Reliability and Validity of Force Platform Measures of Balance Impairment in Individuals With Parkinson Disease
- C.C. Harro, PT, MS, Department of Physical Therapy, College of Health Professions, Grand Valley State University, 301 Michigan St NE, Suite 268, Grand Rapids, MI 49503 (USA).
- A. Marquis, PT, DPT, ATC, Department of Physical Therapy, College of Health Professions, Grand Valley State University.
- N. Piper, PT, DPT, Department of Physical Therapy, College of Health Professions, Grand Valley State University.
- C. Burdis, PT, DPT, Department of Physical Therapy, College of Health Professions, Grand Valley State University.
- Address all correspondence to Professor Harro at: harroc{at}gvsu.edu.
Abstract
Background Complex movement and balance impairments in people with Parkinson disease (PD) contribute to high fall risk. Comprehensive balance assessment is warranted to identify intrinsic fall risk factors and direct interventions.
Objective The purpose of this study was to examine the psychometric properties of 3 balance measures of a force platform (FP) system in people with PD.
Methods Forty-two community-dwelling individuals with idiopathic PD completed the testing protocol. Test-retest reliability was assessed for the Limits of Stability Test (LOS), Motor Control Test (MCT), and Sensory Organization Test (SOT). Intraclass correlation coefficients (ICC [2,1]) were calculated to determine test-retest reliability and minimal detectable change. Validity was assessed by comparing the FP measures with criterion gait and balance measures using Pearson product moment correlations. Multiple regression analyses examined the contribution of PD characteristics to FP measures.
Results All primary FP variables demonstrated excellent test-retest reliability (ICC=.78–.92). The SOT and LOS demonstrated fair to good correlations with criterion measures, whereas the MCT had fair correlations to balance measures only. Both SOT composite equilibrium and MCT average latency were moderately associated with disease severity.
Limitations This study's sample had a relatively small number of participants with a positive fall history, which may limit the generalizability of the findings.
Conclusions This study's findings provide support that FP measures are reliable and valid tests of balance impairment in people with PD. Disease severity was significantly associated with SOT and MCT measures, perhaps reflecting that these tests are meaningful indicators of decline in postural control with disease progression. Force platform measures may provide valuable quantitative information about underlying balance impairments in people with PD to guide therapeutic interventions for fall risk reduction.
Parkinson disease (PD) is a prevalent neurodegenerative disorder affecting approximately 1 million people in the United States.1 Individuals with PD experience injury-related falls, which decreases their quality of life and contributes to rising health care costs.2,3 The etiology of falls in PD is primarily intrinsic in nature and multifactorial, including bradykinesia, postural control and gait impairments, strength decline, and medication side effects.2–4 To reduce personal and health care costs of falls in people with PD, there has been an increased emphasis in rehabilitation on fall risk assessment and prevention.
Postural control deficits seen with PD progression include impaired postural stability, reduced anticipatory postural control during self-generated movements, and slowed postural responses to a loss of balance.5–8 Effective use of sensory strategies in response to varied environmental demands also is impaired. In contexts where vision is absent or inaccurate, people with PD are at increased fall risk.9–11 Gait decline in people with PD, evidenced by reduced speed, short shuffling steps, forward lean of center of mass, and freezing episodes, also can adversely affect mobility and increase fall risk.2–4 Due to the complexity of underlying balance deficits, a comprehensive balance assessment should incorporate a battery of standardized measures that are sensitive to both these intrinsic impairments and functional balance deficits. Findings from these measures can help guide clinicians to design balance interventions to reduce fall risk. Thus, measuring underlying balance impairment with valid and reliable measurement tools in PD is warranted.
Functional performance measures currently used to assess balance abilities in individuals with PD include the Berg Balance Scale, Timed “Up & Go” Test, and Functional Gait Assessment (FGA).12–15 These measures assess functional balance abilities; however, they are limited in their ability to identify underlying balance impairments in proactive, reactive, and sensory strategies. Additionally, these measures have a documented ceiling effect in the early to middle stages of PD.16–18 There are only a few clinical measures that assess underlying balance impairments.16,19–23 The Balance Evaluation Systems Test (BESTest) assesses 6 different components of the balance system19; however, the clinical utility of this test is problematic, as it takes 40 minutes to administer. The Mini-Balance Evaluation Systems Test (Mini-BESTest), an abbreviated version of the BESTest,16 has established reliability and validity in PD16,20,21 and good accuracy in predicting falls in people with PD.16,20,22 There are limited standardized clinical measures to diagnose underlying balance impairments in PD, and the majority of measures utilize subjective scales for scoring. Computerized force platform (FP) measures provide objective, quantitative data regarding balance impairments and, therefore, may be a valuable component in a battery of tests for assessing balance deficits in the PD population.
The NeuroCom SMART EquiTest Clinical Research System/Balance Master System 9.1 (Natus Medical Inc, Pleasanton, California) is a computerized FP system that is able to detect and quantify balance impairment, including postural sway measures, reactive strategies, sensory strategies, and proactive control of balance.24 Three standardized FP tests of the SMART EquiTest system are: the Sensory Organization Test (SOT), the Limits of Stability Test (LOS), and the Motor Control Test (MCT). The SOT assesses the integration of the vision, somatosensory, and vestibular systems for postural control. Sensory Organization Test composite equilibrium scores in people with PD are significantly lower than in age-matched controls and have been correlated with disease severity.10,25,26 Individuals with PD have impaired use of vestibular input and an overreliance on visual inputs for balance.10 Sensory Organization Test composite equilibrium scores in people with PD are responsive to balance training27 and locomotor training28 interventions. No previous research has examined the reliability and validity of SOT measures in PD, although there is evidence for good test-retest reliability and validity in community-dwelling elderly (intraclass correlation coefficient [ICC]=.66–.93) and stroke (ICC=.81) populations.29–32
The LOS assesses voluntary postural control by measuring the individual's active limits of stability, quantifying movement excursion. Individuals with PD have significantly reduced movement velocity and mean endpoint excursion (EPE) compared with healthy controls,10,33 which may be indicative of increased fall risk.7 There is no research evidence demonstrating whether measures of limits of stability can differentiate people with PD based on disease severity or fall history. Test-retest reliability for limits of stability variables are supported in other neurologic populations,30,34,35 but no studies have examined the reliability of EPE in PD, and only one study33 has examined convergent validity in PD. The MCT utilizes anterior and posterior destabilizing movements of the FP to assess reactive postural control strategies. Individuals in the middle and late stages of PD reportedly have impaired reactive strategies, evidenced by slower and hypometric responses compared with healthy controls.5,7 Similar to the LOS, there is a lack of research on reliability and validity of the MCT in people with PD and the community-dwelling elderly population; therefore, further research is warranted.
There is a gap in the literature regarding the reliability and validity of FP measures in the PD population. Furthermore, research is lacking that examines whether these measures are sensitive to detect balance decline and fall risk with disease progression. The primary purpose of this study was to examine the psychometric properties of 3 balance impairment measures on the NeuroCom FP system in people with PD. Specifically, this study examined the test-retest reliability and convergent validity of the SOT, LOS, and MCT with clinical gait and balance measures. A secondary purpose of the study was to investigate whether disease-related variables contributed to FP test performance. We hypothesized that these FP measures would be reliable assessments of balance impairment in PD and would demonstrate good correlation with clinical balance measures and disease severity. Examining these FP tests' properties provides important information for the use of these measures in both future research and clinical application.
Method
Participants
Community-dwelling individuals with PD were recruited from local support groups, exercise classes, and retirement communities and through the regional chapter of the National Parkinson Foundation. Inclusion criteria were: 20 to 80 years of age, idiopathic PD at Hoehn and Yahr (H&Y) stages I through IV,36 stable PD medication for the last 3 months, functional vision with or without corrective lenses, ability to walk 91.4 m (300 ft) and ascend and descend 6 stairs with or without an assistive device or railing, and no more than close guard assist. Exclusion criteria included: history of other neurologic diagnoses or vestibular pathology per self-report, peripheral neuropathy, dementia, deep brain stimulation, orthopedic injury or surgery within the last 3 months that limits walking or stair skills, and inability to speak or understand English. Participants were screened for inclusion criteria via telephone interview and in person. Peripheral neuropathy was screened using the Semmes-Weinstein monofilament examination, applying established criteria for sensory neuropathy.37 The Montreal Cognitive Assessment was administered to determine whether participants had cognitive deficits that met criteria for dementia,38 utilizing a cutoff score of <21/30 to identify dementia in PD (sensitivity=0.81, specificity=0.95).39 Participants who met the study inclusion criteria completed the informed consent process.
Sixty-eight individuals were recruited; 23 were excluded based on screening criteria, and 3 declined to participate (Fig. 1). Forty-two participants were enrolled in the study (22 men, 20 women) (Tab. 1). One participant did not complete the second testing session due to a medical issue; therefore, this participant's data was used for validity analyses but were omitted for test-retest reliability analysis. The participants' mean age was 66.21 years (SD=7.92), their mean disease duration was 4.49 years (SD=3.15), and their mean disease severity score was 47.98 (SD=23.70) based on the Movement Disorder Society–Unified Parkinson's Disease Rating Scale III (MDS-UPDRS),40 which was administered by the principal investigator (C.C.H.), who was trained by the International Parkinson and Movement Disorder Society. An interview was conducted with participants or spouses, or both, to collect information regarding age, medical history, and fall history. Self-report fall history in the past 6 months was used to classify a participant as a “faller” (>2 falls) versus “nonfaller.” A fall was defined as any instance in which the individual lost his or her balance, resulting in the individual dropping to the ground or hitting an object below. The Freezing of Gait Questionnaire was administered to identify individuals who were experiencing freezing of gait during mobility tasks and categorize to participants as “freezers” versus “nonfreezers.”41 Participants were classified as a freezer if they scored greater than 1 on the third question of the Freezing of Gait Questionnaire. Twenty-four percent (10/42) of the participants were classified as fallers, and 24% were classified as freezers. All participants were community ambulators, and 4 participants utilized assistive devices in the community.
Study flow diagram. PD=Parkinson disease.
Participants' Demographic Informationa
Study Design
This study examined the psychometric properties of the LOS, MCT, and SOT of the NeuroCom system. Test-retest reliability and minimal detectable change were examined by administering these tests twice over a 10-day period. Convergent validity was examined by comparing FP measures with criterion gait measures (10-Meter Walk Test [10MWT] and Six-Minute Walk Test [6MWT]) and clinical balance measures (Mini-BESTest and FGA). Factors predictive of FP measures' performance were analyzed by assessing the contribution of demographic variables (eg, age) and disease-related variables (ie, PD severity and duration, H&Y stage, fall history, and freezing of gait) to the FP measures. Discriminative validity was assessed by analyzing whether there were any differences in FP test performance based on faller/nonfaller, freezer/nonfreezer, and PD subtype groupings.
Testing Procedures
Participants completed a 1.5-hour battery of testing in a quiet university research laboratory setting. All testing was completed during participants' “on time” of their PD medications. Additionally, participants were tested at the same time of the day for each of the 2 testing sessions within scheduling restrictions. The assessment battery was conducted in a standardized testing order, which included 3 testing segments: (1) gait measures (10MWT and 6MWT), (2) clinical balance measures (Mini-BESTest and FGA), and (3) FP measures (LOS, MCT, and SOT). A short rest break was given between testing segments. Tests were administered by 3 trained physical therapist student researchers under direct supervision of the principal investigator. To enhance the consistency in testing methods, each researcher was responsible for administering one of the 3 testing segments for all participants. For each test, standardized testing procedures and instructions were followed, as described below.
Assessment Measures
Gait measures.
The 10MWT is a reliable and valid indicator of walking ability in the PD population for both free and fast walking speeds (ICC=.96 and .97, respectively).42 Participants walked on a tiled, 14-m marked walkway, allowing for acceleration and deceleration, and the middle 10 m was timed. Two consecutive trials for both free and fast gait speeds were timed to the nearest hundredth of a second, and the average of the 2 trials was documented.
For the 6MWT, testing procedures followed the American Thoracic Society guidelines.43 Participants walked on a 30.48-m (100-ft) tiled hallway with tape markers on each end to signal when to turn. This measure has excellent test-retest reliability (ICC=.96)42 in people with PD and good concurrent validity with other gait and balance measures.44
Balance measures.
The Mini-BESTest is a 14-item balance test of postural stability that assesses different components of the balance system.16 The Mini-BESTest has excellent test-retest reliability (ICC=.92) and concurrent validity in people with PD.16,20,21 The FGA is an 11-item test of dynamic balance during varied walking tasks and is sensitive to detect balance decline and fall risk in people with PD.15 This test has excellent test-retest reliability in people with PD (ICC=.91).18 The FGA was performed per test administration guidelines, and standardized instructions were provided. Items on the FGA and Mini-BESTest that overlapped were performed only once and scored in accordance with each test's criteria.
Force platform measures.
Force platform measures were standardized and administered using the NeuroCom system.24 Quantitative assessment of proactive, reactive, and sensory postural strategies was done using the LOS, MCT, and SOT, respectively. Although participants were in a harness for safety, a “fall” during FP tests was defined as any time that the participant took a step, used his or her hands for support, or needed the assistance of the researcher to regain balance.
The LOS assesses voluntary postural control, requiring the individual to move his or her center of gravity to 8 different directional targets. Variables measured in this study were average EPE, average reaction time, and number of falls.
The MCT assesses postural responses to small, medium, and large FP perturbations in the anterior and posterior directions.24 Variables measured in this study were average response latency and average amplitude response to large perturbations. These are sensitive measures of postural reactions in people with PD.5,7
The SOT assesses postural stability and use of sensory strategies under 6 different sensory conditions. An equilibrium score (a measure of postural sway) is assessed for each condition, and scores are compared across conditions to analyze the use of the sensory system for balance.24 Variables measured in this study were composite equilibrium score, vestibular ratio score, and number of falls.
Data Analysis
Descriptive statistics were analyzed for demographic characteristics, disease characteristics, gait, and balance variables. Test-retest reliability of the FP measures was calculated using ICCs (2,1).45 The strength of the ICC values was determined according to Andresen.46 Minimal detectable change (MDC) for the primary FP measures was calculated using the formula: MDC = 1.96 × SEM × √2. The standard error of measurement (SEM) was calculated with the formula: SEM=Sx 
, where Sx is the standard deviation of the set of observed test scores of a group of participants and Rxx is the reliability coefficient (test-retest ICC value) for that measurement.47
Convergent validity of FP measures with the criterion gait and balance measures was analyzed using Pearson product moment correlation coefficients. The primary FP variables for validity analyses were: (1) LOS average EPE, (2) MCT average response latency and average amplitude–large, and (3) SOT composite equilibrium score. Each criterion measure was compared individually with these FP variables. The strength of the correlations was determined in accordance with Portney and Watkins.47 Additionally, subtype group analysis was conducted, using the Wilcoxon rank sum test, to examine whether there were any significant differences in primary FP measures between fallers and nonfallers and between freezers and nonfreezers. Subtype group analysis also was done, using one-way analysis of variance, to determine whether there were any differences in FP measures based on PD subtype. Participants were classified into 3 PD subtype groups: tremor dominant (TD), posture instability/gait difficulty (PIGD), and indeterminate, based on MDS-UPDRS score, as described by Stebbins et al.48 Post hoc analysis was completed using pair-wise independent t tests with adjusted Bonferroni corrections for multiple comparisons. The alpha level was set at P<.50 for all subtype group analyses.
Multiple linear regression analyses, using a backward selection procedure, examined the association between demographic and disease-related variables and FP primary variables and determined which variables were predictive of FP measures. The variables entered in the regression model were: age, fallers versus nonfallers, freezers versus nonfreezers, H&Y stage, disease severity (MDS-UPDRS score), and disease duration. All variables were entered into the model, and a backward selection procedure was used to determine the final model, which retained only those factors that significantly contributed (P<.05) to the model. Partial correlations between significant contributing variables in the regression model and the primary FP measure of interest were then calculated, allowing analysis of the contributing factor while holding the other variables constant. Sample size calculation for the study was based on this multiple linear regression model. It was determined that, with a power of 0.80, a large effect size of 0.35, and an alpha of .05, a sample of 46 participants was needed. Power analysis was computed using G*Power version 3.1.9.2 (Universität Düsseldorf, Düsseldorf, Germany). Data analysis was completed using SAS software version 9.4 (SAS Institute Inc, Cary, North Carolina).
Results
Descriptive statistics for the FP measures and for the clinical gait and balance measures are summarized in Table 2. The mean Mini-BESTest and FGA scores were 21.38 (SD=5.23) and 24.88 (SD=4.81), respectively. Almost 20% of the sample were identified as being at fall risk based on the Mini-BESTest scores (8/42 participants scored <19),16,20,22 whereas only 12% were identified as being at fall risk based on FGA scores (5/42 participants scored <18).15 For the SOT, the mean composite equilibrium score was 67.98 (SD=13.04); one-third of the sample (16/42 participants) scored below age-based normative values, and 17 participants experienced a fall during the sensory organization testing.
Descriptive Statistics of Force Platform, Gait, and Balance Measuresa
All FP variables demonstrated excellent test-retest reliability (ICC range=.78–.92), with the exception of LOS average reaction time and LOS number of falls, which both were moderately reliable (ICC=.69 and .63, respectively) (Tab. 3). The MDC for SOT composite equilibrium, LOS average EPE, and MCT average latency were 11.6, 13.8%, and 7.4 milliseconds, respectively.
Test-Retest Reliability and Minimal Detectable Changea
Fair to good relationships were found between FP measures and criterion measures, with the exception of MCT measures, which demonstrated weak to no relationship with criterion measures (Tab. 4). Moderate to good relationships were found between SOT composite equilibrium score and clinical balance measures and 6MWT, whereas fair relationships were found between LOS EPE and balance and gait measures and between MCT average latency and balance measures. Motor Control Test average latency and average amplitude–large demonstrated no significant relationship with the gait measures.
Convergent Validity of Force Platform Measures With Clinical Gait and Balance Measuresa
Multiple regression analyses determined that age was the only significant variable contributing to LOS EPE (F1,40=15.36, P=.0003), with 26.0% of the variance in LOS EPE explained by age. Both age and disease severity were significant contributing variables to SOT composite equilibrium (F2,39=8.88, P=.0007), with an overall adjusted R2 of 27.8% for this model. Partial correlation analyses revealed that when controlling for age, disease severity was a significant contributor to SOT composite equilibrium (P=.0006, r=−.507) and explained 25.8% of the variance in this score (Fig. 2). Age, disease severity, and disease duration were significant variables contributing to MCT average latency (F3,38=6.74, P=.0009), accounting for 29.6% of the variance in this measure. Partial correlation analyses revealed that disease severity had the strongest association with MCT average latency (P=.003, r=.449), accounting for 19.8% of the variance in this measure.
Scatter plots depicting the inverse relationship between Sensory Organization Test (SOT) composite equilibrium score and disease severity (above) and the direct relationship between Motor Control Test (MCT) average latency and disease severity (below). Severity was based on Movement Disorder Society–Unified Parkinson's Disease Rating Scale.
The Wilcoxon rank sum test revealed no significant difference in SOT composite equilibrium score between fallers (n=10) and nonfallers (n=32) (S statistic=161.50, P=.117; 95% confidence interval [CI] of median score for fallers=49.27, 73.47; 95% CI of median score for nonfallers=66.08, 74.05). Likewise, no significant difference was found in LOS EPE between fallers and nonfallers (S statistic=197.00, P=.605; 95% CI of median score for fallers=56.91, 78.54; 95% CI of median score for nonfallers=66.14, 75.85). In contrast, a statistically significant difference was found in MCT average latency (S statistic=291.00, P=.025), as the fallers had significantly slower response latencies than the nonfallers (95% CI of median score for fallers=142.16, 156.04 milliseconds; 95% CI of median score for nonfallers=140.01, 146.30 milliseconds). No significant differences in any of the FP measures were found between freezers (n=11) and nonfreezers (n=31).
The analysis of variance revealed a statistically significant difference in SOT composite equilibrium among the 3 PD subtype groups (F2,39=3.27, P=.048). Post hoc analysis demonstrated that the TD group had significantly higher composite equilibrium scores that the PIGD group (P=.50; 95% CI difference=1.52, 19.42), whereas no differences were found between the indeterminate group and the other 2 groups. Significant differences also were found for the LOS EPE among the 3 PD subtype groups (F2,39=3.40, P=.044); post hoc analysis revealed that the TD group had higher EPE values compared with the PIGD group, approaching significance (P=.052; 95% CI difference=1.67, 20.49). A significant between-group difference also was found for MCT average latency (F2,39=4.4, P=.023); post hoc analysis revealed that the PIGD group had significantly slower latency than the TD group (P=.019; 95% CI difference=2.33, 14.61 milliseconds). No difference in MCT average amplitude–large was found among the PD subtype groups.
Discussion
Our results support the premise that FP measures provide reliable, objective information regarding balance impairments in people with PD. Fair to good correlations were found between FP measures and clinical balance and gait measures, supporting the relationship between balance impairments and functional balance skills. Additionally, disease severity was moderately associated with SOT and MCT variables, perhaps reflecting that these measures may be sensitive indicators of decline in postural control with disease progression.
Reliability and Minimal Detectable Change
Excellent test-retest reliability was found for all primary FP measures, with the MCT variables demonstrating the highest reliability (ICC=.92). This is the first study, to our knowledge, to report on the reliability of MCT in the PD population. Reliability findings for SOT variables were higher than those reported in previous studies in people poststroke and in the community-dwelling elderly population,29,30 whereas reliability findings for LOS EPE were similar to those of previous research in healthy adults.49 High reliability findings may be attributed to the our research protocol, which included a consistent test administrator and standardized instructions for introduction of each FP test and for test administration. Additionally, medication effects were controlled for by requiring participants to be on the “on time” of their PD medication for both testing sessions.
Our study is also the first study, to our knowledge, to report MDC values for FP measures, with the exception of the study by Wrisley et al,50 who reported an MDC of 8 for SOT composite equilibrium in healthy young adults. In comparison, the MDC for this measure in our PD cohort was 11.64, which may have been due to this sample's relatively high variability in SOT performance. The MDC for LOS average EPE (13.8%) was moderately large and may be reflective of participants exploring different strategies to reach targets during the 2 testing sessions. In contrast, the MDC for MCT average latency was quite small (7.43 milliseconds), perhaps evident of the highly reliable, reactive postural control responses used. Minimal detectable change values can assist the clinician to assess whether true change has occurred following directed balance interventions.
Convergent Validity
Evidence is lacking in previous literature to support the relationship between FP measures and clinical balance and gait measures in people with PD, with the exception of LOS and temporal-distance gait measures.33 We found fair to good correlations between FP measures and criterion gait and balance measures. Force platform measures correlated higher with balance than gait measures, which was expected, as these measures assess similar balance constructs. The SOT had the strongest correlation with clinical balance measures, most notably with the Mini-BESTest. This relationship may be explained by the Mini-BESTest design, which assesses multiple components of balance, including sensory organization. In further analysis of this relationship, a significant direct correlation was found between the Mini-BESTest sensory organization items (2 items) and the SOT composite equilibrium score (r=.695). The fair correlation found between LOS EPE and clinical balance measures lends support for the relationship between proactive balance strategies utilized when moving to LOS with functional balance abilities during standing and walking.33 In contrast, the MCT demonstrated a weaker correlation with clinical balance measures and was not significantly correlated with gait measures. This finding may indicate that MCT and clinical balance tests assess different aspects of postural control.
One subsection of the Mini-BESTest assesses reactive postural control similarly to the MCT; however, it utilizes “push-release” subtests that are completed in a manner that brings the individual outside of their limits of stability to deliberately elicit a stepping strategy.16 In further analysis of this relationship, no significant correlation was found between the summative score on the Mini-BESTest reactive balance subtests (×4 items) and MCT average latency (r=−.257) or MCT average amplitude–large (r=−.144). The “push-release” Mini-BESTest subtests are prone to testing administration and reliability errors and provide subjective, qualitative information on postural responses. In contrast, the MCT is a highly sensitive and reliable quantitative measure of postural reactions to assess potential fall risk.
Our study's findings support the premise that FP measures provide valuable information about balance impairments in individuals with PD. The SOT is considered a diagnostic test of sensory integration balance deficits in other clinical populations.29,30,32 Limited research has examined this component of postural control in PD. In this study, approximately one-third of the sample had composite equilibrium scores that fell below age-matched normative values, and 17 participants experienced a fall during sensory organization testing. These findings reflect deficits in sensory integration for effective balance in varied environmental conditions, which is consistent with previous research in PD cohorts.10,25,26 Almost one-third of the participants (15/42) scored below normative values on the vestibular ratio subscore, indicating ineffective use of vestibular cues for postural control, which is consistent with the findings of Rossi et al.10 These deficits in sensory integration may contribute to increased fall risk under sensory conflict conditions. Participants in the PIGD group had significantly lower SOT composite equilibrium scores compared with the TD group, reflecting the sensitivity of this test to identify those individuals with postural control deficits. There was a significant association between PD severity and SOT composite equilibrium, which is in agreement with previous research.10,25,26 The SOT findings may help guide the clinician in designing balance interventions to remediate this deficit and reduce fall risk.
Our findings are consistent with previous research that showed reduced posterior limits of stability in people with PD and anteriorly displaced center of mass during FP testing.5,8 Although the mean number of falls during the LOS was small (0.43), 78.1% of the falls occurred in attempts to reach posterior targets. It is notable that many participants lacked efficient proactive postural strategies to move their center of mass to positional targets, particularly to posterior targets. Participants in the PIGD group had reduced EPE compared with the TD group, lending support to the validity of this test to detect voluntary postural control impairments. The sample's mean EPE was 70.2%, which is consistent with EPE values previously reported in community-dwelling elderly fallers.51 Slowed postural responses have been documented in previous research in people with PD,5,7 whereas only 2 out of 42 participants in the current study had an average latency below normative values, and no falls occurred during the MCT. However, it is notable that disease severity and duration were significantly associated with MCT average latency, with slower responses as severity increased. Additionally, the PIGD group had significantly slower response latencies compared with the TD group, lending support for the discriminative validity of this measure. In this PD cohort, however, no significant differences in the FP measures were found between fallers and nonfallers or between freezers and nonfreezers. The relatively small sample distribution of fallers and freezers may have contributed to this finding.
Clinical Implications
Individuals with PD have multiple intrinsic risk factors that contribute to falls, especially with disease progression. A battery of measures is warranted for balance assessment at both functional and impairment levels to diagnose balance deficits.12 Force platform measures provide objective, quantitative data regarding balance impairments. This information may guide the clinician in designing targeted balance interventions to reduce fall risk and provide sensitive measures to assess treatment effectiveness. Our strong reliability findings support the clinical utility of FP measures to provide a stable baseline measure of balance. Disease severity was significantly associated with both SOT composite equilibrium and MCT average latency. These findings are clinically noteworthy, as they indicate that these FP measures may add valuable objective information regarding the disease progression and the underlying balance deficits that affect safe mobility. The SOT had good clinical utility for identifying sensory organization balance deficits in this PD cohort. Clinicians should consider incorporating the SOT into their balance assessment battery to help identify individuals with these deficits who may be at increased fall risk. The LOS is a sensitive measure of an individual's voluntary postural control. Evaluation of LOS EPE in the backward direction is particularly important in individuals with PD because deficits in this area are highly associated with falls.7
Despite the strong psychometric findings in this study supporting the use of FP measures in people with PD, these tests do require specialized, costly equipment and typically take 20 to 30 minutes to administer. Therefore, clinicians may want to consider a staged assessment approach: first conducting functional performance balance measures, and then if specific underlying balance impairments are suspected from these test findings, progressing to a specific battery of FP measures. Alternatively, the FP measures could serve as a part of an annual physical examination for people with PD, monitoring for any change in balance function.
There were several possible limitations in this study. There was a moderate sample size, with relatively small numbers of fallers and freezers; therefore, the results should be interpreted with caution. The exclusion of individuals with dementia and other neurologic or vestibular diagnoses limits the generalizability of this study's findings. Although test-retest reliability was resilient to confounding factors, it should be noted that fatigue from the extensive battery of testing may be a limitation in this study's protocol, as some participants reported fatigue during the first test session. Additionally, we did not control for other physical activity or exercise that participants engaged in on testing days.
To our knowledge, this was the first study to explore test psychometric properties of FP balance measures in individuals with PD; therefore, these findings provide foundational information for clinical research and to guide future studies. Further research should compare FP tests' psychometric performance between people with PD and age-matched healthy adults. Future studies should examine the sensitivity and predictive validity of FP measures to assess fall risk in people with PD. Research also is needed to examine the responsiveness of FP measures to assess changes in balance following therapeutic interventions.
This study's findings provide evidence to support the use of FP measures for evaluation of balance impairments in individuals with PD. The SOT, LOS, and MCT demonstrated excellent test-retest reliability in an ambulatory PD cohort, and MDC values were reported. The SOT and LOS had fair to good correlation with criterion gait and clinical balance measures, whereas the MCT was weakly correlated with balance measures and had no relationship with gait measures. Disease severity was significantly associated with SOT composite equilibrium score and MCT average latency. Balance decline and increased fall risk are significant health concerns in people with PD. A battery of balance measures is warranted to accurately assess balance deficits at both the functional and impairment levels. Force platform measures may be a valuable addition to this test battery to provide diagnostic information about underlying balance impairments in people with PD, which can guide therapeutic interventions for fall risk reduction.
Footnotes
All authors provided concept/idea/research design, writing, and data collection and analysis. Professor Harro provided project management, participants, and facilities/equipment.
The authors thank Dr Sango Otieno, Department of Statistics, Grand Valley State University, for his expertise and collaboration in the statistical analyses for this study.
This study was approved by the Institutional Review Board at Grand Valley State University.
- Received March 4, 2016.
- Accepted July 7, 2016.
- © 2016 American Physical Therapy Association