Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Wearable Sensor Use for Assessing Standing Balance and Walking Stability in People with Parkinson’s Disease: A Systematic Review



Postural instability and gait disability threaten the independence and well-being of people with Parkinson’s disease and increase the risk of falls and fall-related injuries. Prospective research has shown that commonly-used clinical assessments of balance and walking lack the sensitivity to accurately and consistently identify those people with Parkinson’s disease who are at a higher risk of falling. Wearable sensors provide a portable and affordable alternative for researchers and clinicians who are seeking to objectively assess movements and falls risk in the clinical setting. However, no consensus currently exists on the optimal placements for sensors and the best outcome measures to use for assessing standing balance and walking stability in Parkinson’s disease patients. Hence, this systematic review aimed to examine the available literature to establish the best sensor types, locations and outcomes to assess standing balance and walking stability in this population.


Papers listed in three electronic databases were searched by title and abstract to identify articles measuring standing balance or walking stability with any kind of wearable sensor among adults diagnosed with PD. To be eligible for inclusion, papers were required to be full-text articles published in English between January 1994 and December 2014 that assessed measures of standing balance or walking stability with wearable sensors in people with PD. Articles were excluded if they; i) did not use any form of wearable sensor to measure variables associated with standing balance or walking stability; ii) did not include a control group or control condition; iii) were an abstract and/or included in the proceedings of a conference; or iv) were a review article or case study. The targeted search of the three electronic databases identified 340 articles that were potentially eligible for inclusion, but following title, abstract and full-text review only 26 articles were deemed to meet the inclusion criteria. Included articles were assessed for methodological quality and relevant data from the papers were extracted and synthesized.


Quality assessment of these included articles indicated that 31% were of low methodological quality, while 58% were of moderate methodological quality and 11% were of high methodological quality. All studies adopted a cross-sectional design and used a variety of sensor types and outcome measures to assess standing balance or walking stability in people with Parkinson’s disease. Despite the typically low to moderate methodological quality, 81% of the studies reported differences in sensor-based measures of standing balance or walking stability between different groups of Parkinson’s disease patients and/or healthy controls.


These data support the use of wearable sensors for detecting differences in standing balance and walking stability between people with PD and controls. Further high-quality research is needed to better understand the utility of wearable sensors for the early identification of Parkinson’s disease symptoms and for assessing falls risk in this population.

PROSPERO Registration



Parkinson’s disease (PD) is an age-related neurodegenerative disorder that results from the loss of neurons within the basal ganglia that produce dopamine, an important neurotransmitter involved in the regulation of movement. As medical advances have extended the life expectancy of the average person, clinical and experimental methods need to progress as well in order to improve the management of the symptoms associated with the disease. It is well understood that deficits in balance and gait are common and disabling features of PD that significantly increase an individual’s risk of falling [1]. Subsequently, many clinical assessments have been developed to evaluate these symptoms in this population. The most common assessments include the Berg Balance Scale [2, 3], the Tinetti Gait and Balance assessment [2], the Timed up and Go test [2, 4] and the postural instability and gait disability (PIGD) score derived from the Unified Parkinson’s Disease Rating Scale (UPDRS) [2, 5]. These assessments are suited to clinical settings because they require little equipment to conduct and provide almost immediate outcomes that can be reported to the patient. However, prospective research shows these tests have poor sensitivity and specificity for identifying prospective fallers in the PD population [2] and may not be sufficiently sensitive to detect changes in balance and walking in people with PD who have mild to moderate disease severity [69].

Given the inherent short-comings of the aforementioned clinical tests, previous research has sought to improve the objectivity of these measures to enhance their ability to track symptom progression and evaluate patient risk. Camera-based three-dimensional motion analysis systems have been commonly used in laboratory settings to examine the walking patterns of people with PD [1012]. However, the methods associated with these assessments are often time-consuming and require specific expertise and expensive motion capture systems that are impractical for smaller clinical spaces. Wearable sensors, such as accelerometers or inertial measurement units (IMUs), offer a more portable, flexible and moderately-priced alternative to camera-based motion analysis systems. Moreover, wearable sensors do not require excessive space for normal operation and outcome measures can be output almost immediately without the need for significant post-processing procedures. Given these strengths, research has recently sought to improve the sensitivity of clinical assessments, such as the Timed Up and Go test, by incorporating accelerometers or IMUs to provide continuous measures of walking [1317]. The results of this research demonstrated that it was possible to detect differences in the performances of people with PD compared with controls by instrumenting the Timed Up and Go test with a wearable sensor [1317].

Wearable sensors have recently shown good test-retest reliability for assessing individuals with PD, particularly for acceleration-based measures calculated in the time domain (e.g. jerk; the first time derivative of acceleration) [13]. Furthermore, a growing body of literature supports the use of wearable sensors to assess standing balance or walking for; i) people with PD and controls [13, 14, 1829]; ii) PD fallers and non-fallers [30, 31]; iii) people with different PD sub-types [17, 3235]; iv) carriers and non-carriers of the LRRK2 gene [36]; and v) people at high risk of developing PD (HRPD) [37, 38]. Results from these studies demonstrate that outcomes derived from wearable sensors are effective for detecting differences in standing balance between HRPD patients, people with PD and controls [38]. When combined in a logistic regression model, it was evident that outcome measures derived from wearable sensors can discriminate HRPD patients from controls using an instrumented functional reach test [37]. Furthermore, three-dimensional accelerometers positioned on the head, trunk or pelvis, have highlighted less rhythmic walking patterns for people with PD who retrospectively reported falling than patients without falls [30, 31]. Collectively, these results suggest that wearable sensors may not only be useful for evaluating changes in a patient’s balance or gait patterns, but may also offer a means of screening individuals for various risk factors associated with PD or falls. Nevertheless, scientifically-rigorous prospective research is needed before stronger recommendations can be provided regarding the use of these devices as predictive instruments for clinical populations.

Despite the expanding body of evidence to support the use of wearable sensors for assessing function in people with PD, it is important to recognise that this area of science is still developing. Furthermore, the adoption of such varying methodological approaches in the existing literature makes it difficult to determine which type of sensor is the best to use and which placements and outcome measures are optimal to maximise the utility of these devices. As such, it was the purpose of this systematic review to examine the available literature that utilised wearable sensors to measure standing and walking balance in people with PD and provide a summary of the best sensor types, locations and outcomes based on a consensus of the literature.


This review was registered with the International Prospective Register of Systematic Reviews on September 3, 2014 (PROSPERO Registration: CRD42014010838). The search strategy and study protocol are available at

Search Strategy

An electronic database search of titles and abstracts was performed in January 2015 using PubMed, EMBASE and the Cochrane Library to identify articles measuring standing balance and walking stability with any kind of wearable sensor among adults diagnosed with PD. The following terms were used for the literature search: ‘Parkinson’, ‘Parkinson’s’, ‘walk’, ‘gait’, ‘balance’, ‘stability’, ‘sensor’, ‘gyroscope’, ‘inertial’, ‘acceleration’ and ‘accelerometer’. Specifically, papers that were included in this review were required to have the term ‘Parkinson or Parkinson’s’ AND (‘walk’ OR ‘gait’ OR ‘balance’ OR ‘stability’) AND (‘sensor’ OR ‘gyroscope’ OR ‘inertial’ OR ‘accelerometer’ OR ‘acceleration’) located within the title and/or abstract. In addition to the systematic electronic database search, a targeted search of the bibliographies of relevant articles was also performed to identify any additional studies for inclusion. The research protocol for this systematic review is included as Supporting Information and outlines the procedures followed and the exact search strategy used for this study (S1 File).

Selection Criteria

Only original, full-text articles published in English between January 1994 and December 2014 that assessed standing balance or walking stability with wearable sensors in people with PD were included in this review. Articles were excluded if they; i) did not use any form of wearable sensor to measure variables associated with standing balance or walking stability; ii) did not include a control group or control condition; iii) were an abstract and/or included in the proceedings of a conference; or iv) were a review article or case study. All studies that met the inclusion criteria were considered for review, irrespective of their research design (cross-sectional, randomised controlled trial, etc). After the initial literature search was completed, two assessors (RPH, MHC) independently screened each of the papers based on their title and abstract and made a decision on the suitability of the paper for inclusion in the review. Once both reviewers had completed this process, any and all discrepancies between the two assessments were discussed until a consensus was reached regarding each paper. Full-text articles were retrieved for all of the papers selected for inclusion based on the title and abstract review process and the full-text of these articles was reviewed for suitability by one assessor (RPH). A flow diagram illustrating the study selection process is provided in Fig 1.

Fig 1. Flow diagram outlining the progression of the study’s systematic search strategy and review process, which led to the identification of the articles included in the review.

Data Extraction and Quality Assessment

Upon selection of the articles for inclusion, one assessor (RPH) extracted and collated information concerning the type and number of participants, their mean age, disease duration and symptom severity, as well as the type and location of the wearable sensor(s) used and the major findings of each study (Table 1). The included studies presented a range of outcomes that sought to gain a better insight into the deficits of standing balance and walking stability evident in people with PD and these included; i) the root mean square (RMS) of segmental accelerations; ii) the harmonic ratio; iii) jerk (the first derivative of acceleration); iv) step or stride variability; v) step or stride regularity/symmetry; and vi) other less commonly-used measures of stability.

Table 1. Summarises the major characteristics of the research design, analyses and outcomes for the studies that met the inclusion criteria for this review.

In addition to extracting and compiling these data, a quality assessment was performed by using a modified version of a previously-developed 27-item quality checklist, designed to accommodate both randomised and non-randomised studies [14]. To evaluate the overall methodological quality of each paper, 25 of the criteria on the quality assessment tool were assigned a score of one point if the criterion was met or a zero if the criterion was not met. If it was not possible or unreasonably difficult for the assessors to determine whether the information required for a particular criterion had been provided by the authors, a score of zero was given for that criterion. Of the remaining two questions on the quality checklist, one question evaluating whether potentially confounding variables had been reported by the authors was assessed on a 2-point scale, where the study was given 2 points if confounders were clearly described, 1 point if they were partially described or 0 points if they were not described. The final methodological aspect of the studies that was evaluated was statistical power, which was more heavily weighted than the other criteria and assessed on a 5-point scale. Studies that achieved a statistical power of ≤70% for the standing balance or walking stability measures were given a score of zero, while those that achieved powers of 80, 85, 90, 95 or 99% were assigned scores of 1 to 5, respectively. Where an appropriate statistical power calculation was not provided by the authors, it was necessary to evaluate the statistical power of each study based on the data presented by the authors. If a statistical power calculation was not reported and the raw data were not presented, the paper was given a score of zero for this criterion. After each paper was assessed against these criteria, the scores were summed and divided by the maximum total points to yield a final score that represented the percentage of total possible points earned. This percentage score was used to evaluate the overall quality of the study using quartiles to classify the methodological quality of the article as either very low (≤25%), low (>25%, but ≤50%), moderate (>50%, but ≤75%) or high (>75%). The methodological quality assessment tool (S2 File) and the scoring of each of the studies included in this review (Table A in S2 File) are provided as Supporting Information.


The initial database search identified 335 articles that were potentially eligible for inclusion in this review. Of the 335 studies identified, 98 were excluded as duplicates, 114 were conference abstracts, six were review articles and six were written in a language other than English. The remaining 115 papers were screened by title and abstract, which resulted in 34 being excluded, based on title and 38 being excluded based on abstract. A manual search was conducted of the bibliographies of those papers that were considered appropriate for full-text review, which identified five additional papers for consideration. Following full-text review of the remaining 44 studies, a further 18 studies were excluded, including one that was unattainable, three that had no control group or condition and 14 that had no sensor-based measure of standing balance or walking stability. The remaining 26 articles were selected for inclusion in this systematic review.

Study Design and Methodological Quality

All 26 studies included within this review had a cross-sectional research design with a broad aim of using different types of wearable sensors to observe or identify differences in standing balance or walking stability for Parkinson’s disease compared with controls or a control condition (e.g. on medication vs. off medication, PD subtypes). Given their cross-sectional nature, ten items were excluded from the methodological quality checklist, as they specifically targeted qualities that are unique to intervention studies. The decision to exclude these criteria was made to ensure that the overall quality of the studies included in this review was not unfairly biased by these items that were not relevant to their chosen design.

Based on the appraisal of methodology quality, eight papers were identified as being of low methodological quality (range = 31.8% to 50.0%), 15 papers were of moderate methodological quality (range = 54.5% to 72.7%) and three papers were of high methodological quality (range = 77.3% to 90.9%). In general, the reviewed papers performed poorly on criteria addressing external validity (e.g. representativeness of the sample), internal validity (e.g. identification of and adjustment for potential confounders) and statistical power (e.g. no power calculation and insufficient details to make an informed appraisal).

Sensor Type and Placement

Multiple wearable sensor types were used within the included articles to assess measures of standing balance and walking stability. Of these studies, 69% reported using three-dimensional accelerometers [14, 1723, 3037, 39, 40], 27% used inertial sensors [13, 2428, 38], and 4% used other types of sensors [28, 29]. Similarly, there were multiple protocols described with respect to the placement of the wearable sensors on the human body. Of the 26 included studies, 85% reported placing a wearable sensor on either the lumbar or sacral region of the trunk [13, 14, 1722, 2427, 3140] and 15% reported placing devices on other body landmarks (e.g. head, shank, wrist) [23, 2830]. Details on the studies included in this review that reported using each specific type and placement of sensors are summarised in Table 1.

Assessment of standing balance and walking stability

Of the 26 included studies, 65% used wearable sensors to assess walking during clinical tests, such as the Timed up and Go Test [14, 28] or during assessments of straight-line walking at a self-selected speed [1723, 27, 2931, 3436, 39]. A wide range of sampling frequencies was used to assess walking stability in the reviewed studies, with authors reporting sampling frequencies ranging between 20 and 1024 Hz. The remaining nine studies (35%) assessed standing balance using an instrumented functional reach test [37], dynamic posturography [24] or one of many pre-existing clinical tests conducted during quiet stance (i.e. the Romberg test, tandem stance, semi-tandem stance, standing with eyes open and eyes closed) [13, 25, 26, 32, 33, 38, 40]. Understandably, the wearable sensors used in these studies were generally set to collect data at a slower rate to those used for assessing the dynamic tasks, with reported sampling frequencies ranging from 50 to 128 Hz.

The included studies reported multiple outcomes of standing balance and walking stability that were calculated from the signals provided by the wearable sensors (e.g. accelerations). Of these outcomes, the most commonly-reported measures of standing balance included postural sway velocity (23% of studies) [13, 25, 26, 32, 33, 38], RMS accelerations (19% of studies) [13, 2426, 38] and jerk (19% of studies) [13, 25, 26, 37, 38]. The most commonly-reported measures of walking stability included, the harmonic ratio (31% of studies) [14, 17, 19, 20, 22, 30, 35, 39] and stride timing variability (27% of studies) [17, 19, 22, 2830, 36]. A summary of the studies reporting each of the outcome measures of standing balance and walking stability is provided in Table 2.

Table 2. Summarises and defines the sensor-based measures of standing balance and walking stability used in the studies included in this review.


The purpose of this systematic review was to examine the existing literature to determine the best types of wearable sensors and the most appropriate anatomical placements and outcome measures to assess deficits in balance and gait between people with PD and controls. Using the methodological quality assessment tool adapted from Downs and Black [41], it was determined that the overall quality of scientific reporting in this area is largely of low to moderate quality. In general, the reviewed papers were lacking details concerning the representativeness of the study population (external validity), the approaches adopted to identify and account for confounding variables (internal validity) and an appropriate justification for the chosen sample size. Interestingly, 62% of the included studies received a score of zero for all of the criteria related to at least two of these three areas, while one study (4%) received a score of zero for all three of these areas. The heavier weighting attributed to the sample size criterion is indicative of the importance of ensuring that a study has sufficient statistical power to identify a difference where one exists and, hence, minimise the likelihood of incorrectly accepting the null hypothesis (i.e. Type II error) [42]. Of the 26 studies included in this review, not one reported the results of a sample size calculation, but 13 (50%) had fewer than 15 participants in each of their groups [13, 1921, 2328, 32, 34, 39] and three others (12%) had at least one group with fewer than this number [29, 37, 38]. While it is important to emphasise that a large sample size is not always required to address a specific research question, reporting the outcome of an appropriate a-priori statistical power calculation is beneficial for determining the overall rigor of the reported findings.

Of the other methodological aspects that were poorly reported, the lack of appropriate detail regarding the influence of confounding variables was quite substantial, as failure to account for these factors may result in a study observing a significant change that is simply the manifestation of another variable not adequately controlled for [43]. For example, it is widely recognised that gait and balance variables are influenced by walking speed [4448] and age [4951], hence if groups differ for either or both of these variables, appropriate adjustments should be made to account for this. Of the reviewed studies, 15 (58%) described the principal confounder(s) of their research and reported having made adjustments to their outcomes to account for these variable(s) [17, 19, 22, 2632, 35, 36, 3840]. Of the remaining studies, four (15%) provided a description of the potential confounders, but lacked clear descriptions of how they were accounted for in their analyses [14, 21, 34, 37], while seven (27%) neither reported nor accounted for their potential confounders [13, 18, 20, 2325, 33]. In the study by Fazio et al [18], it was reported that people with PD had significantly lower accelerations and jerk scores than ataxic patients and healthy controls. However, the age of the patients in the PD group (n = 17) ranged from 60–85 years, while the ataxic patients (n = 24) and controls (n = 24) were aged between 20 and 85 years, with more than 60% of these participants aged less than 60 years. Furthermore, the authors reported that the PD and ataxic patients walked significantly slower than the control participants. Given the differences in age and walking speed between the cohorts, it is difficult to determine whether the reported differences in acceleration profiles were indicative of disease-related changes or whether they were simply representative of age-related and/or speed-related factors. Identifying all potential confounders in this type of research and reporting how they have been accounted for in the analyses is critical to ensuring that any changes in outcome can be confidently attributed to the treatment or disease of interest. Collectively, the results of the methodological quality assessment identified that issues related to internal and external validity, as well as statistical power are typically poorly reported in the literature. It should be emphasised that this does not suggest that the authors did not consider some or all of these factors, but rather suggests that these areas should be given more attention in the reporting of future research. To improve the overall methodological quality of research in this area, it is recommended that scientists use existing research reporting guidelines (e.g. CONSORT, STROBE) when designing and planning the reporting of their studies.

Despite the outlined shortcomings in the reporting of the methods, 81% of the studies described differences between different PD groups and/or a healthy control group for one or more of their sensor-based measures of standing balance or walking stability [13, 14, 1722, 2527, 2937, 39, 40]. However, contradictory findings reported in separate studies suggest that some of the reported outcomes may be more robust than others. For example, two studies that compared PD patients with controls using a standing balance assessment reported no significant differences between the groups for jerk scores [37, 38], while three others reported significantly greater jerk scores for PD patients [13, 25, 26]. Similarly, two studies reported no differences between people with PD and controls for RMS accelerations [24, 38], while three studies reported significantly greater RMS accelerations for PD patients [13, 25, 26]. Sway velocity was another common measure used to evaluate standing balance, but similarly only three studies [25, 26, 33] reported differences between people with PD and controls, while the remaining three did not [13, 32, 38]. It is interesting to note, however, that contradictory findings were presented by the three studies reporting differences between patients and controls for sway velocity, as one study reported reduced values for PD patients while standing with eyes closed [33], while the others reported greater values for people with PD while standing with eyes open [25, 26], but not eyes closed [26]. While each of the studies that assessed standing balance derived their outcomes from a wearable sensor positioned on the trunk [13, 2426, 32, 33, 37, 38], there were some methodological differences that may explain the discrepancies observed between the studies’ reported outcomes. The studies unable to report significant differences in jerk scores, RMS accelerations and sway velocities assessed standing balance using a semi-tandem stance test [38], the Sensory Organisation Test [24], the Romberg test [32] or an instrumented version of the functional reach test [37]. In contrast, the studies that reported significant differences for jerk, RMS accelerations and sway velocities assessed participants during quiet standing with the heels separated by 10 cm [13, 25, 26] or while they stood with their feet together or in a semi-tandem stance with their eyes open and closed [33]. Given the available evidence, it seems that the best recommendation for clinicians seeking to assess standing balance using wearable sensors would be to calculate RMS accelerations or jerk scores from trunk accelerations collected while patients stand with their eyes open and their heels 10 cm apart. However, a degree of caution may be required when considering this recommendation, as three of the four studies that reported differences in standing balance for people with PD appear to have used the same patient cohort, due to the reported demographics being the same for each study [13, 25, 26]. As such, it is possible that the overall interpretation of the existing literature in this area may be biased and the transferability of the findings may be more limited than they appear.

In addition to the nine studies that used wearable sensors to assess standing balance, the remaining 65% used these devices to assess walking stability. These studies reported numerous outcome measures derived from the acceleration signals, but the Harmonic Ratio (HR) was the most commonly-reported measure and was calculated for the head [30] and lumbosacral region [14, 17, 19, 20, 30, 31, 35, 39]. The HR seems to be a sensitive and versatile measure of walking stability, as the reviewed literature reports differences between people with PD and controls [14, 19, 20, 30], PD freezers and non-freezers [35], PD fallers and non-fallers [30, 31], PD patients with different dominant symptoms [17] and different methods of cueing for people with PD [39]. Stride timing variability was the second most common outcome measure for the studies that assessed walking stability, but careful review of the included studies suggested that it may not be a dependable measure for discriminating between different populations. Of the seven studies that reported this outcome, three described differences in stride timing variability between PD fallers and non-fallers [30], PD patients and controls [22, 30] or carriers and non-carriers of the LRRK2 gene mutation [36]. In contrast, four studies reported no differences between PD patients and controls [19, 28, 29] or patients with different sub-types of PD [17]. A common characteristic of those studies reporting differences for the HR and stride timing variability was that they each assessed walking stability during straight line walking. As such, it is recommended that clinicians who wish to assess walking stability using wearable sensors calculate the HR from trunk accelerations collected while patients walk in a straight line at a self-selected speed. While there is some evidence to support the use of stride timing variability to assess walking stability, it would only be recommended as a secondary measure due to the inconsistencies evident within the current literature.

While it was not the primary focus of this review to evaluate the effects of anti-parkinsonian medications, such as levodopa, on measures of standing balance and walking stability, it is an important factor that warrants consideration. It is widely recognised that levodopa improves symptoms of PD (based on the UPDRS) [17, 32], spatiotemporal gait characteristics (e.g. stride length) [52, 53] and performance on clinical tests of balance, such as the Berg Balance scale [54]. Of the studies included in this review, five (19%) reported assessing standing balance or walking stability while patients were not medicated [14, 24, 33, 38, 40], 9 (35%) assessed patients on-medication [1821, 30, 31, 35, 36, 39] and three (12%) assessed patients in both on and off states [17, 22, 32]. Of the remaining studies, six (22%) assessed patients who were not yet being medicated for PD [13, 2529], while three (12%) did not report whether their participants were on or off medication at the time of testing [23, 34, 37]. Interestingly, of the studies not reporting differences in standing balance or walking stability between different groups of PD patients and/or healthy controls, two assessed patients while they were off medication [24, 38], while the other did not report whether patients were assessed on or off medication [23]. Of the three studies that assessed patients on and off medication, only two statistically compared their presented outcomes for the two conditions [22, 32]. For a group of idiopathic PD patients, it was reported that the length and maximal distance of postural sway was significantly increased during normal stance, when patients were assessed on medication [32], which would typically be interpreted as a greater amount of sway during the medicated state. During walking, Weiss et al. [22] reported a significant reduction in the width of the dominant harmonic in the acceleration signal when patients were tested on medication, which represented less variability in the gait patterns of medicated patients. While there is a clear need for further research in this area, the presented findings suggest that wearable sensors can be effectively used to evaluate changes in standing balance and walking stability for different patients who are assessed with or without anti-parkinsonian medication.

Considering that 66% of individuals with PD fall at least once in a given year [11, 55] and nearly 50% of these falls occur during locomotion [56, 57], assessing walking stability and falls risk is critical to ensure that high-risk patients can be easily identified by clinicians. However, to date, there is a paucity of research evaluating the capacity for wearable sensors to identify people with PD who are at a higher risk of prospectively falling. Two of the studies included in this review compared people with PD who retrospectively reported having no falls (non-fallers) to those who reported falling at least once (fallers) in the previous 12 months [30, 31]. Both of these studies reported that PD fallers had less rhythmic movements for the pelvis or lower trunk (as assessed using the HR) in both the anterior-posterior (forward-backward) and vertical directions compared with PD non-fallers [30, 31] and controls [30]. While their retrospective nature makes it difficult to determine whether these deficits contribute to the patients falling or whether they are perhaps a consequence of an increased fear of future falls, the results of these studies provide some support for the use of wearable sensors for screening patients for falls risk. Nevertheless, further prospective research is needed to confirm whether sensor-based measures of standing balance or walking stability are suitable for the assessing falls risk and predicting future falls in this population.

There are a number of limitations that should be considered when interpreting the results of this review of literature. First, the results of the methodological quality assessment included in this systematic review are based on the assessor’s (RPH) interpretation of each of the studies. Often, the results reflect the quality of the reporting of the research and, hence, should not be seen as a critique of the significance of the research and its outcomes. Second, given the relatively small number of studies published in this area and the wide variety of research questions addressed using wearable sensors, it is difficult to make strong recommendations regarding the most appropriate equipment, placements and outcomes for assessing standing balance and walking stability in people with PD. In light of these limitations, the results presented in this systematic review should be considered preliminary and additional work will be required as this field of science continues to evolve.

In conclusion, wearable sensors provide a light-weight, portable and affordable alternative to more expensive three-dimensional motion analysis systems and are effective for detecting changes in standing balance and walking stability among people with PD. However, it appears that some outcome measures may be more useful than others for discriminating patient cohorts from controls. Specifically, measures of jerk and RMS acceleration for the trunk appear to be the best sensor-based measures of standing balance, even under less challenging conditions (i.e. feet apart on a firm surface with eyes open). For assessments of walking stability, a trunk-mounted wearable sensor can be used to assess the rhythmicity of dynamic gait patterns using the HR calculated for the three axes of motion. While some studies have provided support for other more complex frequency-based measures of postural stability, additional research is essential to objectively assess the utility of these measures for the PD population. Future research should give careful consideration to the internal and external validity of their methods and provide an appropriate sample size calculation to support their study, as these aspects could have been better reported in the existing literature.

Supporting Information

S1 File. Systematic search strategy and procedures.


S2 File. The quality of methodological reporting assessment tool and the outcomes of this assessment for each of the included studies.



This project was supported by research funding provided by the Australian Catholic University (Project code #2013000584). Dr Michael H. Cole was also supported by an Australian National Health and Medical Research Council Early Career Researcher Fellowship (Project #GNT1016481)

Author Contributions

Conceived and designed the experiments: RPH MHC. Performed the experiments: RPH MHC. Analyzed the data: RPH MHC. Contributed reagents/materials/analysis tools: MHC GAN PAS. Wrote the paper: RPH MHC GAN PAS. Review and critical feedback on manuscript: MHC GAN PAS.


  1. 1. Shulman L, Gruber-Baidini AL, Anderson KE, Vaughan CG, Reich SG, Fishman PS, et al. (2008) The evolution of disability in Parkinson disease. Movement Disorders 23: 790–796. pmid:18361474
  2. 2. Kerr G, Worringham C, Cole M, Lacherez P, Wood J and Silburn PA (2010) Predictors of future falls in Parkinson disease. Neurology 75: 116–124. pmid:20574039
  3. 3. Landers MR, Backlund A, Davenport J, Fortune J, Schuerman S and Altenburger P (2008) Postural instability in idiopathic Parkinson’s disease: discriminating fallers from nonfallers based on standardized clinical measures. Journal of Neurologic Physical Therapy 32: 56–61. pmid:18645292
  4. 4. Mak MK and Pang MY (2009) Balance confidence and functional mobility are independently associated with falls in people with Parkinson’s disease. Journal of Neurology 256: 742–749. pmid:19240961
  5. 5. Lun V, Pullan N, Labelle N, Adams C and Suchowersky O (2005) Comparison of the effects of a self—supervised home exercise program with a physiotherapist—supervised exercise program on the motor symptoms of Parkinson's disease. Movement Disorders 20: 971–975. pmid:15838853
  6. 6. Allen NE, Canning CG, Sherrington C, Lord SR, Latt MD, Close JC, et al. (2010) The effects of an exercise program on fall risk factors in people with Parkinson's disease: a randomized controlled trial. Movement Disorders 25: 1217–1225. pmid:20629134
  7. 7. Ashburn A, Fazakarley L, Ballinger C, Pickering R, McLellan LD and Fitton C (2007) A randomised controlled trial of a home based exercise programme to reduce the risk of falling among people with Parkinson’s disease. Journal of Neurology, Neurosurgery & Psychiatry 78: 678–684.
  8. 8. Goodwin VA, Richards SH, Henley W, Ewings P, Taylor AH and Campbell JL (2011) An exercise intervention to prevent falls in people with Parkinson's disease: a pragmatic randomised controlled trial. Journal of Neurology, Neurosurgery & Psychiatry 82: 1232–1238.
  9. 9. Lin MR, Wolf SL, Hwang HF, Gong SY and Chen CY (2007) A randomized, controlled trial of fall prevention programs and quality of life in older fallers. Journal of the American Geriatrics Society 55: 499–506. pmid:17397426
  10. 10. Cole MH, Silburn PA, Wood JM and Kerr GK (2011) Falls in Parkinson's disease: Evidence for altered stepping strategies on compliant surfaces. Parkinsonism & Related Disorders 17: 610–616.
  11. 11. Cole MH, Silburn PA, Wood JM, Worringham CJ and Kerr GK (2010) Falls in Parkinson’s disease: kinematic evidence for impaired head and trunk control. Movement Disorders 25: 2369–2378. pmid:20737542
  12. 12. Wilcox RA, Cole MH, Wong D, Coyne TJ, Silburn PA and Kerr G (2011) Pedunculopontine nucleus deep brain stimulation produces sustained improvement in primary progressive freezing of gait. Journal of Neurology, Neurosurgery & Psychiatry 82: 1256–1259.
  13. 13. Mancini M, Salarian A, Carlson-Kuhta P, Zampieri C, King L, Chiari L, et al. (2012) ISway: a sensitive, valid and reliable measure of postural control. Journal of Neuroengineering and Rehabilitation 9: 59. pmid:22913719
  14. 14. Palmerini L, Mellone S, Avanzolini G, Valzania F and Chiari L (2013) Quantification of motor impairment in Parkinson's disease using an instrumented timed up and go test. IEEE Transactions on Neural Systems and Rehabilitation Engineering 21: 664–673. pmid:23292821
  15. 15. Salarian A, Horak FB, Zampieri C, Carlson-Kuhta P, Nutt JG and Aminian K (2010) iTUG: A sensitive and reliable measure of mobility. IEEE Transactions on Neural Systems and Rehabilitation Engineering 18: 303–310. pmid:20388604
  16. 16. Weiss A, Herman T, Plotnik M, Brozgol M, Maidan I, Giladi N, et al. (2010) Can an accelerometer enhance the utility of the Timed Up & Go Test when evaluating patients with Parkinson's disease? Medical Engineering and Physics 32: 119–125. pmid:19942472
  17. 17. Herman T, Weiss A, Brozgol M, Giladi N and Hausdorff JM (2014) Gait and balance in Parkinson's disease subtypes: objective measures and classification considerations. Journal of Neurology 261: 2401–2410. pmid:25249296
  18. 18. Fazio P, Granieri G, Casetta I, Cesnik E, Mazzacane S, Caliandro P, et al. (2013) Gait measures with a triaxial accelerometer among patients with neurological impairment. Neurological Sciences 34: 435–440. pmid:22447360
  19. 19. Lowry KA, Smiley-Oyen AL, Carrel AJ and Kerr JP (2009) Walking stability using harmonic ratios in Parkinson's disease. Movement Disorders 24: 261–267. pmid:18973258
  20. 20. Sejdic E, Lowry KA, Bellanca J, Redfern MS and Brach JS (2014) A comprehensive assessment of gait accelerometry signals in time, frequency and time-frequency domains. IEEE Transactions on Neural Systems and Rehabilitation Engineering 22: 603–612. pmid:23751971
  21. 21. Sekine M, Akay M, Tamura T, Higashi Y and Fujimoto T (2004) Fractal dynamics of body motion in patients with Parkinson's disease. Journal of Neural Engineering 1: 8–15. pmid:15876617
  22. 22. Weiss A, Sharifi S, Plotnik M, Van Vugt JPP, Giladi N and Hausdorff JM (2011) Toward automated, at-home assessment of mobility among patients with Parkinson disease, using a body-worn accelerometer. Neurorehabilitation and Neural Repair 25: 810–818. pmid:21989633
  23. 23. Yang CC, Hsu YL, Shih KS and Lu JM (2011) Real-time gait cycle parameter recognition using a wearable accelerometry system. Sensors 11: 7314–7326. pmid:22164019
  24. 24. Baston C, Mancini M, Schoneburg B, Horak F and Rocchi L (2014) Postural strategies assessed with inertial sensors in healthy and parkinsonian subjects. Gait and Posture 40: 70–75. pmid:24656713
  25. 25. Mancini M, Carlson-Kuhta P, Zampieri C, Nutt JG, Chiari L and Horak FB (2012) Postural sway as a marker of progression in Parkinson's disease: a pilot longitudinal study. Gait and Posture 36: 471–476. pmid:22750016
  26. 26. Mancini M, Horak FB, Zampieri C, Carlson-Kuhta P, Nutt JG and Chiari L (2011) Trunk accelerometry reveals postural instability in untreated Parkinson's disease. Parkinsonism and Related Disorders 17: 557–562. pmid:21641263
  27. 27. Sant'Anna A, Salarian A and Wickstrom N (2011) A new measure of movement symmetry in early Parkinson's disease patients using symbolic processing of inertial sensor data. IEEE Transactions on Biomedical Engineering 58: 2127–2135. pmid:21536527
  28. 28. Zampieri C, Salarian A, Carlson-Kuhta P, Aminian K, Nutt JG and Horak FB (2010) The instrumented timed up and go test: Potential outcome measure for disease modifying therapies in Parkinson's disease. Journal of Neurology, Neurosurgery & Psychiatry 81: 171–176.
  29. 29. Van Emmerik RE, Wagenaar RC, Winogrodzka A and Wolters EC (1999) Identification of axial rigidity during locomotion in Parkinson disease. Archives of Physical Medicine and Rehabilitation 80: 186–191. pmid:10025495
  30. 30. Latt MD, Menz HB, Fung VS and Lord SR (2009) Acceleration patterns of the head and pelvis during gait in older people with Parkinson's disease: A comparison of fallers and nonfallers. Journals of Gerontology: Biological Sciences and Medical Sciences 64: 700–706. pmid:19221191
  31. 31. Weiss A, Herman T, Giladi N and Hausdorff JM (2014) Objective assessment of fall risk in Parkinson's disease using a body-fixed sensor worn for 3 days. PLoS ONE 9: e96675. pmid:24801889
  32. 32. Gago MF, Fernandes V, Ferreira J, Silva H, Rodrigues ML, Rocha L, et al. (2014) The effect of levodopa on postural stability evaluated by wearable inertial measurement units for idiopathic and vascular Parkinson's disease. Gait and Posture.
  33. 33. Rocchi L, Palmerini L, Weiss A, Herman T and Hausdorff JM (2014) Balance testing with inertial sensors in patients with Parkinson's disease: assessment of motor subtypes. IEEE Transactions on Neural Systems and Rehabilitation Engineering 22: 1064–1071. pmid:24760912
  34. 34. Sekine M, Akay M, Tamura T, Higashi Y and Fujimoto T (2004) Investigating body motion patterns in patients with Parkinson's disease using matching pursuit algorithm. Medical and Biological Engineering and Computing 42: 30–36. pmid:14977220
  35. 35. Weiss A, Herman T, Giladi N and Hausdorff JM (2014) New evidence for gait abnormalities among Parkinson's disease patients who suffer from freezing of gait: Insights using a body-fixed sensor worn for 3 days. Journal of Neural Transmission 122: 403–410. pmid:25069586
  36. 36. Mirelman A, Heman T, Yasinovsky K, Thaler A, Gurevich T, Marder K, et al. (2013) Fall risk and gait in Parkinson's disease: The role of the LRRK2 G2019S mutation. Movement Disorders 28: 1683–1690. pmid:24123150
  37. 37. Hasmann SE, Berg D, Hobert MA, Weiss D, Lindemann U, Streffer J, et al. (2014) Instrumented functional reach test differentiates individuals at high risk for Parkinson's disease from controls. Frontiers in Aging Neuroscience 6: 286. pmid:25386137
  38. 38. Maetzler W, Mancini M, Liepelt-Scarfone I, Müller K, Becker C, van Lummel RC, et al. (2012) Impaired trunk stability in individuals at high risk for Parkinson's disease. PLoS ONE 7: e32240. pmid:22457713
  39. 39. Lowry KA, Carrel AJ, McIlrath JM and Smiley-Oyen AL (2010) Use of harmonic ratios to examine the effect of cueing strategies on gait stability in persons with Parkinson's disease. Archives of Physical Medicine and Rehabilitation 91: 632–638. pmid:20382298
  40. 40. Palmerini L, Rocchi L, Mellone S, Valzania F and Chiari L (2011) Feature selection for accelerometer-based posture analysis in Parkinson's disease. IEEE Transactions on Information Technology in Biomedicine 15: 481–490. pmid:21349795
  41. 41. Downs SH and Black N (1998) The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions. Journal of Epidemiology and Community Health 52: 377–384. pmid:9764259
  42. 42. Nayak BK (2010) Understanding the relevance of sample size calculation. Indian Journal of Ophthalmology 58: 469–470. pmid:20952828
  43. 43. Skelly AC, Dettori JR and Brodt ED (2012) Assessing bias: The importance of considering confounding. Evidence-Based Spine-Care Journal 3: 9–12. pmid:23526903
  44. 44. Alexander NB (1996) Gait disorders in older adults. Journal of the American Geriatrics Society 44: 434–451. pmid:8636592
  45. 45. Elble R, Thomas SS, Higgins C and Colliver J (1991) Stride-dependent changes in gait of older people. Journal of Neurology 238: 1–5. pmid:2030366
  46. 46. Helbostad JL and Moe-Nilssen R (2003) The effect of gait speed on lateral balance control during walking in healthy elderly. Gait and Posture 18: 27–36. pmid:14654205
  47. 47. Hirasaki E, Moore ST, Raphan T and Cohen B (1999) Effects of walking velocity on vertical head and body movements during locomotion. Experimental Brain Research 127: 117–130. pmid:10442403
  48. 48. Latt MD, Menz HB, Fung VS and Lord SR (2008) Walking speed, cadence and step length are selected to optimize the stability of head and pelvis accelerations. Experimental Brain Research 184: 201–209. pmid:17717650
  49. 49. Hageman PA, Leibowitz JM and Blanke D (1995) Age and gender effects on postural control measures. Archives of Physical Medicine and Rehabilitation 76: 961–965. pmid:7487439
  50. 50. Steffen TM, Hacker TA and Mollinger L (2002) Age-and gender-related test performance in community-dwelling elderly people: Six-Minute Walk Test, Berg Balance Scale, Timed Up & Go Test, and gait speeds. Physical Therapy 82: 128–137. pmid:11856064
  51. 51. Gill J, Allum JH, Carpenter MG, Held-Ziolkowska M, Adkin AL, Honegger F, et al. (2001) Trunk sway measures of postural stability during clinical balance tests effects of age. The Journals of Gerontology: Biological Sciences and Medical Sciences 56: 438–447.
  52. 52. O'Sullivan JD, Said CM, Dillon LC, Hoffman M and Hughes AJ (1998) Gait analysis in patients with Parkinson's disease and motor fluctuations: influence of levodopa and comparison with other measures of motor function. Movement Disorders 13: 900–906. pmid:9827613
  53. 53. Schaafsma JD, Giladi N, Balash Y, Bartels AL, Gurevich T and Hausdorff JM (2003) Gait dynamics in Parkinson's disease: Relationship to Parkinsonian features, falls and response to levodopa. Journal of the Neurological Sciences 212: 47–53. pmid:12809998
  54. 54. Nova IC, Perracini MR and Ferraz HB (2004) Levodopa effect upon functional balance of Parkinson's disease patients. Parkinsonism and Related Disorders 10: 411–415. pmid:15465397
  55. 55. Wood B, Bilclough J, Bowron A and Walker R (2002) Incidence and prediction of falls in Parkinson's disease: A prospective multidisciplinary study. Journal of Neurology, Neurosurgery & Psychiatry 72: 721–725.
  56. 56. Ashburn A, Stack E, Ballinger C, Fazakarley L and Fitton C (2008) The circumstances of falls among people with Parkinson's disease and the use of falls diaries to facilitate reporting. Disability and Rehabilitation 30: 1205–1212. pmid:18608387
  57. 57. Berg WP, Alessio HM, Mills EM and Tong C (1997) Circumstances and consequences of falls in independent community-dwelling older adults. Age and Ageing 26: 261–268. pmid:9271288