Objective Measurement of Walking Activity Using Wearable Technologies in People with Parkinson Disease: A Systematic Review

Parkinson’s disease (PD) is a complex neurodegenerative disease with a multitude of disease variations including motor and non-motor symptoms. Quality of life and symptom management may be improved with physical activity. Due to technological advancement, development of small new wearable devices recently emerged and allowed objective measurement of walking activity in daily life. This review was specifically designed to synthesize literature on objective walking activity measurements using wearable devices of patients with PD. Inclusion criteria included patients with a diagnosis of PD and exclusion criteria included studies using animal models or mixed syndromes. Participants were not required to undergo any type of intervention and the studies must have reported at least one output that quantifies daily walking activity. Three databases were systematically searched with no limitation on publication date. Twenty-six studies were eligible and included in the systematic review. The most frequently used device was the ActiGraph GT3X which was used in 10 studies. Duration of monitoring presented a range from 8 h to one year. Nevertheless, 11 studies measured walking activity during a 7-day period. On-body sensor wearing location differed throughout the included studies showing eight positions, with the waist, ankle, and wrist being the most frequently used locations. The main procedures consisted of measurement of walking hours during a 2-day period or more, equipped with a triaxial accelerometer at the dominant hip or ankle. It is also important for further research to take care of different factors such as the population, their pathology, the period, and the environment.


Introduction
Neurodegenerative diseases such as Parkinson's disease (PD) can lead to motor [1,2] and non-motor symptoms [3]. The latter also often occur in in the general elderly population but people with PD show a stronger decline in a number of cognitive domains when compared to age-matched healthy adults (e.g., executive, attentional, and visuospatial domains) [4]. Motor symptoms on the other hand express themselves as bradykinesia, rigidity, tremor, and eventually even affect the ability to walk or maintain balance [1,5]. The most common motor-related deficits are gait disorders [6,7] which can lead to a loss of independence and increase the incidence of falls [8]. In addition, people with PD suffer from impaired functional abilities [9][10][11], based on a reduced level of strength [11,12] and lower physical activity levels [13,14].

Study Selection
A total of 72 studies were identified. Additional content (n = 4) was included with this selection from other sources and the article selection process is detailed in a flowchart ( Figure 1). After removing duplicates, a total of 60 studies were screened based on the abstracts. This first step excluded 50% of abstracts according to inclusion and exclusion criteria. Then, the thirty eligible articles remaining were assessed on full text. Four studies were excluded after full reading and finally twenty-six studies were eligible and included in the systematic review.

Study Characteristics
The earliest research dates back to 2004, and interest has sparked over the last years with no less than 14 studies published since 2017 ( Figure 2). Fifty percent of the included studies originates from the USA followed by Sweden and the United Kingdom ( Figure 3).
To compare accuracy of wearable sensors according to on-body location [40] and environment [34]; 5.
To compare PD parameters assessed with the ActiGraph GT3X+ (AGT3X) accelerometer and processed with two different filter settings [40];

6.
To investigate the reliability and validity of a device [43].

Outcomes of Interest
In this review, 24 studies linked to the use of wearable devices in the context of walking activity were included. According to inclusion criteria, all studies somehow measured walking activity. The included studies have reported various outcomes such as steps per day, gait speed, and time spent walking but also more general outcomes such as sedentary time, active time, activity counts, walking activity, light physical activity, moderate to vigorous physical activity, and high intensity physical activity.
Only one study [35] mentioned declared specifically to have recorded weekdays, while four studies [24,25,33,34] did not specify whether weekdays or the weekend was measured. Only two studies indicated clearly if they measured weekdays or weekend days and presented results for each period [23,37].

Discussion
The aim of this systematic literature review was to document the use of wearable technology for objective measurement of walking activity in people with PD [18].
Based on the included studies, triaxial accelerometers were the most represented wearable devices [19][20][21][22][23][31][32][33][34][35][36][37][38][39][40]42], followed by microprocessor-linked devices [24][25][26][27][28]41,43]. Both devices allow for an investigation of walking as well as physical activity measures depending on the underlying software/algorithm packages. At this point it has to be mentioned that some of the stated devices only provide raw sensor data, i.e., accelerations that need custom post-processing to compute the desired parameters. Systems such as the AGT3X sensor or the FitBit devices are extensively used to measure step numbers and physical activity intensity even though the parameters are not necessarily validated in the studied cohort. As an example, Riel et al. (2016) validated the ActiGraph step count for post-stroke survivors at specific walking velocities [45], but not yet for patients with PD [46]. Nonetheless, the devices prove beneficial when estimating steps under freeliving conditions [47], similarly to the FitBit at least in healthy female adults [48]. Overall, commercial wearable devices are accurate within their application specification [49] and it is necessary to understand for which clinical cohort the devices are actually validated and if they can be used in the clinical setting. Furthermore, the implementation of wearables in clinical settings requires formation and training of health professionals and patients when using wearable activity trackers [50,51].
While some wearable devices may be of preference for research, Bodine and Gemperle (2003) highlighted that the function of any wearable tool must outweigh any physical or social discomfort felt when wearing it [52]. This directly influences the sensor placement, which as a consequence affects the reliability and validity of the sensor outcomes. Kim et al. (2019) compared the number of steps recorded by an AGT3X attached at the wrist (least affected hand) and one attached at the waist (right hip) in a cohort of PD patients. At moderate speed (1.05-1.3 m/s), results showed an overestimation of daily step count for the wrist and an underestimation for the device recording at waist level. Similarly, in the laboratory environment, waist-worn sensors showed higher accuracy compared to wrist worn sensors, but all activity monitors underestimated the number of steps [53].
Yet, most of our walking activity happens outside the laboratory and the included studies provide a range from eight hours [33] to one year [26] for the monitoring period. While none of the studies justified their monitoring period, Cavanaugh et al. (2012) provided evidence that walking activity of PD patients did not differ if you measure longitudinal data over the course of one year, compared to shorter monitoring periods. Their findings are supported by Paul et al. (2016) who determined that two consecutive days of monitoring are sufficient to estimate daily activity reliably during a representative week in people with PD [54]. Moreover, the amount of ambulatory activity was greater on weekdays than weekends in this study which is supported by Christiansen et al. (2017) who found a significant difference for number of steps between weekdays and weekend days. This does not only hold true for PD patients but also for adolescent girls who present a greater activity and greater moderate vigorous intensities on weekdays than on weekends [55]. The concept of weekend warriors and couch potatoes is well established and depending on the employment, educational level, and household income, activity behaviors and patterns during the weekend differ, which is supported by [56,57].
However, in PD patients, the situation is more complicated. While overweight/obesity are common [58], age, gender, education, disease duration, Hoehn and Yahr stage, UPDRS-II and UPDRS-III scores, and dosage of levodopa do not correlate with physical activity [59].
In addition, behavioral and environmental factors are known to affect physical performance. A relationship between sleep and physical activity exists, suggesting that sleep quality could deteriorate walking activity [60]. Dog owners also showed greater walking activity compared to people not owning a dog [61,62]. However, walking activity decreased significantly with increasing wind speed, precipitation, and humidity [63].
To conclude, the selection of an appropriate sensor ultimately depends on the purpose of the study, methodological considerations, and the population characteristics [64].
Even though this systematic review highlights major findings to evaluate walking activity, these findings must be interpreted with caution. Only the outcome steps per day was consistent throughout studies, and all other reported outcomes were exclusively reported by the respective study. Future research could analyze these individual outcomes to improve our understanding of the role of the device when measuring walking activity. Data processing techniques vary greatly and make the comparison between studies rather difficult. In addition, this review already includes a large sample of PD patients; however, the variability of PD severity between each group increases the difficulty to compare the specified protocols.

Conclusions
To conclude, this systematic review documents the most frequently used wearable devices as well as data collection procedures and data processing in PD patients. Walking activity is mainly assessed during a 2-day period or more, using a triaxial accelerometer preferably located at the hip or ankle. These findings may be taken into account when evaluating walking activity in PD patients.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.