Objectively Assessed Physical Activity and its Association with Balance,Physical Function and Dyskinesia in Parkinson’s Disease

Abstract

Background: The desirable effects of physical activity in individuals with Parkinson’s disease are well-known, although according to results from previous studies factors associated with objectively assessed physical activity are not fully investigated.

Objective: To investigate demographic, disease-related and mobility-related factors that associate with objectively measured physical activity, in a sample of older adults with mild to moderate Parkinson’s disease.

Methods: Demographic, disease-related and mobility-related factors were gathered by interview from a total of 91 older adults with Parkinson’s disease, followed by an evaluation of balance control using the Mini-BESTest. After initial testing, participants wore a tri-axial accelerometer during a week of free-living. Correlation analysis and multiple linear regression was used to investigate factors associated with total PA, represented by total activity counts, and time in brisk walking.

Results: Motor impairment, physical function, body mass index and dyskinesia contributed to the variance of total physical activity, explaining 34 % of the variance, while physical function and balance control were significant factors associated with brisk walking, explaining 22 %.

Conclusions: This study identified factors that have not been shown to associate with objectively measured physical activity previously, such as dyskinesia, balance control and self-rated physical function. The findings also demonstrated that associated factors differ, depending on the activity behavior being investigated. However, other factors than those included in this study may also be of importance.

Keywords

Accelerometry dyskinesias motor activity postural control walking

INTRODUCTION

Older adults with Parkinson’s disease (PD) are especially exposed to negative health effects related to insufficient physical activity (PA), when considering not only higher age but also the symptoms of PD, such as bradykinesia and rigidity. These symptoms may lead to deterioration of balance and walking abilities, followed by a decline in physical function [1]. As previously reported, the amount of PA of individuals with PD may decrease with increased disease severity, although this is not the sole explanatory factor. In fact, individuals with PD have shown to be about 30 percent less physically active than age-matched controls, in spite of effective symptomatic therapies [2, 3]. Therefore, in order to develop effective interventions for PA promotion in PD [4], knowledge regarding factors associated with PA need to be established. For adults in general, demographic factors associated with PA and exercise behavior have been thoroughly investigated. A review on the topic states that some of the most consistent factors significantly associated with higher PA in adults are lower age, male gender, lower weight status and higher self-efficacy [4]. Nevertheless, knowledge regarding factors associated with PA in individuals with PD is limited.

Individuals with PD who exercise have less severe disease, are less depressed, have higher self-rated mobility, higher education, and higher self-efficacy, in comparison to those not regularly exercising [5]. Additionally, low outcome expectations and fear of falling may act as a barrier to exercise [6] and falling experiences as well as the feeling of unsteadiness when turning have been linked to lower self-related self-efficacy and an avoidance or reduced level of activity [7 –9]. Exercise is a sub-category of PA and according to the literature potential disease related factors that associate with PA in PD are incompletely explored [2]. Previous studies have revealed a significant relationship between disease severity, gait impairment, level of disability, higher age, lower physical fitness and PA in individuals with PD [2 , 11]. Furthermore, after adjustment for disease severity there is still a tangible association between level of gait impairment and reduced level of activity [12]. For individuals with PD, mobility-related impairments (e.g. balance and gait) could mediate falls and injuries, increased fear of falling and activity limitations [13 –15]. This highlights the importance of studying the association between these factors and PA. Adding on, previous studies investigating associations have only partly explained the variance in PA and some have utilized self-rated PA or pedometry, both inferior methods in comparison to accelerometry [16].

In the context of PA behavior, the mild to moderate stage of PD is of specific interest. Although disabling features such as falls have arisen and mobility problems are present, afflicted are still ambulatory and considered independent [17]. Therefore, levels of PA in daily living are low at this stage of PD and in order to promote PA, modifiable factors associated with PA need to be further investigated. Thus, the aim of this study was to investigate factors associated with accelerometer-measured PA in older adults with mild to moderate PD. In particular, a wide range of demographic, disease-related and mobility-related factors were included in a multivariate model predicting the total volume of PA as well as minutes of brisk walking.

METHODS

This cross-sectional study was founded on the baseline assessment of a randomized controlled trial (the BETA-PD study; clinical trial number NCT01417598) investigating the effects of balance training (the HiBalance program) in older adults with mild to moderate PD [18]. Data collection covered three parts: i) an interview including self-reported assessment, ii) laboratory testing of gait and balance performance and iii) objective measurement of physical activity in free-living condition.

Participants

A total of 100 community-dwelling older adults (41 women) with PD were recruited via patient organizations, the Karolinska University Hospital, outpatient neurological clinics in Stockholm county and advertisements in local newspapers. The inclusion criteria were impaired balance based on clinical assessment, Hoehn & Yahr (H & Y) score of 2 or 3 [19], age≥60 years, the ability to ambulate independently indoors and≥3 weeks of stable medication for PD. Exclusion criteria were a Mini-Mental State Examination score of <24 [20], atypical PD according to Hughes et al. [21], and/or other medical condition that would substantially influence balance performance. Ethical approval was acquired from the Regional Board of Ethics in Stockholm (DNR 2006/151-31, 2009/819-32, and 2011/37-32). All participants signed informed consent before inclusion.

Physical activity

The accelerometer Actigraph GT3X+ (ActiGraph, Pensacola, FL) was used to measure PA [22]. The GT3X+ is an accelerometer that captures time-varying acceleration in three axes and specifies this in an arbitrary unit named counts. Participants were assigned the accelerometer attached to a belt and instructed to wear it around the hip and positioned lateral to the spine, for the duration of seven days, only removing it for showering, swimming, bathing and at night. Participants were also asked to fill in a wear-time diary during the period.

The ActiLife 6 software (ActiGraph, Pensacola, FL) was used to cleanse, filter and compute accelerometer data. Through software settings a 15 second epoch was chosen and episodes of≥90 minutes of consecutive zeroes was considered non-wear time and not included in the analysis. Small window length (up- or downstream consecutive zero-count window) and spike tolerance (allowance of an interval of non-zero counts) was set to 30 minutes and 2 minutes respectively, to filter out artefactual movements [23]. ActiLife settings contain two alternatives of band-pass filters (low frequency extension-filter or normal data filter). Following the recommendations of Benka Wallén et al. [24], the normal filter option was utilized. Subsequently, data was transferred to a Microsoft Excel spreadsheet (Microsoft, Washington, U.S) for visual evaluation and comparison with reported wear-time from participant diaries.

After visual comparison of reported wear-time from participant wear-time diaries with accelerometer data, days with <540 minutes of wear time were excluded. Data from a minimum of four and a maximum of seven days was used to summarize activity data and data from individuals with a lesser amount of days were considered invalid and therefore excluded, all according to published recommendations [25 –27].

To represent total PA per day, total activity counts (TAC) / day (of the vector magnitude) was used [28]. As a measure of minutes of brisk walking, minutes of walking >1.0 m/s /day was calculated. TAC / day and minutes of walking >1.0 m/s / day were defined as outcome variables. Calculation of minutes of brisk walking was based on a calibration study that generated accelerometer cut points of the vertical (Y) axis counts from older adults with PD walking in different speeds [29]. Sedentary time was calculated using cut points from Aguilar-Farías et al. (2014) [30].

Demographic factors

Data on age (years), gender and civil status (living with partner versus alone) were collected using structured questions. Body mass index was calculated (weight in kilograms / length in meters²) via weight measured on a portable digital scale and length measured using a stadiometer. Participants wore light clothes and in-door shoes while measured.

Disease related factors

Disease duration (years) and daily levodopa equivalency dose (LED) [31] were collected during the interview. Motor impairment was assessed using the H & Y-scale and the Unified Parkinson’s Disease Rating Scale motor score (UPDRS-motor) [32]. Dyskinesia was evaluated on a five grade scale (from none to severe) during motor assessment and thereafter dichotomized (dyskinesia in any body part or no dyskinesia) due to uneven distribution.

Mobility-related factors

Balance control and gait were assessed in a movement laboratory using the Mini Balance Evaluation Systems Test (Mini-BESTest) and a GAITRite electronic walkway system (CIR Systems, Inc., Havertown, PA, USA), respectively. The Mini-BESTest is a clinical test incorporating four domains of balance control and each item is scored from 0 to 2, where a higher score equals greater balance control (maximum score = 28) [33]. For the gait assessment, mean walking velocity (m/s) and mean step length out of six trials was analyzed.

The Swedish Short Form-36 Health Survey (SF-36) and the Parkinson’s Disease Questionnaire (PDQ-39) were distributed to the participants. Two dimensions of these surveys were extracted and used as indicators of physical function, namely the Physical Function (PF) from SF-36 and the Mobility-score from the PDQ-39. These dimensions cover 10 questions regarding the patient’s ability to be mobile in daily living, though the interpretation of sum scores differ; a high PF-score equals fewer limitations whereas a high Mobility-score equals greater problems [34, 35]. The Falls Efficacy Scale-International (FES-I) was used to assess concerns about falling. As a measure of depressive symptoms, the Geriatric Depression Scale (GDS) was distributed [36]. Participants also reported number of falls the last 12 months during the interview.

Statistical analysis

Data regarding demographic-, disease-related and mobility-related factors were compiled fordescriptive purposes. Mean difference between dichotomous independent variables were investigated with an independent samples T-test, or Mann-Whitney U test if non-normally distributed. To investigate potential correlates of total PA and brisk walking, Spearman’s rank correlation test was performed between independent variables (demographic, disease-related and mobility-related factors) and the dependent variables (total PA and brisk walking) separately. Hence all independent variables were assessed for association (or if dichotomous mean difference was tested) to the dependent variables. All independent variables with correlations with a p-value below <0.25 (or those dichotomous variables showing a near significant difference) were carried forward to a backward entry multiple linear regression model (one for each dependent variable). Thereafter, using forced entry of the independent factors (Enter method), age was included to control for its potential effect on the outcome [37]. Since heteroscedasticity of residuals and standardized residuals greater than±3 were observed in both regression models, the outcome variable for both the first and the second model were square root transformed, which abolished these deviations. Subsequently, the assumptions of linearity, independence of errors, homoscedasticity, unusual points and normality of residuals were met. All analyses were performed with SPSS version 22 for Windows (SPSS Inc., Chicago, IL, USA). Values of p < 0.05 were considered significant.

RESULTS

After excluding participants with invalid acce-lerometer data, a total of 91 individuals (39 women) remained. Age ranged from 61 to 87 years and BMI ranged from 16 to 38, with 6% being underweight, 40% normal weight and 54% being overweight or obese. A vast majority of the individuals were retired. Participant demographics as well as disease-specific- and mobility-related factors are presented in Table 1.

Table 2 shows PA-variables within the sample. The percentage reaching one 10 minute bout / per day or more of walking >1.0 m/s was 32% and maximum amount of bouts per day was 5. Participants were sedentary for 75% of the total wear time / day, and brisk walking represented about 4% of wear time / day (Table 2).

The Spearman’s rank correlation test performed between potential correlates and the dependentvariables TAC / day (total PA) and minutes / day walking in>1.0 m/s resulted in nine and eight variables with a p-value <0.25, respectively (Table 3). For dichotomous variables, the result revealed higher total PA for those with dyskinesia compared to without dyskinesia (p = 0.02) and for those with a partner compared to those without partner (p = 0.19), as well as a greater amount of minutes of brisk walking among those with a H&Y score 2 compared to 3 (p = 0.22).

The multiple linear regression model with TAC / day as the dependent variable consisted of four significant factors (motor impairment, physical function, BMI and dyskinesia) and one near significance (balance control, p = 0.06; significant before age adjustment) and had an adjusted R² of 0.34 (R = 0.63, R² = 0.39) (Table 4).

The second model with the dependent variable minutes / day walking >1.0 m/s contained two significant factors and had an adjusted R² of 0.22 (R = 0.50, R² = 0.25) (Table 5).

DISCUSSION

Although studies on associations to PA in PD have been performed previously, this paper adds new factors and relations of interest. The findings revealed that motor impairment, poorer self-rated physical function and higher BMI are associated with lower total PA, whereas dyskinesia has a positive correlation to accelerometer output. For brisk walking, self-rated physical function and balance control were significant factors. The difference in associated factors between the total measure of PA and the measure of brisk walking constitutes a unique finding, and welcomes further investigation.

Associated factors of PA in individuals with PD vary among studies. In Dontje et al’s study (2013) significant differences in activity (measured by accelerometer) between both gender and different disease severity of PD as indicated by the H & Y score were found [11]. Furthermore, van Nimwegen et al. (2011) investigated the influence of disease related factors of PD on self-reported daily PA. The instrument used was an interview-based questionnaire and a significant difference between males and females was also found in this study, although compared to Dontje et al., the difference was reversed (women were more active than men) [2]. In comparison to the current study, all used a measure of total PA and found that motor impairment is an important factor for daily PA. The discrepancies however, are several; first, the inclusion of analysis of factors associated with brisk walking in the current study led to unique findings. Second, the above mentioned studies found gender differences in activity, which we could not produce. Third, physical function, balance control, dyskinesia and BMI were correlated to PA in the current study, associations not investigated in the previous ones.

The measure of physical function was defined by the PF dimension of the SF-36 and PDQ-39 Mobility sub-score. Surprisingly, the more generic measure from SF-36 had a higher correlation to PA in our population and qualified for an independent factor in the linear regression. The PF-sub score is a self-rated measure of the experienced bodily function related to mobility, and covers areas related to activities in daily living such as walking, carrying groceries, bending down or bathing. Adding on, balance control was associated with brisk walking which implies that balance control may have a pivoting role for the execution of functional activities (e.g. brisk walks). Balance control, as well as PF, may be enhanced by specific interventions focusing on postural balance, mobility and strength. In point of fact, the effect of progressive balance training on PA has been reported in a previous study [38], proposing a causal direction and suggesting further investigation.

There may be a variety of reasons for accelerometer data of individuals with PD to differ from healthy older adults, such as shortened stride length (hypokinesia), freezing and shuffling of gait. Evidenced by our results, levodopa-induced dyskinesia could be associated with total PA. Although dyskinesia can be defined as PA, being aware of this symptom is important when studying PA in PD. Despite that onset of dyskinesia may be delayed by postponing treatment, limiting levodopa dosage or by dopamine agonist medication, about 40% of patients experience this symptom after 4–6 years of treatment [39]. Proposedly, if the individual being measured suffers from dyskinesia, this should be taken into consideration when quantifying PA, specifically when measuring total activity, which in previous research has been the most widely utilized variable - as opposed to structured activity, exercise or specific higher intensities. However, it should be considered that the assessment of dyskinesia in the current study is limited, since it is based on one occasion in a clinical context which may be biased. The impact of dyskinesia on activities in daily life needs to be determined by other approaches, such as self-assessed dyskinesia using diaries combined with utilization of modern technology. Algorithms that separate signals from regular movements and dyskinesia have been developed for motion sensors, hence identifying what part of the recorded movement is involuntary [40, 41]. Although not yet available for the most commonly used accelerometers, this is a modern approach that holds promise for future studies of objectively measured PA in PD. Despite limitations in the assessment, the potential interference of dyskinesia on the measure of total PA highlights that a measure such as brisk walking, which does not seem to be affected by dyskinesia, may add valuable information.

Since PD progression leads to increased bradykinesia and gait dysfunction, the inverse relationship between UPDRS-motor score and total PA is intuitive. Adding on, there was no difference in PA between H & Y score 2 and 3. This resonates with the fact that H & Y classification is - although well-established - quite a crude measure, where the main difference between scores 2 and 3 are based upon the clinician’s interpretation of how the subject responds to the pull-test. Since both UPDRS-motor score (including the H&Y grading), lack the patient’s own perception of her/his impairment, self-rated physical function is a valuable additive to the assessment.

The negative association between total PA and BMI is an interesting finding. Although previous studies have shown an increase in body weight and an inverse association between this increase and physical activity level in people with early PD, body weight and BMI in general is lower in people with PD in comparison to healthy age-matched individuals [42]. Low body weight, in combination with poor balance is viewed as a risk factor for bone fractures, subsequent pressure ulcers and pneumonia [43]. Hypothesized causes for weight-loss in PD are reduced energy intake due to dysphagia, anorexia, intestinal hypomobility and/or depression, and increased energy expenditure due to levodopa-induced dyskinesia, tremor and/or rigidity [43]. All this considered, in general it is recommended that people with PD should be advised against weight loss.

This study incorporates some limitations. The participant sample was recruited based on inclusion in the HiBalance program, a balance training intervention. Hence participants were recruited partly based on specific balance abilities in line with the training program. Also, some of the measures were initially chosen to capture potential effects of this intervention, and not primarily as hypothetical factors associated with PA. This constitutes a limitation since other non-motor symptoms or demographical factors may associate to PA, but was not measured in the study. Further limitations is that only individuals with mild-to-moderate disease severity were included, hence limiting generalisation. Nonetheless, since disease related characteristics within the population of individuals with PD can withhold great variation, it would be difficult to generalize in this group of individuals without some classification, such as the H & Y score. Also, although the level of explained variance in the resultant models are above or in line with some previous studies on individuals with PD or stroke [2, 44], it signifies there are several other potential factors associated to PA in this population. Finally, since the study is cross-sectional, the causal direction between variables is unknown, which should be taken into consideration when interpreting the results.

Although limitations exist, the current study contributes with some important information and new methodology. The use of an objective and for the population calibrated measure of PA on a representative Swedish group with mild to moderate PD, gave way to new associations and potentially modifiable associated factors such as balance control and physical function. The objective measure allowed separation of different levels of PA and a double-model approach, where the results propose that associated factors differ when comparing total PA to a relative measure such as brisk walking. The study also confirms previous findings regarding the importance of motor function to stay physically active.

CONCLUSIONS

This cross-sectional study confirms previous findings that motor impairment is negatively associated with PA in older adults with PD. Intriguingly, total PA and brisk walking had principally different associated factors, a finding that is likely explained by the fact that they reflect two different behaviors. The study also identified factors that have not been shown to associate with PA previously. Two of these factors, balance control and physical function, are of particular clinical interest because being modifiable, they are potential targets of interventions aiming to increase PA in this population.

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

Footnotes

ACKNOWLEDGMENTS

The authors wish to thank the following organisations for financial support: The Doctoral School of Health Care Sciences, Karolinska Institutet; The Swedish Parkinson Foundation; The Strategic Research Area of Neuroscience (StratNeuro); The Norrbacka-Eugenia foundation; and The Swedish Research Council. We also wish to thank doctoral candidate Niklas Löfgren for help with data collection and valuable intellectual input.

References

Marsden

(1994) Parkinson’s disease. J Neurol Neurosurg Psychiatry, 57, 672–681.

van Nimwegen

, Speelman

, Hofman-van Rossum

, Overeem

, Deeg

, Borm

, van der Horst

, Bloem

, & Munneke

(2011) Physical inactivity in Parkinson’s disease. J Neurol, 258, 2214–2221.

Lord

, Godfrey

, Galna

, Mhiripiri

, Burn

, & Rochester

(2013) Ambulatory activity in incident Parkinson’s: More than meets the eye?J Neurol, 260, 2964–2972.

Trost

, Owen

, Bauman

, Sallis

, & Brown

(2002) Correlates of adults’ participation in physical activity: Review and update. Med Sci Sports Exerc, 34, 1996–2001.

Ellis

, Cavanaugh

, Earhart

, Ford

, Foreman

, Fredman

, Boudreau

, & Dibble

(2011) Factors associated with exercise behavior in people with Parkinson disease. Phys Ther, 91, 1838–1848.

Ellis

, Boudreau

, DeAngelis

, Brown

, Cavanaugh

, Earhart

, Ford

, Foreman

, & Dibble

(2013) Barriers to exercise in people with Parkinson disease. Phys Ther, 93, 628–636.

Nilsson

, Drake

, & Hagell

(2010) Assessment of fall-related self-efficacy and activity avoidance in people with Parkinson’s disease. Bmc Geriatrics, 10.

Bryant

, Rintala

, Hou

, & Protas

(2015) Relationship of falls and fear of falling to activity limitations and physical inactivity in Parkinson’s disease. J Aging Phys Act, 23, 187–193.

Bloem

, Grimbergen

, Cramer

, Willemsen

, & Zwinderman

(2001) Prospective assessment of falls in Parkinson’s disease. J Neurol, 248, 950–958.

10.

Skidmore

, Mackman

, Pav

, Shulman

, Garvan

, Macko

, & Heilman

(2008) Daily ambulatory activity levels in idiopathic Parkinson disease. J Rehabil Res Dev, 45, 1343–1348.

11.

Dontje

, de Greef

, Speelman

, van Nimwegen

, Krijnen

, Stolk

, Kamsma

, Bloem

, Munneke

, & van der Schans

(2013) Quantifying daily physical activity and determinants in sedentary patients with Parkinson’s disease. Parkinsonism Relat Disord, 19, 878–882.

12.

Tan

, McGinley

, Danoudis

, Iansek

, & Morris

(2011) Freezing of gait and activity limitations in people with Parkinson’s disease. Arch Phys Med Rehabil, 92, 1159–1165.

13.

Boonstra

, van der Kooij

, Munneke

, & Bloem

(2008) Gait disorders and balance disturbances in Parkinson’s disease: Clinical update and pathophysiology. Curr Opin Neurol, 21, 461–471.

14.

Duncan

, & Earhart

(2012) Randomized controlled trial of community-based dancing to modify disease progression in Parkinson disease. Neurorehabil Neural Repair, 26, 132–143.

15.

Leddy

, Crowner

, & Earhart

(2011) Utility of the Mini-BESTest, BESTest, and BESTest sections for balance assessments in individuals with Parkinson disease. J Neurol Phys Ther, 35, 90–97.

16.

Helmerhorst

, Brage

, Warren

, Besson

, & Ekelund

(2012) A systematic review of reliability and objective criterion-related validity of physical activity questionnaires. Int J Behav Nutr Phys Act, 9, 103.

17.

Hely

, Morris

, Reid

, & Trafficante

(2005) Sydney Multicenter Study of Parkinson’s disease: Non-L-dopa-responsive problems dominate at 15 years. Mov Disord, 20, 190–199.

18.

Conradsson

, Lofgren

, Stahle

, & Franzen

(2014) Is highly challenging and progressive balance training feasible in older adults with Parkinson’s disease?Arch Phys Med Rehabil, 95, 1000–1003.

19.

Hoehn

, & Yahr

(1967) Parkinsonism: Onset, progression and mortality. Neurology, 17, 427–442.

20.

Folstein

, Folstein

, & McHugh

(1975) Mini-mental state. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res, 12, 189–198.

21.

Hughes

, Daniel

, Kilford

, & Lees

(1992) Accuracy of clinical diagnosis of idiopathic Parkinson’s disease: A clinico-pathological study of 100 cases. J Neurol Neurosurg Psychiatry, 55, 181–184.

22.

John

, & Freedson

(2012) ActiGraph and Actical physical activity monitors: A peek under the hood. Med Sci Sports Exerc, 44, S86–S89.

23.

Choi

, Liu

, Matthews

, & Buchowski

(2011) Validation of accelerometer wear and nonwear time classification algorithm. Med Sci Sports Exerc, 43, 357–364.

24.

Wallen

, Nero

, Franzen

, & Hagstromer

(2014) Comparison of two accelerometer filter settings in individuals with Parkinson’s disease. Physiol Meas, 35, 2287–2296.

25.

Trost

, McIver

, & Pate

(2005) Conducting accelerometer-based activity assessments in field-based research. Med Sci Sports Exerc, 37, S531–S543.

26.

Matthews

, Ainsworth

, Thompson

, & Bassett

Jr (2002) Sources of variance in daily physical activity levels as measured by an accelerometer. Med Sci Sports Exerc, 34, 1376–1381.

27.

Benka Wallen

, Franzen

, Nero

, & Hagstromer

(2015) Levels and patterns of physical activity and sedentary behavior in elderly individuals with mild to moderate Parkinson disease. Phys Ther, 95, 1135–1141.

28.

Bassett

, Troiano

, McClain

, & Wolff

(2015) Accelerometer-based physical activity: Total volume per day and standardized measures. Med Sci Sports Exerc, 47, 833–838.

29.

Nero

, Benka Wallen

, Franzen

, Stahle

, & Hagstromer

(2015) Accelerometer cut points for physical activity assessment of older adults with Parkinson’s disease. PLoS One, 10, e0135899.

30.

Aguilar-Farias

, Brown

, & Peeters

(2014) ActiGraph GT3X+ cut-points for identifying sedentary behaviour in older adults in free-living environments. J Sci Med Sport, 17, 293–299.

31.

Tomlinson

, Stowe

, Patel

, Rick

, Gray

, & Clarke

(2010) Systematic review of levodopa dose equivalency reporting in Parkinson’s disease. Mov Disord, 25, 2649–2653.

32.

Fahn

, & Elton

(1987) Members of the UPDRS Development Committee(1987), Unified Parkinson’s Disease Rating Scale. In Recent Developments in Parkinson’s Disease, Vol 2, Fah

, Marsde

, Caln

, Goldstein , eds. Macmillan Health Care Information, Florham Park, NJ.

33.

Franchignoni

, Horak

, Godi

, Nardone

, & Giordano

(2010) Using psychometric techniques to improve the Balance Evaluation Systems Test: The mini-BESTest. J Rehabil Med, 42, 323–331.

34.

Ware

Jr , & Sherbourne

(1992) The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care, 30, 473–483.

35.

Peto

, Jenkinson

, Fitzpatrick

, & Greenhall

(1995) The development and validation of a short measure of functioning and well being for individuals with Parkinson’s disease. Qual Life Res, 4, 241–248.

36.

Gottfries

, Noltorp

, & Norgaard

(1997) Experience with a Swedish version of the Geriatric Depression Scale in primary care centres. Int J Geriatr Psychiatry, 12, 1029–1034.

37.

Bauman

, Sallis

, Dzewaltowski

, & Owen

(2002) Toward a better understanding of the influences on physical activity: The role of determinants, correlates, causal variables, mediators, moderators, and confounders. Am J Prev Med, 23, 5–14.

38.

Conradsson

, Lofgren

, Nero

, Hagstromer

, Stahle

, Lokk

, & Franzen

(2015) the effects of highly challenging balance training in elderly with Parkinson’s disease: A randomized controlled trial. Neurorehabil Neural Repair, 29, 827–836.

39.

Ahlskog

, & Muenter

(2001) Frequency of levodopa-related dyskinesias and motor fluctuations as estimated from the cumulative literature. Mov Disord, 16, 448–458.

40.

Ramsperger

, Meckler

, Heger

, van Uem

, Hucker

, Braatz

, Graessner

, Berg

, Manoli

, Serrano

, Ferreira

, Hobert

, & Maetzler

, SENSE-PARK study team (2016) Continuous leg dyskinesia assessment in Parkinson’s disease -clinical validity and ecological effect. Parkinsonism Relat Disord, 26, 41–46.

41.

Pulliam

, Burack

, Heldman

, Giuffrida

, & Mera

(2014) Motion sensor dyskinesia assessment during activities of daily living. J Parkinsons Dis, 4, 609–615.

42.

van der Marck

, Dicke

, Uc

, Kentin

, Borm

, Bloem

, Overeem

, & Munneke

(2012) Body mass index in Parkinson’s disease: A meta-analysis. Parkinsonism Relat Disord, 18, 263–267.

43.

Bachmann

, & Trenkwalder

(2006) Body weight in patients with Parkinson’s disease. Mov Disord, 21, 1824–1830.

44.

Michael

, Allen

, & Macko

(2005) Reduced ambulatory activity after stroke: The role of balance, gait, and cardiovascular fitness. Arch Phys Med Rehabil, 86, 1552–1556.