Utah Chapter of the American Parkinson Disease Association. The authors acknowledge the participants in this research as well as Katy Hendron, PT, NCS and Heather Boies for their assistance and persistence in educating participants, programming the monitors and ensuring the return of the monitors.