Talking the talk on walking the walk

Abstract

Background

The Multiple Sclerosis Walking Scale (MSWS-12) was developed to measure the impact of multiple sclerosis on walking. Many other disabling neurological conditions affect patients’ ability to walk, and a generic measure of walking could provide valuable insights into patients’ perceptions in clinical trials and epidemiological studies as well as routine clinical practice.

Objective

To evaluate the clinical usefulness and psychometric properties of the Walking Impact Scale (Walk-12), a modified version of the MSWS-12, in patients with neurological conditions.

Design

A prospective, observational study of 120 consecutive patients admitted for rehabilitation. The Walk-12 was used to measure the impact of neurological disability on walking. Traditional psychometric methods (data quality, scaling assumptions, targeting, reliability, validity and responsiveness) were used to assess the Walk-12. Transition questions were used on discharge to measure perception of change. Outcome was also measured using the timed walk test (TWT), Barthel Index (BI) and Functional Independence Measure (FIM).

Results

For the total group, missing data were few, scaling assumptions were satisfied, and internal consistency was 0.94. Correlations between the Walk-12 and TWT, BI and FIM motor score were moderate (r = −0.58, −0.26, −0.31). Responsiveness of the Walk-12 was high (effect size = 1.12). Relationships between effect size and patients’ and physiotherapists’ opinion of change in walking demonstrated good concordance. Preliminary subgroup analyses indicate satisfactory psychometric properties across different neurological conditions; however, sample numbers in these analyses are small.

Conclusions

In this sample of neurologically disabled patients the Walk-12 was clinically useful and satisfied standard psychometric criteria. This provides preliminary evidence that it may be suitable as a generic measure of walking ability.

Appendices

Appendix 1: Walking impact scale (Walk-12)

These questions ask about limitations to your walking due to your illness during the past two weeks. For each statement, please circle the one answer that best describes your degree of limitation. Please answer all questions even if some seem rather similar to others, or seem irrelevant to you. If you cannot walk at all, please tick this box: □

  

In the past two weeks, how much has your illness ...	Not at all	A little	Moderately	Quite a bit	Extremely
Limited your ability to walk?	1	2	3	4	5
Limited your ability to run?	1	2	3	4	5
Limited your ability to climb up and down stairs?	1	2	3	4	5
Made standing when doing things more difficult?	1	2	3	4	5
Limited your balance when standing or walking?	1	2	3	4	5
Limited how far you are able to walk?	1	2	3	4	5
Increased the effort needed for you to walk?	1	2	3	4	5
Made it necessary for you to use support when walking indoors (e.g. holding on to furniture, using a stick, etc)?	1	2	3	4	5
Made it necessary for you to use support when walking outdoors (e.g. using a stick, a frame, etc)?	1	2	3	4	5
Slowed down your walking?	1	2	3	4	5
Affected how smoothly you walk?	1	2	3	4	5
Made you concentrate on your walking?	1	2	3	4	5

1. Please check that you have circled ONE answer for EACH question
2. © 2005 Neurological Outcome Measures Unit, University College London

Appendix 2: Analysis plans

Data quality

The Walk-12 data were examined for percentage missing items and the percentage of the sample for whom total scores could be calculated. For responders with missing items, a total score can be calculated if at least 50% of the items (n ≥ 6) have been completed. Each missing item is replaced with an imputed score, the patient specific mean score, which is the mean score across completed items for that patient [35]. The analyses of percentage missing items did not include imputed items.

Scaling assumptions

It has been proposed by others [24, 30], and is generally accepted [27] that a series of criteria should be satisfied for a set of items to be summed, legitimately, to form a total score. We tested the Walk-12 against these criteria, which are:

• 1. Items should be roughly parallel, that is, measure at the same point on the scale and have similar variance, otherwise they do not contribute equally to the variance of the total score and should be standardised before combination [26]. A set of items is considered parallel when their item response option frequency distributions, and their item mean scores and standard deviations are roughly similar [24].

• 2. Items should measure the same underlying construct, otherwise it is not appropriate to combine them to generate a total score. A set of items is considered to be measuring the same construct when each item’s corrected item-total correlation, which is the correlation between each item and the total score computed from the remaining items in that scale, exceeds 0.30 [26, 30].

• 3. Items in the scale should contain a similar proportion of information concerning the construct being measured. This criterion is considered satisfied the corrected item-total correlations exceed 0.30 [34].

Targeting

Targeting refers to the match between the distribution of walking problems in the sample and the range of walking problems measured by the scale. The better this match, the greater the potential for precise measurement. Targeting was evaluated by examining score distributions, skewness statistics and floor and ceiling effects. Floor effects are the percentage of patients scoring 100 (greatest impact on walking) and ceiling effects are the percentage of patients scoring zero (least impact on walking). It is recommended that floor and ceiling effects should be less than 20% each [20].

Reliability

Scales should generate reliable estimates of the construct being measured (internal consistency). Cronbach’s alpha coefficient was used to determine this criterion [8]. Although a range of minimum values has been suggested, it is widely accepted that Cronbach’s alpha should exceed 0.70 [27] or 0.80 for group comparison studies [30].

Validity

Convergent and discriminant construct validity [4] were evaluated by examining the extent to which Spearman’s correlations between the Walk-12 and the velocity and cadence of walking, the BI, and the FIM motor and cognitive scores were consistent with expectations. It was predicted that correlations between the Walk-12 and the velocity and cadence, BI and FIM motor score would be moderate (between 0.3 and 0.7) and exceed the correlation with the FIM cognitive score. It was also expected that the correlations between the Walk-12 and the velocity and cadence (indicators of walking ability) would be greater than the correlations with the BI and FIM motor score (indicators of overall disability).

Responsiveness

Responsiveness was evaluated for the outcome measures using the effect size statistic, calculated as the mean change from admission to discharge divided by the standard deviation of the admission score [21]. Larger effect sizes indicate greater change measured by the scale. Effect sizes were interpreted using Cohen’s criteria, where 0.20 is a small effect, 0.50 is moderate and 0.80 is large [5]. The significance of change scores for the Walk-12 was also assessed using Wilcoxon Signed Ranks test.

Effect sizes were calculated for the sample as a whole, for each of the main diagnostic subgroups, and for patients in each response category of the change in walking ability transition question [28]. This relates the degree of change measured by the scale with patients’ own opinion of their change in walking ability. Similarly, we compared change detected by the Walk-12 with physiotherapists’ opinions of change in walking ability at discharge.