| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Research Reports |
A Nieuwboer, PT, PhD, is Researcher, Neuromotor Unit, Department of Rehabilitation Science, Faculty of Physical Education and Physiotherapy, University of Leuven, Tervuursevest 101, 3001 Heverlee, Belgium (alice.nieuwboer{at}flok.kuleuven.ac.be). Address all correspondence to Dr Nieuwboer
W De Weerdt, PT, PhD, Professor, Department of Rehabilitation Science, Faculty of Physical Education and Physiotherapy, University of Leuven
R Dom, MD, is Neurologist and Professor, Department of Neuroscience and Psychiatry, Faculty of Medicine, University of Leuven
K Bogaerts, MSc, is Statistical Consultant, Biostatistical Centre of Clinical Trials, Faculty of Medicine, University of Leuven
G Nuyens, PT, PhD, is Research Assistant, Neuromotor Unit, Department of Rehabilitation Science, Faculty of Physical Education and Physiotherapy, University of Leuven
Submitted April 20, 2000;
Accepted July 11, 2000
 | Abstract |
|---|
=11.7 years, SD=4.9, range=6–22) took part in the study. Methods. The subjects' performance on a 10-item scale was videotaped. The videotapes were then scored by 2 independent raters, and the scores were used to determine interrater reliability. The stability of 6 repeated measurements was examined in the home situation, taking into account self-rated fluctuations of motor performance. Results. Weighted Kappa values of agreement (.86–.98) confirmed the reliability between testers. Measurement during the "on" phase (when medication was working optimally) and the "off" phase (when the action of medication was strongly decreased or absent) led to different measurements. Measuring frequently within "on" and "off" phases gave relatively stable measurements for total function, bed transfers, and gait akinesia, the latter during the "off" phase only (intraclass correlation coefficients [ICCs]=.70–.93). However, more modest repeatability applied to transfers from a chair (ICC=.65–.67). Conclusion and Discussion. To ensure valid results in future effect studies, clinical differentiation between "on" and "off" phase measurements is proposed on the basis of patients' own perception of their medication status.
Key Words: Disability • Motor fluctuation • Parkinson disease • Rating scale • Reliability
| Introduction |
|---|
 TopAbstract Introduction MethodResultsDiscussionConclusionAppendixReferences |
|---|
Currently available clinical scales used for the evaluation of people with PD, such as the Unified Parkinson's Disease Rating Scale (UPDRS)6 and the Hoehn and Yahr Scale7 designed for people in all stages of the disease, appear to us to be essentially global measures covering components of disability as well as impairment. Although the reliability and validity of measurements obtained with the UPDRS are well-established, the scale includes several items that we believe are not very relevant to physical therapy (eg, items on speech, facial expression, salivation, swallowing).8–10 Moreover, the relatively long time it takes to obtain the measurements (the motor examination part alone includes 27 ratings), in our view, makes the UPDRS impractical for patients who are prone to sudden fluctuations of dopaminergic status. Therefore, we developed a 10-item scale to characterize the functional problems of people who are in the middle and advanced stages of PD.
Reliability of measurements of motor performance in people with PD has been studied mainly in patients in the early and middle Hoehn and Yahr stages.11–13 In these people, measurements of musculoskeletal impairments as well as timed functional tests such as turning 360 degrees and walking and rising from a supine position to a standing position were found to be sufficiently repeatable for retesting after several days.11 Similarly, reliability was evaluated for a number of balance tests during the peak-dose period in patients with and without a history of falls.12 Both immediate test repetitions of spatiotemporal gait measures (conducted every 15 minutes) and delayed test repetitions of spatiotemporal gait measures (conducted after a 24-hour interval) resulted in stable measurements of walking speed, cadence, stride length, and double support duration, provided that patients were tested in the "on" phase and had a relatively short disease duration (mean=8.3 years).13 However, O'Sullivan et al14 and MacKay-Lyons15 found that measurements of gait-related variables obtained in people with chronic PD (mean disease durations of 13 and 15 years, respectively) differed in major ways when comparing "on," "wearing off," and "off" conditions. The intrasubject variability, in our opinion, illustrates a dramatic and inconsistent effect of medication on gait performance.
In summary, it seems that measuring outside of the periods in which medication is working optimally leads to greater variability in measurements. This is not considering the confounding influence of "peak-dose" dyskinesias, which are most likely to occur at the time of a presumed stable effect of medication on functional or motor performance (or mobility). As most researchers have addressed either impairment variables or isolated measures of motor performance, the question for us remains whether the performance of functional activities can be measured reliably in the context of L-dopa–induced variability. Therefore, we conducted 2 related studies. In study 1, we examined the intertester reliability and internal consistency of measurements obtained with a newly developed activity scale. In study 2, we examined whether randomly occurring fluctuations ("on-off" and dyskinesias), as rated by the patients themselves, led to differences in functional performance. Functional performance refers to motor activities such as basic transfer movements and walking. We also investigated the test-retest stability of 6 consecutive measurements during one day within the self-rated "on" and "off" phases.
| Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendixReferences |
|---|
|
The items of the Parkinson Activity Scale (PAS) are described in the Appendix (page 1096) and are divided into 4 categories: chair transfers, gait akinesia, bed mobility, and bed mobility with cover. A 5-point scoring system (0–4) was devised, corresponding to that used in the UPDRS. The maximum score indicated test performance without difficulties. Carrying out the overall task more slowly than people without health problems was seen by us as without difficulty in the context of chronic disease. A minimum score implied that the subject required physical help. In-between scores (1–3), usually expressed in nonexplicit terminology such as "mild," "moderate," and "severe," were defined according to the degree of movement efficiency and to the extent to which an adequate end position was reached. In the case of bed movements, grades represented the number of problems occurring in these areas.
Procedure
Study 1.
Two physical therapists, one (aged 23 years) with 1 year of experience in neurology and one (aged 38 years) with 10 years of experience in neurology, participated in the interrater reliability study. The subjects carried out the test items of the PAS while being videotaped. Following completion of testing, the 2 raters scored the videotapes independently of each other (initial reliability test). Six weeks after improving the definitions of the scoring criteria, the procedure was repeated (final reliability test). The subjects performed the tests mostly during the "on" phase and in 3 cases during the "off" phase.
Study 2.
In this study, "on-off" variability and test-retest repeatability of the scale were examined. The junior therapist involved in study 1 tested the same subjects at their home on 6 occasions dispersed over 2 medication cycles within one day. The first test was carried out 90 minutes after the subject took the first morning dose of medication. This test was followed by 2 additional evaluations repeated at 1-hour intervals. The hour between repetitions was timed from the beginning of the previous test. The same test sequence was used in the afternoon starting 90 minutes after the midday dose. The subjects took their medication as normal. In an effort to minimize variability, we used the PAS in a standardized fashion adapted to each home situation. For chair rise (item 1), an armchair with a seat height of between 40 and 45 cm was chosen. The order of test categories was randomized so that fatigue or stiffness would not influence performance. On average, 20 minutes (range= 15–28) was needed to obtain the measurement. Missing values were generated on 2 occasions during the sixth test due to fatigue.
Additional scores were given to record the occurrence of "on-off" fluctuations of motor performance during the 6 measurements. Because the "on" and "off" phases are acknowledged to be highly individual and sometimes occur randomly, describing a variety of conditions for different patients, a self-rating procedure was used in agreement with the Core Assessment Program for Surgical Interventional Therapies (CAPSIT) recommendations.28 Our purpose in adopting a self-monitoring method was our belief that it would allow discriminating between "on" and "off" phases and thus controlling for the confounding effects of medication independent of time. Therefore, the "on-off" score (0–2) represented the subjects' own evaluation of their response to medication, in agreement with the researcher, using their normally experienced fluctuations as a point of reference.
Before recording the occurrence of "on-off" fluctuations of motor performance, we explained the operational definitions of the "on-off" phenomena to the subjects. We defined the "on" phase (2) as the typical stable level of motor performance present at peak dose, when the action of medication was considered optimal. We defined the "off" phase (0) as the typical stable level of motor performance at the end of dose, when the action of medication was strongly decreased or absent. We defined "wearing off" (1) as the unstable period in which a patient changes from the "on" phase to the "off" phase. Similarly, the subjects monitored the presence and severity of dyskinesias as absent (2), mild, without interference with the motor tasks involved (1), or moderate, interfering with the efficiency of performance (0). In our descending scoring scale, the highest score (2) indicates the best condition for the patient, and the lowest score (0) indicates the worst condition for the patient. These variables were scored after the measurement of the PAS items so as not to bias the observer.
Data Analysis
Study 1.
Interrater reliability for each item was calculated using the weighted Kappa coefficient29 (Kw), which is appropriate for ordinal data. A Cronbach alpha and item total correlations were used to assess the internal consistency of the PAS total score (range= 0–40) and the 4 subscale scores: chair transfers (range=0–8), gait akinesia (range=0–8), bed mobility (range=0–12), and bed mobility with cover (range= 0–12).
Study 2.
Subsequent to the analysis of internal consistency, our statistical analysis considered the 4 summed subscale scores and the total score as continuous data, requiring the use of parametric statistics. The gait akinesia subscale scores were not normally distributed, which led us to perform a logarithmic transformation. To explore the test-retest reliability of the 6 measurements obtained during one day, a linear mixed model30 was used, enabling correction for missing values and confounding variables such as age, sex, and disease duration. The separate estimation of regression lines for each subject's "on" and "off" phases allowed for the random occurrence of fluctuations within the subjects (also for having dyskinesia). The "wearing-off" phase was not included in the analysis because of its very infrequent occurrence. The significance and size of the difference between "on" and "off" phase performance and having dyskinesias were analyzed. We then calculated separate test-retest reliability values for the "on" and "off" phase status. We used an intraclass correlation coefficient (ICC), the ratio of the between-subject variability over the total variability. Total variability was the sum of the within-subject variability (the variability between the subject-specific intercepts in the "on" or "off" phase) and the between-subject variability (the measurement error in the "on" or "off" phase).32 The formula for the ICC is comparable to the one described by Shrout and Fleiss,31 but it differs in that it is based on PROC MIXED,32 an SAS procedure that can incorporate unequal numbers of repetitions by fitting a regression model for each subject and combining the results of all subjects. As a measure of the size of variation within the "on" or "off" phase, the square root of the within-subject variability was computed as a percentage of the range of scores.
We examined the possibility of a fatiguing effect (systematically descending slope) or a training effect (systematically ascending slope) over the 6 occasions for the "on" phase status. Using the linear mixed method, we tested whether the slopes differed significantly from zero.
| Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendixReferences |
|---|
|
Study 2
Descriptive data.
Thirteen subjects (44.8%) were without "on-off" fluctuations of motor performance during the home assessments. Three subjects (10.3%) who reported having predictable fluctuations at the onset of the study experienced unexpected "off" phases. Of all the evaluations carried out, 78.2% (136) were within an "on" phase, 17.2% (30) were within an "off" phase, and 4.6% (8) were within a "wearing-off" phase. Dyskinesias were absent in 75.9% (132) of the assessments, mild to moderate in 20.1% (35) of the assessment, and functionally significant (ie, having a score of 0, indicating moderate dyskinesias interfering with movement) in 4% (7) of the assessments. In one subject, mild dyskinesias were observed during an "off" phase.
Impact of fluctuations on movements.
Analyzing the effects of "on-off" variation and dyskinesias on the data, we found that movements were different during the perceived "on" and "off" phases (Tab. 3). Mean differences, expressed as a percentage of the range of scores, varied from 48.2% for bed mobility with handling the cover and 29.6% for transfers from the chair. In contrast, dyskinesias did not have an effect on the movements, except for bed mobility with cover where a mean difference of –10.8% was found (P=.05). The negativity of the figure illustrated the fact that experiencing dyskinesias was associated with improved performance.
|
|
| Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendixReferences |
|---|
We believe that the most important finding of our study was that, when measuring repeatedly within either the self-reported "on" or "off" phase, all but one aspect of the movements studied provided relatively consistent results. High temporal repeatability was previously ascertained on measures of gait, balance, musculoskeletal impairment, and timed tests in patients with mild to moderate PD in the "on" phase.11–13 We add to these findings that acceptable reproducibility also could be achieved in patients with considerable disease duration when measuring transfer movements and walking in the home several times. Test-retest reliability was similar within the "off" and "on" phases. There was a large difference in the number of evaluations performed in the "off" phase (30) compared with the "on" phase (136). Although this difference did not pose a statistical problem, it implied that the reliability estimates calculated for the "off" phase were less accurate than for the "on" phase. The large between-subject variability observed for the total PAS score and bed mobility in the "off" phase was probably responsible for the fact that in these cases the consistency in the "off" phase exceeded that in the "on" phase. More variable results might have been obtained by inclusion of the "wearing-off" phase. The poor result (ICC=.41) for gait initiation and turning (gait akinesia) during the "on" phase may be explained by the skewed distribution of scores not fully ameliorated with transformation. This raises the question of whether it is relevant to test this area of mobility during an "on" phase. Despite the fact that 75.9% of subjects reported having regular "freezing" or starting difficulties, these problems were difficult to provoke consistently in a standardized test situation even within a familiar environment. Earlier work emphasized the susceptibility of freezing to psychological factors, in that the higher levels of arousal and attention inherent to a test situation might prevent it from occurring.23,34
Having demonstrated overall satisfactory standards of test-retest reliability, the magnitude of within-subject variability remained considerable within "on" and "off" phases, especially for the subcategory scores (8.5%–15.2%). The fact that this variation was smaller for the total scores might be explained by the fact that fluctuations occurred randomly in one area of function and were cancelled out by stability in other areas. Recently, high degrees of variability in gait measurements within each "on" and "off" phase were reported in patients with longer disease duration than in our study, leading to the suggestion that measures of general functional performance might produce more stable results.15 Our results would favor this contention.
Our results showed differences in performance of activities between self-reported "on" and "off" phases. A recent study by O'Sullivan et al14 demonstrated similar swings of "on-off" responses in relation to gait variables, timed tests, and tapping counts following overnight withdrawal of antiparkinson medication in patients with established motor fluctuations. Mean "on-off" changes demonstrated in O'Sullivan and colleagues' study varied between 29.3% and 35.4%, percentages that overlap the range of between 30% and 48% in our study. We found that the occurrence of mild to moderate dyskinesias did not have an effect on performance of activities except on double task performance (bed mobility with cover), illustrating the activating effect of medication. Our results, however, were obtained while we excluded patients with severe dyskinesias.
Our results highlight the importance of implementing strategies to deal with the problem of confounding drug effects in clinical trials. Such methods might include (1) adopting rigorous inclusion criteria, (2) testing patients at the same time of day within the therapeutic window, and (3) carrying out repeated measurements within the same day as well as over several days to gain a valid picture of mean performance. Yet, all of these measures have a number of practical and validity problems.
Excluding patients with dyskinesias and "on-off" fluctuations of motor performance not only may seriously reduce the number of patients available for study but also may introduce selection bias, diminishing the clinical relevance of results. Notwithstanding its feasibility, measuring patients at the same time of day requires rigorous control of patients living in the community. Lack of adherence to the drug regimen as well as daily variations in response to medication make this method far from ideal. Disappointing results on adherence were recently published on a group of individuals with moderate depression but no other health problems, a large proportion (57%) of whom adhered only 50% of the time to a once-a-day drug regimen at home.35 These findings may be extrapolated to people with advanced PD, who are characterized by a high incidence of depression36 and a long-term requirement of multiple doses. Finally, carrying out repeated measurements may create an intolerable burden on elderly people with disabilities, who often experience problems of transportation, despondency, and fatigue.
Although we found that the most likely source of variation is due to L-dopa status, other reasons for within-subject variability were also explored in this study. Even though no evidence of fatigue was apparent, 2 subjects were not able to complete the sixth test due to exhaustion. Fatigue is a common complaint in almost 50% of people with PD,37 a factor that needs to be managed carefully when planning repeated measurements. We found no training effect in our study, which is not surprising given that the test items consisted of well-learned daily activities.
Our results point to an additional strategy to control for fluctuations of motor performance. Subjects' self-perceptions of being in an "on" or "off" phase were used as a tool to document and thus control for their changing L-dopa status at regular intervals during the test procedure. Subjects evaluated their responses to medication using their normal fluctuations as a point of reference, a method also recommended in the CAPSIT.28
We believe that it makes sense to ask people with long-standing disease, who are sensitive and aware of fluctuations as part of their daily experience, to distinguish between their "on" and "off" phases. We believe that the validity of measurements obtained with this method was verified by the fact that similar levels of "on-off" differences, imposed by withdrawal of medication, were established on a comparable population, but on different variables.14 Other researchers38 have suggested that self-reports of disability from patients with PD can be accurate when compared with reports from their spouses, even in the presence of depression and mild cognitive decline. Furthermore, in the CAPSIT recommendations,28 an "off" phase was operationally defined as the standardized condition observed after 12 hours after withdrawal of medication. Interfering with medication regimens was not feasible in the context of a home-based study and, more importantly, did not concur with our aim of testing for reliability against the background of naturally rather than artificially occurring fluctuations of motor performance. Further study is needed to confirm the validity of self-determinations of "on-off" status as a tool for improving reliability.
Our results not only illustrate that fluctuations have implications for measurement but also underscore the impact of these side effects for patients and clinical practice. The estimated "on-off" swings of function ranged between 30% and 48%, results that cannot be generalized to all patients with middle- and late-stage PD because patients with severe dyskinesias and unpredictable "on" and "off" phases were excluded. However, the participants represented a clinically relevant population of "possible candidates" for rehabilitation taking medication regimens that are common in advanced disease. The largest size of "off" phase deterioration was found for bed mobility with a blanket, often indicating total dependence on physical assistance.
| Conclusion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendixReferences |
|---|
| Appendix |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendixReferences |
|---|
|
| Footnotes |
|---|
This work was supported by the Research Council of the Katholieke Universiteit Leuven and the Fund of Scientific Research-Flanders.
| References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendixReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Tyson and L. Connell The psychometric properties and clinical utility of measures of walking and mobility in neurological conditions: a systematic review Clinical Rehabilitation, November 1, 2009; 23(11): 1018 - 1033. [Abstract] [PDF]
|
|
|
|
|
|
A. Nieuwboer, W. De Weerdt, R. Dom, and K. Bogaerts Prediction of outcome of physiotherapy in advanced Parkinson's disease Clinical Rehabilitation, August 1, 2002; 16(8): 886 - 893. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
