PHYS THER
Vol. 88, No. 12, December 2008, pp. 1554-1567
DOI: 10.2522/ptj.20070186
Hierarchical Properties of the Motor Function Sections of the Fugl-Meyer Assessment Scale for People After Stroke: A Retrospective Study
J Lesley Crow and
Barbara C Harmeling-van der Wel
JL Crow, MSc, MCSP, DipTP, is Research Physiotherapist, Physiotherapy Department Ha104, Erasmus MC, University Hospital Rotterdam, Post Box 2040, 3000 CA Rotterdam, The Netherlands
BC Harmeling-van der Wel, Diploma of Physiotherapy, is Physiotherapist, Physiotherapy Department Ha104, Erasmus MC, University Hospital Rotterdam
Address all correspondence to Ms Crow at: berg.crow{at}hetnet.nl
Submitted June 28, 2007;
Accepted July 30, 2008
 |
Abstract
|
|---|
Background and Purpose: The upper-extremity (UE) and lower-extremity (LE) sections (excluding balance) of the motor function domain of the Fugl-Meyer (FM) assessment scale (a construct referred to here as the FM motor scale) are recognized as a robust part of the scale for use with people after stroke. However, it is frequently criticized as a lengthy and time-consuming measurement tool. The aims of this study were to support a shortened method of administration for the FM motor scale and to provide arguments for the use of a summed score. In pursuit of these aims, the hierarchical properties of both the UE and LE sections of the FM motor scale were investigated.
Participants and Methods: A retrospective analysis of data from 62 people with a previous stroke was performed. Guttman scale analysis considered the hierarchy of items within each subsection and each stage, between subsections and stages, and across all of the scale items (ignoring the stage divisions) of the FM motor scale.
Results: For the within-stage and subsection analyses and between-stage and subsection analyses, all of the results met or exceeded the acceptable levels for the coefficient of reproducibility and the coefficient of scalability. When stage divisions were ignored, the coefficient of reproducibility for both extremities was just below acceptable levels.
Discussion and Conclusion: The results support the use of the UE and LE sections of the FM motor scale as a stagewise and subsectionwise hierarchical assessment and outcome measure. This allows the use of a shortened method of administration, which can potentially reduce the time required for scale administration, and appropriate scores can be allocated for untested items, such that a legitimate total summed score can be used. A limitation of this study was that the study population consisted predominantly of older people with such severe disability that they were unable to function independently.
|
Introduction
|
|---|
With the continuing emphasis on evidence-based practice in health care, the need for accurate tools to measure progress and outcome is increasingly more pertinent. For people recovering from stroke, the major consideration is the testing of motor ability. For this purpose, the Fugl-Meyer (FM) assessment scale1 (also known as the Brunnstrom Fugl-Meyer test) is an instrument commonly administered by physical therapists (and others) in both clinical and research fields to evaluate people after stroke. The original scale, which consisted of 4 domains (motor function, sensation, joint mobility, and pain), has undergone rigorous investigations for reliability, validity, and responsiveness for change.2,3 However, the lack of robustness of the latter 3 domains has resulted in the predominant use of just the motor function domain of the scale. Together, the upper-extremity (UE) and lower-extremity (LE) sections of the motor function domain4,5 are recognized as a robust part of the scale.
The UE section has 4 subsections: shoulder-arm, wrist, hand, and coordination. The LE section has 2 subsections: leg and coordination. In the original FM assessment scale, a balance subsection was grouped in the LE section of the motor function domain. However, because the reproducibility and validity of this subsection have been questioned,5 it is now seldom administered. Furthermore, the recovery of balance after stroke is based on an underlying construct different from that of motor recovery, which underlies the other subsections of the UE and LE sections. Therefore, this article focuses on the UE and LE sections of the motor function domain, excluding balance; this construct is referred to here as the FM motor scale.
For both the UE and the LE sections of the FM motor scale, the progression of recovery is based on the stereotypical stages of recovery after stroke, as originally reported by Twitchell6 and Brunnstrom.7 Twitchell described an ordered, predictable, stepwise progression of the restoration of motor function in the arm and leg after stroke, from the initial stage of flaccid paralysis to the development of a basic stereotypical synergy of voluntary movements and then to normal patterns of voluntary movements. On the basis of these observations, Brunnstrom divided the stepwise progression into 6 sequential stages of motor recovery. These theories, along with other poststroke treatment philosophies,4 provided the underlying construct for the FM motor scale.
Fugl-Meyer et al1 used only 5 of the 6 stages defined by Brunnstrom,7 because in the initial flaccid paralysis stage, no voluntary movements or reflexes would be recorded. They created a hypothesized rank ordering of difficulty, by generating scale items from the easiest to the most difficult to perform, within each stage and each subsection. They emphasized that progression from one stage to the next began when motor recovery of the previous stage reached at least 60% in the shoulder-arm and 80% in the leg.1 Both Brunnstrom and Fugl-Meyer et al described the recovery of wrist and hand functions as independent of the stages of shoulder-arm recovery; therefore, wrist, hand, and shoulder-arm are listed separately as subsections. The testing of coordination for both the UE and the LE by Brunnstrom required minimal specific movements of the wrist and hand. Therefore, Brunnstrom incorporated coordination in the final, sixth stage for each extremity. However, Fugl-Meyer et al treated coordination as a separate subsection, probably because of the specific wrist and hand movements required to perform the task for the UE. Figure 1 compares the stage definitions used by Twitchell,6 Brunnstrom, and Fugl-Meyer et al and demonstrates the alignment of the stages with the stepwise progression of recovery for both the UE and the LE.
View larger version (48K):
[in this window]
[in a new window]
|
Figure 1. Comparison of sequence of stepwise recovery described by Twitchell6 and Brunnstrom7 with the stages and scale items used by Fugl-Meyer et al.1
|
|
If the FM motor scale has a true hierarchy, then it has the advantage that, with some prior knowledge of a patient's ability, the assessor does not have to start with the first item of an assessment scale but can begin the testing at a suitable item of the scale that the assessor believes the patient can achieve. If a patient passes the selected item of the scale, pass scores can be awarded for the preceding scale items. If a patient subsequently fails 2 or 3 consecutive items, the assessment can be stopped with the knowledge that the patient will fail the remaining items of the scale.8,9 The administration of a smaller number of scale items is less demanding for the patient and examiner and may reduce the time needed to complete testing. However, when an assessment measure is not fully administered, the results can be invalidated.10,11 Therefore, showing that such a ranked scale has acceptable hierarchical properties (ie, reproducibility and scalability) would permit a reduction in the number of scale items that must be tested. Because all of the scale items are considered to be relevant at some stage of a patient's recovery, item redundancy is not our interest. Nevertheless, having the flexibility to choose which scale items need to be administered for a patient, depending on that patient's level of recovery, may shorten the time required for administration.
Scores on the FM motor scale are frequently summed to provide a total score per limb or side of the body. The total numerical value indicates a patient's level of motor function. However, the summing of ordinal scores (even if ranked) is considered inappropriate by some and may mask discrepancies in abilities because of the lack of consistency in intervals between items.12 Establishing a hierarchy helps to legitimize the use of a summed score because the rank ordering of scale items is confirmed. The scale items not tested still require the arbitrary allocation of a score, which is based on performance on other items of the scale. Easier or more difficult scale items need not be tested, but the appropriate number of points (2 or 0) needs to be allocated for the untested scale items.
To date, few publications have investigated whether the rank ordering within and between stages and subsections of the FM motor scale is hierarchical, that is, from the easiest to the most difficult to perform. Since the completion of our analysis, Woodbury et al13 have published an investigation into the unidimensionality and the construct validity of the UE section of the FM motor scale. The resulting Rasch analysis of item difficulty (hierarchical) order was not consistent with the expected scale item order of Fugl-Meyer et al1 (see "Discussion" section for more details). However, the study of Woodbury et al assumed a linear continuum, which is very different from the stepwise progression that is part of the philosophy underlying the construct of the original scale.
Establishing a hierarchy will legitimize the shortening of the scale administration and confirm that it is appropriate to allocate scores for untested scale items, such that a meaningful total summed score can be used. In this study, Guttman scale analysis14 was used to investigate the hierarchical properties of the FM motor scale, both within each stage and each subsection and stepwise between stages and subsections.
|
Method
|
|---|
Participants and Procedure
We conducted a retrospective analysis of data obtained as part of a previous study15 investigating the reliability and validity of the Modified Ashworth Scale (MAS), an assessment measure for spasticity.16 The original study, which took place in April and May 2000, investigated the construct validity of the MAS by correlating the scores of the FM motor scale with those of the MAS. Therefore, for all patients, each item of the FM motor scale was administered in full. All adult patients receiving physical therapy at Erasmus MC, University Hospital Rotterdam, or one of the participating nursing homes were screened for inclusion in the study. The inclusion criteria were: ischemic or hemorrhagic stroke diagnosed by a physician; occurrence of the stroke at least 3 months previously; possible presence of previous strokes, transient ischemic attacks, or reversible ischemic neurological deficits; paresis of the arm or the leg; and granting of written informed consent. The exclusion criteria were: coexisting different type of neurological condition; venous thrombosis in the affected limbs; and inability of the patient to attain the starting position for the testing procedure or maintain it for 30 minutes.
Just one assessor administered the FM motor scale. This experienced physical therapist has worked with patients with neurological conditions for over 25 years and has routinely incorporated the FM motor scale as part of her clinical practice. The testing procedure was conducted in a standardized manner, according to the written instructions originally published, but with some additional standardizations approved by a national interest group of neuro-physical therapists from university hospitals (FANN)* (Appendix 1).1,17 These standardizations, which served to ensure the accurate reproducibility of some aspects of the testing procedure, did not alter the scoring criteria in any way. The testing of reflexes was reduced from 3 to 2 reflexes for both the UE and the LE. The finger and hamstring reflexes were deemed least relevant in clinical practice and were omitted, a step that also served to clarify the scoring of these items. Furthermore, the testing of coordination for the UE was not clear and therefore was standardized (Appendix 2).18
The administration of the FM motor scale was not possible for all 72 participants admitted to the original study; 62 participants were in the UE group, and 59 participants were in the LE group. Ten participants were completely excluded because they were unable to follow instructions adequately. In addition, 3 participants had other pathologies relating to the LE, making testing impossible. The mean age of the participants was 73 years (SD=10.5, range=42–94). The mean time after stroke was 17 months (SD=14.26, range=3 months–10 years). Other demographic details are provided in Table 1.
Data Analysis
The initial phase of the analysis was a visual inspection of the raw data to ensure that the general pattern of a patient's recovery followed the expected stagewise progression. The frequencies of "scale" patients (those whose motor recovery progressed as expected in the rank order of the scale items and stages) and of "non-scale" patients (those whose motor recovery did not progress as expected in the rank order or stagewise manner) were noted. Likewise, the frequencies of "extreme" patients, as defined by Adams et al,19 were summed, either as those with a ceiling effect of passing all of the scale items or those failing to score on any scale items. To establish the hierarchical properties of the FM motor scale, we performed Guttman scale analysis14 to confirm that the scale items were in the correct order. This analysis initially was performed manually and later was confirmed with Microsoft Excel 1997
for the following stepwise permutations: (1) transformation of raw data into a dichotomous score; (2) ordering of items within each stage and subsection, defined by the total number of passes per item, to form a rank ordering of item difficulty; (3) prediction of pass order per subject from their total number of passes; (4) calculation of total number of errors from the mismatch of actual order and predicted order of passes; and (5) calculation of statistical values with standard formulas.14 Because a minimum of 3 items are required to perform a scalogram analysis, this analysis was not possible within all of the stages. In particular, stages 1 and 5 (reflexes) for both the UE and the LE were affected. Both stage 3 and stage 4 for the LE have only 2 scale items each, and these stages were combined to create a joint stage.
With Guttman scale analysis, the initial calculation was the coefficient of reproducibility (CR).14 This value provides an estimation of the accuracy with which a scale score predicts which items are passed or failed. Any discrepancies from the score predictions are counted as scaling errors. A CR of .9 or higher indicates a valid cumulative and unidimensional Guttman scale and indicates that a subject's scores can be legitimately summed20; that is, a hierarchy of scale item difficulty is established from the rank order such that it is appropriate to use a cumulative total score. However, if many participants happen to pass or fail all items of a scale (ie, "extreme" participants), then the CR will be spuriously high. This situation also can occur if all participants pass a scale item or if a scale item is too difficult for all participants (ie, "extreme" items). Therefore, the minimal marginal reproducibility (MMR),19 which corrects for the chance appearance of a hierarchy, is calculated. The MMR can be based on the extremeness of items or the extremeness of participants. Menzel21 recommended using the smaller of the 2. Finally, the coefficient of scalability (CS) was calculated. This value indicates how much better the CR is than the MMR (ie, a true and not a chance hierarchy) and indicates the proportion of responses that can be correctly predicted from the total summed score, that is, allowing for the relative frequencies with which different scale items are passed. A CS of 0.6 is required to confirm the validity of Guttman scale analysis.20
For the scalogram analysis, the raw data had to be transformed from a 3-point scoring system for each item to dichotomous pass or fail scores. Data transformation converted all scores of 0 to fail scores and combined scores of 1 or 2 points to pass scores. Guttman scale analysis was then performed in 3 phases. The initial phase was done to establish that the correct rank ordering was evident within each stage and subsection of the FM motor scale (within-stage and subsection analysis). More information was obtained from the second phase, in which the scaling properties between the stages and subsections were analyzed (between-stage and subsection analysis). The between-stage analysis for the UE was done only for stages 2 and 3 and for stages 3 and 4 of the scale; stages 1 and 5 were excluded because they do not contain the minimum of 3 scale items required to perform a scalogram analysis. This between-stage and subsections analysis was repeated, initially including the wrist and hand subsections and finally including all of the subsections. For the LE, the between-stage analysis was performed with stages 1 and 5 excluded and with and without the coordination subsection. The final phase considered an entire limb; therefore, all of the stage divisions were ignored. Furthermore, to verify the allocation of appropriate scores for untested scale items, the exact rank ordering of the test items in the hierarchy was examined.
|
Results
|
|---|
The initial visual inspection of the raw data confirmed that a stepwise progression appeared to be evident for the majority of the participants. For the UE analysis, 9 participants (14.5%) failed 3 items, followed by another pass in the rank ordering (ie, non-scale types), and no one passed all of the items (ie, no ceiling effect). Similarly, for the LE analysis, 1 subject failed to score on the entire scale (ie, extreme patient type), and just 4 participants (6.5%) were non-scale types. No participants passed or failed all of the items for both the UE and the LE. Furthermore, no scale item was passed by all of the participants, and no scale item was too difficult for all of the participants. These findings indicated that the calculation of the MMR should be based on the extremeness of items, as recommended by Menzel.21 Figure 2 shows the spread of scores across all of the stages and subsections for both extremities.
View larger version (31K):
[in this window]
[in a new window]
|
Figure 2. Spread of raw scores across all stages and subsections of the upper-extremity (UE) and lower-extremity (LE) sections of the FM motor scale.
|
|
These results provided an adequate incentive to progress further with the investigation and to perform Guttman scale analysis. For calculation of the CR, some items were reordered as part of the ranking procedure. Further analysis then was performed on the reordered data within each stage and subsection (Tabs. 2, 3, and 4).
View this table:
[in this window]
[in a new window]
|
Table 2. Original and New Rank Ordering of Items in the Shoulder-Arm Subsection of the Upper-Extremity (UE) Section of the FM Motor Scale
|
|
View this table:
[in this window]
[in a new window]
|
Table 3. Original and New Rank Ordering of Items in the Wrist, Hand, and Coordination Subsections of the Upper-Extremity (UE) Section of the FM Motor Scale
|
|
The within-stage and subsection analysis for the UE and the LE provided acceptable CR and CS values for all of the individual stages and subsections. Likewise, for the between-stage and subsection analysis, acceptable CR and CS values were obtained for all combinations of stages 2, 3, and 4 with coordination and with the wrist and hand for the UE. For Guttman scale analysis, across all stage items for the UE and the LE (ignoring stage divisions), the CR value was just below the acceptable level, although the CS value was acceptable (Tab. 5).
Further inspection of the new rank ordering of stages 2, 3, and 4 (Tabs. 2, 3, and 4) considered the allocation of scores for untested scale items. The hierarchical order demonstrated that when a chosen stage is fully scored (ie, each scale item within the stage is awarded a score of 2 points such that a maximum score is obtained), all of the items in any previous untested stages should have already occurred in the hierarchy and therefore can legitimately be allocated a full score. Likewise, when a person fails to score for all of the scale items of an entire stage, no pass scores would be expected for scale items in higher stages, because these items appear further along in the hierarchy. Therefore, the allocation of 0 points for any remaining untested scale items is acceptable. Because it was not appropriate to include the reflex stages (1 and 5) in the inspection, they were excluded from this aspect of the investigation.
|
Discussion
|
|---|
The present study supports the use of the UE and LE sections (balance excluded) of the motor function domain of the FM assessment scale (FM motor scale) as a stagewise and subsectionwise hierarchical assessment and outcome measure. Therefore, the scale can be administered in a shortened manner, and it is appropriate to use a summed score.
The FM motor scale has been widely used in comparison studies of measurement scales2,5 or different treatment techniques.22–24 It is an inexpensive assessment measure that requires minimal equipment or space; therefore, it is ideal for assessing the recovery of motor function after stroke. As an assessment measure, it provides information on the status of a patient. As an outcome measure, it can evaluate change over time and, because of its predictive properties, can be used in the setting of achievable objectives of treatment.25 Nevertheless, in clinical practice, the FM motor scale is often criticized as being time-consuming. Shortening the time required to administer the assessment, without a loss of information, could lead to more-frequent use of the scale in clinical practice.
In the present study, there appeared to be 2 anomalies in the hierarchy (Tabs. 2, 3, and 4). Stage 4 of the UE section, the item pronation/supination (with shoulder flexion and elbow extension), emerged earlier in the rank ordering than expected, namely, before stage 3 was completed. It is likely that this occurrence was attributable to an incorrect translation from the English scale. The translated version allows additional support of the upper arm, making the item easier to perform and perhaps leading to an incorrect placement of this item in the hierarchy.
The other anomaly relates to reflexes. For the UE, the reflex testing of stage 1 remained at the beginning of the hierarchy, whereas for the LE, reflexes became mixed in with stage 2 scale items. According to the Fugl-Meyer et al guidelines,1 stage 5 reflexes are tested only if all stage 4 items are passed. In the present study, stage 5 was always tested, whatever the results from stage 4; this method led to the unexpected placing of stage 5 in the hierarchy. The new rank ordering of stage 5 also could be related to the unreliability of the testing of reflexes26 rather than to the hierarchy or could be related to some participants having no reflex activity on both sides, resulting in a potentially false pass score because of the symmetry of reflex activity. Furthermore, diabetes, medication, stress, and infections also may affect reflex behavior. In clinical practice, little information about voluntary motor function is gained from the reflex items. This fact, in combination with the unreliability issues raised above, leads to the recommendation that the reflex items be removed from the FM motor scale.
The results of the initial phase of investigation, which considered within-stage hierarchies, confirmed a hierarchy within all of the stages and subsections of the UE and LE sections. Some changes in rank ordering were seen within most of the stages and subsections of the UE and LE sections.
The between-stage and subsection Guttman analysis for the UE resulted in acceptable CR and CS values, thereby supporting the hypothesis that the stepwise progression of the scale, by stages and subsections, is hierarchical. According to the original guidelines published by Fugl-Meyer et al,1 the wrist and hand are considered to recover somewhat independently from the rest of the UE. The between-stage and subsection results of the present study suggest that the hierarchy continues into the wrist, hand, and coordination subsections. When the stage divisions for the shoulder-arm and leg subsections were ignored, the scalogram analysis for all items of the 5 stages produced less-favorable results. Although the CS values were acceptable (
0.6) and, therefore, supported general scalability, the CR values were just below the acceptable level (.9) and, therefore, were unable to support the reproducibility of the hierarchy. These results suggest that the stepwise progression of the scale in stages and subsections is an integral and important aspect of the hierarchical properties of the scale, supporting the need for these divisions. Therefore, because the hierarchy that we have demonstrated in the present study is stepwise by stages and subsections, rather than linear, it is only possible to abbreviate scale administration in terms of complete stages, rather than after 2 or 3 consecutive passes or fails, as described in other studies.8,9 Because there were only a few participants with scores in the wrist, hand, and coordination subsections, further work is needed to support the notion that the hierarchy truly progresses into these subsections. Therefore, we still recommend that items in the UE wrist and hand subsections always be tested.
With a hierarchy established, scale administration can be legitimately shortened without affecting the validity of the scale. Scores can be allocated for untested items, whether omitted (with the assumption of a pass and 2 points allocated) or not attempted (with the assumption of a fail and 0 points allocated), such that a summed score can be legitimately used. However, in the present study, we used data transformation, in which a fail was assigned a score of 0 and scores of 1 and 2 were combined to form a pass. This analysis specifically supported the hypothesis that when an entire stage of the scale is passed, the remaining scale stages need not be tested and a score of 0 can be allocated, as indicated by the rank ordering of items in Tables 2, 3, and 4. The hypothesis of a retrospective allocation of 2 points when an entire stage is passed also was supported by this new rank ordering. However, the alternative method of data transformation would have been more specific (ie, combining scores of 0 and 1 as a fail and leaving only a score of 2 as a pass). Therefore, a reanalysis of the data with this alternative data transformation was performed. Similar results were obtained; therefore, the inclusion of the latter analysis was deemed unnecessary.
With an accurate choice of a starting point, a shortened assessment can be completed in 5 minutes instead of the approximately 20 minutes required for a full assessment.27 However, when an individual attains partial scores in stage 3, a full assessment is still necessary. The optimal benefit occurs when a person attains maximal scores in stage 4 or minimal scores in stage 2. Full guidelines for the hierarchical administration of the FM motor scale are provided in Appendix 2.
The use of a summed score with ordinal data remains an open debate.10,12 Establishing the hierarchy of the FM motor scale justifies the use of such a summed score. Problems may be encountered even for a hierarchical scale with good CR and CS values when a summed score is considered. Collin stated, "A particular score on a hierarchical test should always indicate the same level of ability."12(p93) Because the FM motor scale is a 3-point scale, this situation may not necessarily be the case, because patients with different patterns of disability can achieve the same total score.28 Ebrahim et al9 suggested when such ranked scales are used to compare patients’ summed scores over time, not only the number of items passed but also the most difficult items passed should be compared. With the hierarchy of the FM motor scale established in stages and subsections, in clinical practice it may be sufficient to identify the highest stage achieved by a patient. For research purposes, it may be necessary to identify the most difficult item achieved. Numbering the scale items with the new rank ordering on a score sheet can identify this item.
In the ordinal FM motor scale, there are only a few items in each of the stages and subsections, such that the data are suited to Guttman scalogram analysis, which is based on a deterministic model. Alternatively, Rasch analysis, which is based on a probabilistic model, could be used to shorten the administration time. However, Rasch analysis is more appropriate for a long measurement scale and a large study population.29 Such an analysis was recently reported by Woodbury et al13 but considered only the UE section of the FM motor scale.
A comparison of the 2 studies highlights some interesting facts. In common with the present study, Woodbury et al13 also identified reflex items as being problematic. Like us, they recommended the removal of these items from the scale.
The present study consisted of 3 phases. The first and second phases established hierarchies within and between the stages and subsections of the FM motor scale. Only the third phase can be compared with the study of Woodbury et al.13 In that phase, when the scale divisions were ignored and a hierarchy was not established, some interesting differences and similarities in the new rank ordering were evident between the 2 studies. Woodbury et al assumed a linear continuum and therefore ignored stage and subsection divisions, which the present study suggested are a fundamental part of the construct of the FM motor scale. A striking difference in the linear rank ordering, from easiest to most difficult, of the 2 studies is that in the present study, the scale items were reordered within a stage and there was no mixing between stages. However, in the study of Woodbury et al, the scale items were truly mixed. For example, all of the items of stage 3 and one item of stage 4 appeared before stage 2 was totally completed. Their new rank ordering of all of the scale items appeared to break from the underlying conceptual framework suggested by Fugl-Meyer et al.1 In contrast, the present study is supportive of the phylogenetic construct underlying the hierarchy of the FM motor scale.
Differences in the new rank ordering may have occurred because of the different statistical methods used to analyze the data. In addition, the study population of Woodbury et al13 consisted predominantly of people with a mild to moderate stroke, whereas the study population in the present study consisted mostly of people with moderate to severe disability. The limitations in both studies also may have contributed to the differences in rank ordering.
The similarities in the new rank ordering in both studies were related to the wrist and hand items being spread throughout the stages. This finding can be explained by the independent recovery of the wrist and hand as described by Fugl-Meyer et al.1 Furthermore, both studies demonstrated a mixing up of the flexor and extensor synergy items in stage 2. However, this finding has no impact on test administration because all of the scale items in this stage were tested with just 2 synergistic movements.
One obvious anomaly in the item difficulty hierarchical order of Woodbury et al13 is that of coordination (movement without tremor) appearing so close to the easiest item. The original guidelines of Fugl-Meyer et al1 for the testing of coordination items failed to provide a starting position. Therefore, in the present study, the standardized test administration recommended by a national interest group of neuro-physical therapists was used.17 This strategy further emphasizes the importance of good standardization of test administration procedures, which also was mentioned by Woodbury et al.
The present study was a retrospective study of data available from a previous study in which the FM motor scale was administered in its entirety to people with chronic stroke. Further work is now needed to generalize the findings to a wider population of people with stroke and to compare the test score results obtained with the original method of administration and the shortened version for the same participants. Such a prospective study will identify the risk of missing non-scale–type participants. Because the 2 methods of data transformation evaluated in the present study produced the same results, a dichotomized pass/fail scoring system could be considered in the future. Ultimately, using the shortened method of administration of the FM motor scale should be investigated for its reproducibility and responsiveness.
|
Conclusion
|
|---|
Our study supports the hypothesis that the UE and LE sections (excluding balance) of the motor function domain of the FM assessment scale fulfill the statistical criteria of a Guttman scale and may be administered as a stagewise and subsectionwise hierarchical scale. With such a hierarchy established, it is now appropriate to start the test administration at a stage considered appropriate to the obvious level of recovery of a patient. If a patient passes the chosen stage, then maximum scores can be allocated for any previous (untested) stages. If a patient fails all items in a stage, testing can cease because it is very unlikely that further items will be successfully passed. Some caution is required when the hierarchy is extended into the wrist, hand, and coordination subsections because of the small number of participants achieving scores in these subsections in the present study. It appears that less time is now required to administer the FM motor scale, and a summed score can be legitimately calculated. Further work is needed to consolidate the results for a wider population with stroke and to reconfirm the robustness of the scale when a shortened administration process is used.
|
Appendix 1.
|
|---|
View larger version (37K):
[in this window]
[in a new window]
|
Appendix 1. Standardization of the Administration of the Upper-Extremity (UE) and Lower-Extremity (LE) Sections of the FM Motor Scale
|
|
|
Appendix 2.
|
|---|
View larger version (87K):
[in this window]
[in a new window]
|
Appendix 2. Guidelines for the Hierarchical Administration of the Upper-Extremity (UE) and Lower-Extremity (LE) Sections of the FM Motor Scale
|
|
|
Footnotes
|
|---|
Both authors provided concept/idea/research design, writing, and project management. Ms Crow provided data analysis. Ms Harmeling-van der Wel provided data collection, participants, facilities/equipment, and institutional liaisons.
The authors acknowledge the support of Professor Gert Kwakkel, Sylvia Gorter, and Dr Diederik WJ Dippel for proofreading this article and Dr Marij E Roebroeck for advice on statistics. They thank the student physical therapists who were involved in the coordination of the original study.
Ethics approval for this study was granted by the Medical Ethics Testing Committee, Erasmus MC, University Hospital Rotterdam.
* FANN=Fysiotherapeuten Academische Ziekenhuizen Neurologie Neurochirugie. 
Microsoft Corp, One Microsoft Way, Redmond, WA 98052-6399.
|
References
|
|---|
- Fugl-Meyer AR, Jääskö L, Leyman I, et al. The post-stroke hemiplegic patient, 1: a method for evaluation of physical performance. Scand J Rehabil Med. 1975;7:13–31.[Medline]
- Poole JL, Whitney SL. Motor assessment scale for stroke patients: concurrent validity and interrater reliability. Arch Phys Med Rehabil. 1988;69:195–197.[Web of Science][Medline]
- Duncan P, Propst M, Nelson SG. Reliability of the Fugl-Meyer assessment of sensori-motor recovery following cerebrovascular accident. Phys Ther. 1983;63:1606–1610.[Abstract/Free Full Text]
- Gladstone DJ, Danells CJ, Black SE. The Fugl-Meyer assessment of motor recovery after stroke: a critical review of its measurement properties. Neurorehabil Neural Repair. 2002;16:232–240.[Abstract/Free Full Text]
- Beckerman H, Vogelaar TW, Lankhorst GJ, Verbeek AL. A criterion for stability of the motor function of the lower extremity in stroke patients using the Fugl-Meyer Assessment Scale. Scand J Rehabil Med. 1996;28:3–7.[Web of Science][Medline]
- Twitchell TE. The restoration of motor function following hemiplegia in man. Brain. 1951;74:443–480.[Free Full Text]
- Brunnstrom S. Motor testing procedures in hemiplegia: based on sequential recovery stages. Phys Ther. 1966;46:357–375.[Medline]
- Lincoln N, Leadbitter D. Assessment of motor function in stroke patients. Physiotherapy. 1979;65:48–51.[Medline]
- Ebrahim S, Nouri F, Barer D. Measuring disability after stroke. J Epidemiol Community Health. 1985;39:86–89.[Abstract/Free Full Text]
- Michels E. Measurement in physical therapy: on the rules for assigning numerals to observations. Phys Ther. 1983;63:209–215.[Abstract/Free Full Text]
- DeWeerdt WJG, Harrison MA. Measuring recovery of arm-hand function in stroke patients: a comparison of the Brunnstrom Fugl-Meyer test and the action research arm test. Physiother Can. 1985;37:65–70.[CrossRef]
- Collin C. Scales and scaling techniques used in measurement in rehabilitation medicine. Clin Rehabil. 1992;6:91–96.[Free Full Text]
- Woodbury ML, Velozo CA, Richards LG, et al. Dimensionality and construct validity of the Fugl-Meyer Assessment of the upper extremity. Arch Phys Med Rehabil. 2007;88:715–723.[Medline]
- Guttman LA. Basis for scaling qualitative data. Am Sociol Rev. 1944;9:139–150.[CrossRef]
- Van Dormolen I, de Hoogh P. Een Onderzoek naar de Betrouwbaarheid en Validiteit van de Modified Ashworth Scale [thesis]. Rotterdam, The Netherlands: Hogeschool Rotterdam; 2000.
- Bohannon RW, Smith MB. Interrater reliability of a modified Ashworth scale of muscle spasticity. Phys Ther. 1987;67:206–207.[Abstract/Free Full Text]
- Koolstra M, Smeets CJ, Harmeling-van der Val BC, et al. Klinimetrie na een Beroete, een Praktische Handleiding. 2nd ed. Amersfoort, The Netherlands: Nederlands Paramedisch Instituut; 2004:28–53.
- Oosterhuis HJGH. Klinische Neurologie, een Beknopt Leerboek. Houten/Diegem, The Netherlands: Bohn Stafleu Van Loghum; 1997:29.
- Adams SA, Ashburn A, Pickering RM, Taylor D. The scalability of the Rivermead motor assessment in acute stroke patients. Clin Rehabil. 1997;11:42–51.[Abstract/Free Full Text]
- Nouri FM, Lincoln NB. An extended activities of daily living scale for stroke patients. Clin Rehabil. 1987;1:302–305.
- Menzel H. A new coefficient of scalogram analysis. Public Opinion Q. 1953;17:268–280. Cited in: Adams SA, Ashburn A, Pickering RM, Taylor D. The scalability of the Rivermead motor assessment in acute stroke patients. Clin Rehabil. 1997;11:42–51.[Abstract/Free Full Text]
- Winstein CJ, Rose DK, Tan SM, et al. A randomised controlled comparison of upper extremity rehabilitation strategies in acute stroke: a pilot study of immediate and long term outcomes. Arch Phys Med Rehabil. 2004;85:620–628.[CrossRef][Web of Science][Medline]
- Duncan P, Studenski S, Richards L, et al. Randomized clinical trial of therapeutic exercise in subacute stroke. Stroke. 2003;34:2173–2180.[Abstract/Free Full Text]
- Van Peppen RPS, Kwakkel G, Wood-Dauphinée S, et al. The impact of physical therapy on functional outcomes after stroke: what's the evidence? Clin Rehabil. 2004: 18:833–862.[Abstract/Free Full Text]
- Kwakkel G, Kollen B. Predicting improvement in the upper paretic limb after stroke: a longitudinal study. Restor Neurol Neurosci. 2007;25:453–460.[Web of Science][Medline]
- Manschot J, van Passel L, Algra A, van Gijn J. Mayo and NINDS scales for assessment of tendon reflexes: between observer agreement and implications for communication. J Neurol Neurosurg Psychiatry. 1998;64:253–255.[Abstract/Free Full Text]
- Crow JL, Harmeling BC. Development of a consensus and evidence-based standardised clinical assessment and record form for neurological in-patients: the Neuro Dataset. Physiotherapy. 2002;88:33–46.
- Barer DH, Murphy JJ. Scaling the Barthel: a 10-point hierarchical version of the activities of daily living index for use with stroke patients. Clin Rehabil. 1993;7:271–277.[Abstract/Free Full Text]
- Hambleton RK, Jones RW. An NCME instructional module on comparison of classical test theory and item response theory and their implications to test development. Educational Measurement: Issues and Practice. 1993;12:38–47.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Copyright © 2008 by the American Physical Therapy Association.
Top
Introduction
Method
