Formal statement concerning ethical approval: This study received ethical consent from the Hunter New England Area Health Service's Human Research Ethics Committee and the University of Newcastle's Human Research Ethics Committee.
Assessment of the upper limb in acute stroke: The validity of hierarchal scoring for the Motor Assessment Scale
Version of Record online: 8 SEP 2009
© 2009 The Authors. Journal compilation © 2009 Australian Association of Occupational Therapists
Australian Occupational Therapy Journal
Volume 57, Issue 3, pages 174–182, June 2010
How to Cite
Pickering, R. L., Hubbard, I. J., Baker, K. G. and Parsons, M. W. (2010), Assessment of the upper limb in acute stroke: The validity of hierarchal scoring for the Motor Assessment Scale. Australian Occupational Therapy Journal, 57: 174–182. doi: 10.1111/j.1440-1630.2009.00810.x
Contributors This research project was the initiative of a multiprofessional, multicentred team. RLP designed and undertook the study, monitored data collection, analysed the data and wrote the paper. KB and IJH assisted in the design of the study, monitored progress and critically reviewed the paper prior to submission. MP supervised the study, provided data, monitored progress and critically reviewed the paper for submission.
Rebekah L. Pickering BOccThy (Hons); Occupational Therapist. Isobel J. Hubbard PhD Candidate, MOccThy, DAppSc(OT); Academic Researcher. Kerry G. Baker PhD, BAppSc(PT); Senior Lecturer. Mark W. Parsons PhD, BMed, FRACP; Staff Specialist in Neurology, Associate Professor.
- Issue online: 21 MAY 2010
- Version of Record online: 8 SEP 2009
- Accepted for publication 31 May 2009.
- Motor Assessment Scale;
- upper limb;
Background/aim: Stroke is the greatest contributor to disability in Australian adults and much of this disability results from a stroke-affected upper limb. This study aimed to determine the validity of hierarchal scoring for the upper limb subscale of the Motor Assessment Scale (UL-MAS) in acute stroke using Rasch analysis.
Method: This study applied Rasch analysis to 40 UL-MAS assessment results across 25 subjects to determine the validity of the hierarchy of the three upper limb subsets: upper arm function (six), hand movements (seven) and advanced hand activities (eight). Rasch analysis examines the relationship between ‘item difficulty’ and ‘person ability’ and produces an output which represents the difficulty of each item in relation to each other.
Results: As hypothesised, the hierarchy was upheld within subset 6. In subset 7, the hierarchy was not upheld. Results indicated that item 3 was the least difficult, followed by items 1, 4, 2, 5 and 6 in order of increasing difficulty. In subset 8 the hierarchy was not upheld. Results indicated that item 1 was the least difficult, followed by item 6, then 2 and 5 of equal value and then 3 and 4 of equal value.
Conclusions: The hierarchal scoring is not supported for subsets 7 and 8 and future research is required to explore the validity of alternate scoring methods. At present, the authors recommend that the UL-MAS should be scored non-hierarchally, meaning that every item within the subsets should be scored regardless of its place within the hierarchy (UL-MAS-NH).
Stroke is the leading cause of disability in Australia (National Stroke Foundation (NSF), 2005), and upper limb motor dysfunction is one of the primary causes (Gillen & Burkhardt, 2004). Current research has found that the brain is neuroplastic and capable of reorganising itself, enabling motor function to be regained in some instances, particularly in the first month post-stroke (Carey & Seitz, 2007; Williams, Galea & Winter, 2001). As upper limb dysfunction often contributes to decreased independence in stroke survivors, it is vital that therapists and researchers alike are able to accurately measure changes in upper limb motor function. Not only will this enable stroke survivors and therapists to measure improvements post-stroke, it will also afford researchers the opportunity to identify the most effective and efficient upper limb interventions.
The Motor Assessment Scale (MAS) was developed by Carr, Shepherd, Nordham and Lynne (1985) to assess motor function following stroke. This outcome measure has been widely used in research as an assessment and as part of the inclusion/exclusion criteria for some studies (Barker, Brauer & Carson, 2008; English, Hillier, Stiller & Warden-Flood, 2006; Horsley, Herbert & Ada, 2007). It is also used as a ‘gold standard’ upon which other outcome measures are assessed (Tyson & DeSouza, 2004). The MAS is taught in physiotherapy and occupational therapy programs across Australian universities (Ada, Canning, Dean & Moore, 2004). As an assessment tool it is commonly used in Australia and internationally in acute and rehabilitation settings (Lannin, 2004; National Stroke Foundation, 2005; Okkema & Culler, 1998).
The MAS was originally designed to assess eight subsets of motor function and one subset of muscle tone. The upper limb subscale (UL-MAS) consists of subset 6: ‘Upper Arm Activity’, subset 7: ‘Hand Movements’, and subset 8: ‘Advanced Hand Activities’. A full outline of the assessment and scoring criteria can be found in Carr et al. (1985). The scoring criteria for the UL-MAS are outlined in Appendix 1. Each subset has six items (criteria) which are scored from 0: unable to perform any of the items, through to 6: able to perform all six of the items (Carr et al.).
Over the last 30 years, many outcome measures have been developed to assess motor function following stroke, such as the Fugl Meyer Assessment (Fugl-Meyer, Jaasko, Leyman, Olsson & Steglind, 1975) and the Action Research Arm Test (Lyle, 1981). There are also outcome measures that assess disability, such as the Functional Independence Measure (Centre for Functional Assessment Research and the Uniform Data System for Medical Rehabilitation, 1990). Identifiable problems with these tools are that they lack time efficiency, accessibility and clinical utility in the assessment of functional tasks (Lannin, 2004). Many of them also have notable floor and ceiling effects. These effects refer to the situation where a person may have a small degree of motor function but still score zero, or conversely where a person may obtain a full score but still have some difficulty with complex motor tasks (Lannin). Requiring 15–30 min to administer, the MAS is the only outcome measure recommended by the North American Post-Stroke Rehabilitation Clinical Practice Guidelines which contains functional test items (Okkema & Culler, 1998). Others have recommended the MAS for use following stroke as it measures both motor dysfunction and its impact on daily activities (Lannin; Loewen & Anderson, 1988; Williams, Galea & Winter, 2001).
Previous studies have examined the reliability and validity of the MAS and found it to be within acceptable ranges, as outlined in Table 1 and Table 2, respectively. As reported by many researchers, the MAS has significant floor and ceiling effects (Lannin, 2004; Hsueh & Hsieh, 2002; Williams, Galea & Winter, 2001). As a result of previous research the general tonus item, designed to measure muscle tone, has been removed from the MAS due to low reliability (Aamodt, Kjendahl & Jahnsen, 2006; Loewen & Anderson, 1988).
|Total MAS||0.95||Carr et al., 1985||0.87–1.00||Carr et al., 1985|
|Subset 6||0.93||Filiatrault et al., 1992||1.00||Filiatrault et al., 1992|
|Subset 7||1.00||Filiatrault et al., 1992||1.00||Filiatrault et al., 1992|
|Subset 8||1.00||Filiatrault et al., 1992||1.00||Filiatrault et al., 1992|
|Content validity||Determined by expert neurological physiotherapists and extensive literature review||Source|
|Carr et al., 1985|
|Tyson & DeSouza, 2004|
|Predictive validity||Valid predictor of function at one month post-stroke||Hsueh & Hsieh, 2002|
|Filiatrault et al., 1992|
|Construct validity||Contains a single underlying construct||Tyson & DeSouza, 2004|
|Center for Functional, 1990|
|Concurrent validity||With Fugl-Meyer|
|Total MAS||0.88||Kielhofner, 2006|
|0.91||Loewen & Anderson, 1988|
|0.93||Loewen & Anderson, 1988|
The aspect of validity which remains under scrutiny in the MAS is the hierarchal scoring system. Hierarchies are valuable as they reduce the time and effort required to administer the instrument (Okkema & Culler, 1998). However, there is discrepancy within the literature regarding the recommended method of scoring the MAS. Numerous researchers evaluating or utilising the MAS report inconsistencies with the hierarchy in the upper limb subsets or the UL-MAS (Dean & Mackey, 1992; English et al., 2006; Lannin, 2004; Malouin, Pichard, Bonneay, Duran & Corriveau, 1994; Poole & Whitney, 1988). It is also reported that some therapists currently using the MAS do not apply the hierarchy when scoring subscale 8, ‘Advanced Hand Activities’, and instead, simply add the number of items the patient is able to achieve, regardless of their place within the subset (Williams, Galea & Winter, 2001). There is currently a lack of evidence to either refute or support this scoring method. Many researchers conclude that the validity of the hierarchy for the UL-MAS requires further investigation (Dean & Mackey, 1992; English et al., 2006; Lannin, 2004; Malouin et al., 1994; Poole & Whitney, 1988). Only two studies have examined the validity of this hierarchy (Aamodt et al., 2006; Sabari et al., 2005).
In 2006, Aamodt et al. reported on their study assessing the unidimensional and hierarchal constructs of the MAS in the subacute stage post-stroke. They found that, in subacute stroke, (up to a few weeks post-stroke), the MAS was unidimensional and based on one underlying construct where all items measure the same trait, in this case motor function. They also found that subset 8 was not ordered correctly because of the difficulty of item 4 in comparison to items 5 and 6. The authors recommended that scoring be researched further before confirming its validation.
In 2005, Sabari et al. reported on their study assessing the hierarchal scoring for the UL-MAS. Participants included stroke survivors in the acute, subacute and chronic stages post-stroke, with time since stroke onset ranging from 3 days to 6.5 years. The study confirmed that the UL-MAS was unidimensional and found that, across all stages post-stroke, the hierarchy was valid for subset 6, but that there was a large gap in the level of difficulty between items 5 and 6. Their study also found that the hierarchy was not valid for subsets 7 and 8, and reported significant ceiling and floor effects in all three subsets.
From previous research, there is a lack of direction regarding the most valid scoring method for the UL component of the MAS. This lack of direction affects the interpretation of the relevant evidence by clinicians and researchers. This creates uncertainty as to the application of the UL-MAS in practice and the interpretation of research results.
The primary aim of this study was to determine the validity of the hierarchal scoring for the UL-MAS in the first month post-stroke only, as this is the phase when the MAS is most often applied and when there is the most potential for improvement (Carey & Seitz, 2007). This would also clarify whether or not the hierarchal scoring is influenced by the time post-stroke.
The authors hypothesised that for stroke survivors in the acute stage of recovery:
- 1The items in subset 6 of the UL-MAS will be ordered in a valid hierarchy.
- 2The items in subsets 7 and 8 of the UL-MAS will not be ordered in a valid hierarchy.
This study applied Rasch analysis to UL-MAS assessment results to determine the validity of the hierarchy of each subset, as described below. This study received ethical approval from two regional human research ethics committees: the Area Health Service and the University.
The data used for this study were 40 UL-MAS assessments that had been routinely collected from 25 stroke survivors. This sample size was the maximum number of completed assessments available for inclusion at the time of the study. A sample size of 30 is considered a minimum for Rasch analysis to be valid (Kielhofner, 2006). These assessment scores came from 25 patients in the first month of recovery as described below. Demographic data are outlined in Table 3. All UL-MAS results were collected by occupational therapists and physiotherapists trained in using the MAS and procedural protocols were adhered to. For data to be included in the study, the patient must have been admitted to a stroke unit with upper limb weakness resulting from a stroke. One dataset was excluded from the study as the person had presented with a recent wrist fracture.
|Participants: n = 25|
|Gender: n (%)|
|Age in years: mean (SD)||69.96 (11.97)|
|Time since stroke onset in days*|
|Mean (SD)||4.58 (2.93)|
|Affected side: n (%)|
Data for this study were obtained from two sources. The first source was data (n = 15) collected as part of a concurrent, functional magnetic resonance imaging (fMRI), acute stroke and upper limb study (Hubbard, Budd & Parsons, 2008). The fMRI study was investigating the changes in brain activation patterns following stroke and their correlation with upper limb impairment and intervention. It used the UL-MAS as an inclusion criterion and as an outcome measure. The UL-MAS scores sourced from the fMRI study included those recorded at admission and at 1 month post-stroke. These scores are considered to be separate results, as they represent different levels of motor function. Aamodt et al. (2006) used this method in a similar study described above. The second source was data routinely collected on acute stroke patients from three regional hospitals (n = 10). Only routinely collected data from the admission MAS were included in this study from the second source.
Rasch analysis was the statistical method selected for this study. Rasch analysis is based on Item Response Theory that focusses on the individual items that make up a scale, rather than the scale as a whole (Kielhofner, 2006; Rasch, 1993). However, it is important to note that the Rasch model requires logarithmic transformation to be brought into line with other Item Response Theory models (Baker & Kim, 2004). The Rasch model computes the relationship between ‘person ability’ and ‘item difficulty’ and produces an output such as in Figure 1. The output is a calibration of ‘person ability’ and ‘item difficulty’ on the one continuum, allowing for analysis of the relationship between each. The calibrations represent how much of the underlying trait is present. Therefore, it is expected that items that are more difficult represent more of the trait being measured, and in turn, are positioned higher on the continuum (Kielhofner, 2006). The bar graph on the right of the figure represents the frequency of persons who achieved the corresponding level of ability. The program used for analysis was JMP® version 6.0.0 (SAS Institute, Cary, NC, USA). For Rasch analysis to be valid, it is important to determine that the outcome measure is unidimensional. As discussed earlier, previous studies have determined that the UL-MAS measures the one underlying trait — motor function (Aamodt et al., 2006; Lannin, 2004).
The results of this study indicate that within the first month post-stroke, the hierarchy in subset 6 was upheld, but was not valid for subsets 7 and 8, as hypothesised. These results will be presented and discussed in terms of the individual subsets.
Subset 6: Upper arm function
As hypothesised, the hierarchy in subset 6 was upheld within this sample and within acute stroke, as seen in Figure 1. Item 1 was the least difficult, followed by items 2, 3, 4, 5 and 6 (see Appendix 1 for item descriptions). There is a large gap between items 1 and 2. This difference in difficulty level is most likely due to the fact that for item 1 the assessor is able to assist the subject to maintain the position, whereas in item 2, the subject must be able to maintain this position independently, requiring greater stabilisation of the glenohumeral joint. Items 3 and 4 are clustered on the continuum, indicating that a similar level of difficulty is required to complete both items. This is not unexpected, as both require a similar level of muscle control around the shoulder girdle and both include elbow extension. These results also support the concept that following stroke, the effects of gravity make functional movements more difficult, as seen in the difference in difficulty between lying with the arm in forward flexion (item 2) and sitting with the arm in this same position (item 4). It can also be seen that the difficulty of the task increases with the duration of time for which the position is required to be held. This relationship is seen between items 4 and 5, as shoulder forward flexion is required to be held for 2 seconds in item 4, but for 10 seconds in item 5. Ceiling effects were seen within this subset as 17 assessments indicated the subject achieved the maximum score.
Subset 7: Hand movements
As hypothesised, the hierarchy for subset 7 was not upheld within this sample, as seen in Figure 2. Item 3 was the least difficult, followed by items 1, 4, 2, 5 and 6. The results indicate that following stroke, pronation and supination (item 3) are less difficult than wrist extension (item 1) and radial deviation (item 2). This is possibly because both wrist extension and radial deviation require movement against gravity, whereas pronation/supination is less impacted by gravity. It is also noted that the (only) bilateral task of picking up, carrying and releasing a large ball (item 4) is less difficult than radial deviation and only slightly more difficult than wrist extension. This may be because picking up and placing down a ball does not require complex or controlled wrist and hand movements, apart from some degree of wrist flexion, but instead, require controlled shoulder and elbow movement. Item 5 was more difficult as it involved complex wrist and hand control, e.g. wrist extension, radial deviation and finger flexion. Item 6 remained the most difficult, where the only movement involved finger opposition. This may support the concept that, following stroke, control over more proximal, large muscle movements returns sooner than control over complex fine motor movements. Ceiling effects were again seen in this subset, where 12 results reported the subject achieved the maximum score.
Subset 8: Advanced hand activities
As hypothesised, the hierarchy was not upheld for subset 8, as seen in Figure 3. Item 1 was the least difficult, followed by item 6, then 2 and 5 of equal value, and then 3 and 4 of equal value. These results indicate that following stroke, grasping and releasing a small object (item 1) are less difficult than functional tasks requiring greater ranges of movement and more precision, such as combing the back of the hair (item 6). These results also indicate that functional tasks such as feeding (item 5) or combing hair (item 6) are significantly less difficult to perform than complex, timed, fine motor tasks such as writing (item 3 and 4). It also demonstrated that functional tasks requiring higher levels of precision, such as feeding (item 5), are more difficult to perform than those tasks requiring greater ranges of movement, such as combing the back of the hair (item 6). Floor effects were found within this subset, where 13 results reported subjects were unable to achieve a score greater than 0.
This study aimed to assess the validity of the hierarchal scoring system of the UL-MAS in the first month post-stroke. The results indicate that the hierarchal scoring system was invalid for this sample, as the items within subset 7 and 8 were not ordered in a valid hierarchy. This was the first study to examine this aspect of validity and to report on results that are particularly related to the first month post-stroke, and, for the most part, support results from previous studies (Aamodt et al., 2006; Sabari et al., 2005) enabling valid conclusions to be drawn.
This study contributes new evidence concerning the first month post-stroke; however, while the results are similar to those from previous studies, the ordering of the hierarchy in the acute phase revealed some differences. This suggests that post-stroke, the pattern of motor recovery may differ across recovery stages. This in turn suggests that while hierarchies have clinical utility in that they reduce the time and effort required for administration, in the case of the UL-MAS assessment applied in the first month post-stroke, it undermines the assessment's validity.
The findings from this study present some similarities and differences when compared to those of Sabari et al. (2005), as outlined below:
Subset 6: Upper arm function
With regard to subset 6, the only difference between the two sets of results was a transposition of items 2 and 3, suggesting that time post-stroke may have influenced the hierarchal scoring. Our analysis of subset 6 upheld the original ordering as developed by Carr et al. (1985), suggesting that in the first month post-stroke, subset 6 has a valid hierarchy.
Subset 7: Hand movements
Similarly, for subset 7 the only difference between the two sets of results was a transposition of items 2 and 5. Sabari et al. (2005) ordered this subset as 3, 1, 4, 5, 2, 6, whereas our analysis indicated that the order for subset 7 should be 3, 1, 4, 2, 5, 6. Item 2 involves radial deviation of the wrist, and item 5, picking up and moving a polystyrene cup. It is unclear why this ordering is different; however, it again suggests that time post-stroke influences the hierarchal scoring.
Subset 8: Advanced hand activities
For subset 8 there were a number of differences between Sabari et al.'s (2005) results and our own. Sabari et al.'s (2005) results ordered these items as 1, 2, 5, 6, 4, 3, whereas our study found that item 1 was the least difficult, followed by item 6, then 2 and 5 of equal value, and then 3 and 4 of equal value. The primary difference is the order of item 6. This item involves hair combing, which Sabari et al. (2005) found to be easier than item 4, making dots with a pencil, but more difficult than item 5, taking a spoonful of liquid to the mouth. However, our study found item 6 to be relatively easy and only more difficult than item 1, picking up and moving a pen top. Again, this may reflect differences in the pattern of motor recovery at different stages post-stroke.
These findings also indicate that some sections of the UL-MAS assess items that are in fact measuring very similar levels of difficulty in upper limb, motor function. Sabari et al. (2005) recommended that items be removed where their level of difficulty was similar to that of another item. For example, they recommended that in subset 7, items 1, 4 and 5 should be eliminated. If this recommendation was taken-up, our study would indicate that there are other items that could be eliminated because of their similar values.
In response to these findings and those of Sabari et al. (2005), we propose two possible modifications to the UL-MAS which could resolve the hierarchal scoring issue. We also acknowledge that any modification will need further research before validation, such as the research currently being performed by Sabari, Lim, Duhan, Chu and Cesaire (2008).
Option 1 modification: Apply non-hierarchal scoring to subsets 7 and 8 (UL-MAS-NH)
There has been some ‘flexibility’ concerning the MAS hierarchal scoring from its inception. In their original publication, Carr et al. (1985) stated that ‘all items except general tonus are constructed so that point 6 indicates the optimal motor behaviour’ (p. 175); while in a more recent publication, Carr and Shepherd (1998) stated that ‘the items do not have to be tested in any order; they are not in any way hierarchally organised’ (p. 50). It should be noted that the terminology between item, subset, criterion and tasks is confusing as it is used inconsistently between researchers. This has resulted in a lack of consistency with the recommended scoring methods.
The findings from our study placed alongside those from Sabari et al. (2005) confirm that the original constructs of subsets 7 and 8 as developed by Carr et al. (1985) are non-hierarchal, and that these two subsets cannot therefore be reliably tested or scored hierarchally. Given that all research to date has found subset 6 to be ordered in a valid hierarchy, it follows that there will be no change in the outcome if it is scored hierarchally or non-heirarchally when combined with subsets 7 and 8. Even though the analyses of subset 6 confirm its hierarchal properties, we nevertheless recommend that, for consistency of application: all items in all three subsets within the UL-MAS should be tested and scored non-hierarchically.
This means that the assessor should test all items and give a score for each item the subject is able to perform, regardless of its position within the subset and that the scoring methodology be adequately identified (UL-MAS-NH).
It is recommended that further research study the reliability and validity of the UL-MAS-NH, as the non-hierarchal scoring system for subset 6, 7 and 8 may impact other psychometric properties.
Option 2 modification: Collapse subsets 7 and 8 into one subset (UL-MAS-SF)
Some may propose that these subsets cannot simply be used to assess patients in a non-hierarchal manner, because each item cannot be evaluated without reference to its order within the subset, and each item has a numerical value corresponding to its position. Option 2, therefore, would be to remove some of the items with the same degree of difficulty as advocated by Sabari et al. (2005, 2008) and collapse the original subsets 7 and 8 into a new subset 7[SF] measuring ‘Hand Function’. Redundancy would apply to items confirmed as ‘overlapping’, for example items 2 and 5 and items 3 and 4 (Fig. 3).
Combining subsets 7 and 8 is plausible because both of the original subsets: 7 ‘Hand Movements’ and 8 ‘Advanced Hand Activities’ are measuring functional tasks. It is not easy to distinguish the two from one another on the basis of movement and function alone, nor on the basis of increased difficulty (Figs 2 and 3). Combining these two subsets and removing redundant items could result in a potential new hierarchal subset 7[SF] measuring ‘Hand Function.’ It could contain items reordered and renumbered to accurately reflect their new item difficulty. A new hierarchal subset 7[SF] would resolve the problems associated with the scoring of the UL-MAS. It could also shorten the time it takes to apply the UL-MAS. Therefore option 2 recommends collapsing subsets 7 and 8 into a new subset 7[SF] titled ‘Hand Function’, which, along with subset 6, would be scored hierarchically.
Further research is recommended if this modification is to be selected as again, the reliability and validity of the UL-MAS-SF will need to be established.
Limitations and future implications
This study is limited by the inclusion of admission and discharge scores from the participants in the fMRI study and this may have biased the findings. However, the utilisation of both sets of UL-MAS data enabled a greater scope of motor function to be included in the Rasch analysis. This study is also limited by the fact that the authors have chosen to recommend two, non-aligning modifications, and, as proposed initially, failed to resolve this issue beyond doubt. Future research should continue to explore the scoring system of the UL-MAS (Sabiri et al., 2008) as we work towards resolving this area of discrepancy.
The full MAS and the UL-MAS have been used in many studies and are considered a ‘gold standard’ outcome measure. While this issue remains unresolved, it potentially casts a shadow over the evidence that has utilised these assessments; however, it is beyond the scope of this study to review the relevant evidence. Therapists measuring changes in motor function are also routinely selecting the MAS and the UL-MAS post-stroke (National Stroke Foundation, 2008), which in turn, further presses the need to resolve this issue.
Based on the findings of this study and taken in light of other relevant evidence, the authors recommend that, until evidence indicates otherwise, therapists and clinicians should select option 1, scoring the UL-MAS non-hierarchally and adequately identifying that they have used this scoring methodology (UL-MAS-NH).
Clinically, the MAS is designed to assist therapists in determining which aspects of motor function are stroke-affected and which are not, and to measure change over time (Dean & Mackey, 1992). It is used by both clinicians and researchers and has devolved into the UL-MAS, which is particularly favoured by occupational therapists.
The findings of this study indicate that within the first month post-stroke, the hierarchy in subset 6 was upheld, while the hierarchy was not upheld for subsets 7 and 8, as hypothesised. The authors recommend two options in response to this evidence, but at present, give preference to option 1 until this issue is further resolved.
Option 1 recommends that therapists and/or researchers using the upper limb component of the MAS score all three subsets non-hierarchically while clearly identifying that this is how the assessment was scored (UL-MAS-NH).
The authors wish to thank the acute stroke research team and therapists who provided the data for this study; Kate Day, the research assistant who collated the data; Jenny White who assisted in the development of the project and Kim Colyvas who provided assistance with data analysis.
- 2006). Dimensionality and scalability of the Motor Assessment Scale (MAS). Disability and Rehabilitation, 28, 1007–1013. , & (
- 2004). Training physiotherapy students’ abilities in scoring the Motor Assessment Scale for stroke. Journal of Allied Health, 33, 267–270. , , & (
- 2004). Item response theory: Parameter estimation techniques. New York: Marcel Dekker. & (
- 2008). Training of reaching in stroke survivors with severe and chronic upper limb paresis using a novel nonrobotic device: A randomised clinical trial. Stroke, 39, 1800–1807. , & (
- 2007). Functional neuroimaging in stroke recovery and neurorehabilitation: Conceptual issues and perspectives. International Journal of Stroke, 2, 245–264. & (
- 1998). Neurological rehabilitation: Optimizing motor performance. Oxford: Boston Butterworth-Heinemann. & (
- 1985). Investigation of a new Motor Assessment Scale for stroke patients. Physical Therapy, 65, 175–180. , , & (
- Center for Functional Assessment Research and the Uniform Data System for Medical Rehabilitation (1990). Guide for use of the Uniform Data Set for Medical Rehabilitation including the Functional Independence Measure (FIM), Version 3.1. Buffalo, NY: State University New York.
- 1992). Motor Assessment Scale scores as a measure of rehabilitation outcome following stroke. Australian Journal of Physiotherapy, 38, 31–35. & (
- 2006). The sensitivity of three commonly used outcome measures to detect change amongst patients receiving inpatient rehabilitation following stroke. Clinical Rehabilitation, 20, 52–55. , , & (
- 1992). Motor function and activities of daily living assessments: A study of three tests for persons with hemiplegia. American Journal of Occupational Therapy, 45, 806–810. , , & (
- 1975). The post-stroke hemiplegic patient: A method of evaluation of physical performance. Scandinavian Journal of Rehabilitation Medicine, 7, 13–31. , , , & (
- 2004). Stroke rehabilitation: A function-based approach (2nd ed.). St Louis, MO: Mosby Inc. & (
- 2007). Four weeks of daily stretch has little or no effect on wrist contracture after stroke: A randomised controlled trial. Australian Journal of Physiotherapy, 53, 239–245. , & (
- 2002). Responsiveness of two upper extremity function instruments for stroke inpatients receiving rehabilitation. Clinical Rehabilitation, 16, 617–624. & (
- 2008). Arm function, fMRI and early reorganisation mapping in stroke (AFfERMS): Findings from the pilot phase. Internal Medicine Journal, 38(S4), A80. , & (
- 2006). Research in occupational therapy: Methods of inquiry for enhancing practice. Philadelphia, PA: F. A. Davis Company. (
- 2004). Reliability, validity and factor structure of the upper limb subscale of the Motor Assessment Scale (UL-MAS) in adults following stroke. Disability and Rehabilitation, 26, 109–115. (
- 1988). Reliability of the modified Motor Assessment Scale and the Barthel Index. Physical Therapy, 68, 1077–1081. & (
- 1990). Predictors of stroke outcome using objective measurement. Stroke, 21, 78–81. & (
- 1981). A performance test for assessment of upper limb function in physical rehabilitation treatment and research. Internal Journal of Rehabilitation Research, 4, 483–492. (
- 1994). Evaluating motor recovery early after stroke: Comparison of the Fugl-Meyer Assessment and the Motor Assessment Scale. Archives of Physical Medicine and Rehabilitation, 75, 1206–1212. , , , & (
- National Stroke Foundation (2005). Clinical guidelines for stroke rehabilitation and recovery. Melbourne, Vic.: Australian Government.
- National Stroke Foundation (2008). National stroke audit post acute services. Melbourne, Vic.: Australian Government.
- 1998). Functional evaluation of upper extremity use following stroke: A literature review. Topics of Stroke Rehabilitation, 4, 54–75. & (
- 1988). Motor Assessment Scale for stroke patients: Concurrent validity and interrater reliability. Archives of Physical Medicine and Rehabilitation, 69, 195–197. & (
- 1993). Probabilistic models for some intelligence and attainment tests. Chicago, IL: MESA Press. (
- 2005). Assessing arm and hand function after stroke: A validity test of the hierarchal scoring system used in the Motor Assessment Scale for stroke. Archives of Physical Medicine and Rehabilitation, 86, 1609–1615. , , , , & (
- 2008). Developing a new scoring hierarchy for hand items on the Motor Assessment Scale for stroke. Physical Disabilities, Special Interest Section Quarterly, 31, 1–4. , , , & (
- 2004). Development of the Brunel Balance Assessment: A new measure of balance disability post stroke. Clinical Rehabilitation, 18, 801–810. & (
- 2001). What is the functional outcome for the upper limb after stroke? Australian Journal of Physiotherapy, 47, 19–27. , & (
APPENDIX 1: Criteria for scoring the UL-MAS
Subset 6: Upper-Arm Function
Item 1. Lying, protract shoulder girdle with arm in elevation. (Therapist places arm in position and supports it with elbow in extension.)
Item 2. Lying, hold extended arm in elevation for 2 seconds. (Therapist should place arm in position and patient must maintain position with some external rotation. Elbow must be held within 20° of full extension.)
Item 3. Flexion and extension of elbow to take palm to forehead with arm as in Item 2. (Therapist may assist supination of forearm.)
Item 4. Sitting, hold extended arm in forward flexion at 90° to body for 2 seconds. (Therapist should place arm in position and patient must maintain position with some external rotation and elbow extension. Do not allow excess shoulder elevation.)
Item 5. Sitting, patient lifts arm to above position, holds it there for 10 seconds, and then lowers it. (Patient must maintain position with some external rotation. Do not allow pronation.)
Item 6. Standing, hand against wall. Maintain arm position while turning body toward wall. (Have arm abducted to 90° with palm flat against the wall.)
Subset 7: Hand Movements
Item 1. Sitting, extension of wrist. (Therapist should have patient sitting at a table with forearm resting on the table. Therapist places cylindrical object in palm of patient's hand. Patient is asked to lift object off the table by extending the wrist. Do not allow elbow flexion.)
Item 2. Sitting, radial deviation of wrist. (Therapist should place forearm in mid-pronation–supination, i.e. resting on ulnar side, thumb in line with forearm and wrist in extension, fingers around a cylindrical object. Patient is asked to lift hand off table. Do not allow elbow flexion or pronation.)
Item 3. Sitting, elbow into side, pronation and supination. (Elbow unsupported and at a right angle. Three-quarter range is acceptable.)
Item 4. Reach forward, pick up large ball of 14-cm (5-in) diameter with both hands and put it down. (Ball should be on table so far in front of patient that he has to extend arms fully to reach it. Shoulders must be protracted, elbow extended, wrist neutral or extended. Palms should be kept in contact with the ball.)
Item 5. Pick up a polystyrene cup from table and put it on table across other side of body. (Do not allow alteration in shape of cup.)
Item 6. Continuous opposition of thumb and each finger more than 14 times in 10 seconds. (Each finger in turn taps the thumb, starting with index finger. Do not allow thumb to slide from one finger to the other, or to go backwards.)
Subset 8: Advanced Hand Activities
Item 1. Picking up the top of a pen and putting it down again. (Patient stretches arm forward, picks up pen top, releases it on table close to body.)
Item 2. Picking up one jellybean from a cup and placing it in another cup. (Teacup contains eight jellybeans. Both cups must be at arms’ length. Left hand takes jellybean from cup on right and releases it in cup on left.)
Item 3. Drawing horizontal lines to stop at a vertical line 10 times in 20 seconds. (At least five lines must touch and stop at the vertical line.)
Item 4. Holding a pencil, making rapid consecutive dots on a sheet of paper. (Patient must do at least two dots a second for 5 seconds. Patient picks pencil up and positions it without assistance. Patient must hold pen as for writing. Patient must make a dot not a stroke.)
Item 5. Taking a dessert spoon of liquid to the mouth. (Do not allow head to lower towards spoon. Do not allow liquid to spill.)
Item 6. Holding a comb and combing hair at back of head.
Reprinted from Carr, J. H., Shepherd, R. B., Nordholm, L., & Lynne, D. (1985). Investigation of a new motor assessment scale for stroke patients. Physical Therapy, 65, 179 (Public Domain).