Unidimensional Scale for Assessing Mobility Disability in Multiple Sclerosis - Prof. Gonzá, Apuntes de Psicometría

¡Descarga Unidimensional Scale for Assessing Mobility Disability in Multiple Sclerosis - Prof. Gonzá y más Apuntes en PDF de Psicometría solo en Docsity! Unidimensional Scale for Disability Assessment in MS 1 1 The Development of an Unidimensional Scale of the Guttman Type for the Assessment of Mobility Disability in Multiple Sclerosis. L H DE SOUZA, PROFESSOR OF REHABILITATION, CENTRE FOR RESEARCH IN REHABILITATION, DEPARTMENT OF HEALTH STUDIES, BRUNEL UNIVERSITY, OSTERLEY CAMPUS, ISLEWORTH, MIDDLESEX TW7 5DU. UK. Address for Correspondence: Professor L H De Souza, Centre for Research in Rehabilitation, Brunel University e-mail: Lorraine.desouza@brunel.ac.uk Running Title: Unidimensional Scale for Mobility Disability Assessment in MS ABSTRACT Unidimensional Scale for Disability Assessment in MS 2 2 Objective: The aim of the study was to develop a valid and reliable unidimensional scale of the Guttman type for the assessment of mobility disability in multiple sclerosis (MS). Subjects: Sixty-eight subjects with a definite diagnosis of MS participated.They were attending as outpatients at a MS unit at a District General Hospital. Thirty had the primary progressive pattern of disease, and 38 had the relapsing-remitting pattern. Methods: Formal assessments used for neurological disability were inspected, and 14 test items of gross motor function were extracted and ordered according to two criteria. These were that actions progressed from lying, to sitting, to standing and walking tasks, and that they progressed from broader to narrower bases of support. All subjects carried out all test items which were scored as ‘pass’ or ‘fail’. Analysis: Data were tested for internal consistency, reliability, inter item correlation, reproducibility and scalability. On the basis of the results, the items were re-ordered in rank, and reduced to eleven tests. The eleven item scale was re-analysed. Results: Results showed that the scale had an internal consistency of 0.88 (alpha coefficient) and a coefficient of reproducibility (CR) of 0.95 and above for both MS subject groups. The coefficient of scalability (CS) for items was 0.78 for primary progressive subjects and 0.74 for the relapsing-remitting group. Reliability ranged from good (kappa = 0.49) for one item, to perfect for six items. Conclusion: The scale was demonstrated to be a hierarchical scale of the Guttman type exhibiting homogeneous unidimensionality and good reliability. The high CR indicated that scores may be summed, and the very acceptable levels of CS indicated that the cumulative scores are meaningful within the defined concept of hierarchy used in this study. Unidimensional Scale for Disability Assessment in MS 5 5 assessment of neurological disability by a number of authors who have developed formal assessment tools. Some test items, when included in various assessments for stroke, require the action to be performed towards affected and unaffected sides of the body. This is not a valid distinction in MS, and such actions were therefore required to be carried out both unilaterally, and bilaterally. The method of assessment for each of the fourteen test items has been published in more detail elsewhere 25. The 14 test items were arranged in a rank order similar to the neurodevelopmental sequence of motor control 9 10. The rank order had regard for two features of the sequence of actions. Firstly, the activities progressed from lying, to sitting, to standing and walking tasks, and secondly, the actions progressed from broader to narrower bases of support. In all 68 people with a definite diagnosis of MS consented to participate and were assessed. Thirty had the primary progressive pattern of MS and 38 had the relapsing-remitting pattern. They were attending a MS Clinical and Research Unit at a District General Hospital having been referred for assessment and advice for their MS. Patients carried out all 14 test items three times at yearly intervals. They either passed (score = 1), or failed (score = 0) each item, thus data were encoded in binary form at the time of collection. Inter-observer reliability was tested between two observers, with experience in carrying out the 14 item assessment, on 20 of the MS patients with stable disease. Testing was carried out according to the published protocol 25. All assessments were carried out in a physiotherapy gymnasium using a consistent test environment. The walking tests were carried out using the same outdoor route (cement surfaced) marked at 20 and 50 metres. Patients were tested for walking once with those progressing past 20 metres walking onwards to attempt 50 metres. ANALYSIS Unidimensional Scale for Disability Assessment in MS 6 6 Inter observer reliability was determined using weighted kappa. The correlation between each scale item on the remaining items in the scale was based on linear multiple correlation to estimate the lower bound of reliability (or lamda6) for items26, and on inter item correlation co-efficients to determine relationships between items. Internal consistency was estimated using Cronbach’s co-efficient alpha27. Co- efficients of Reproducibility and Scalability were calculated using published methods29. Analyses were carried out using SPSS - X release 3.1 for VAX/VMS. RESULTS The lower bound of reliability 26 based on linear multiple correlation for the 14 items was 0.95 indicating high correlation between each item in the scale on the remaining items. For the inter observer reliability estimation, there were 12 disagreements in scoring out of the 280 paired observations. The greatest number of disagreements was for the test of unilateral lower limb stance (item 14) which had four disagreements (weighted kappa = 0.49; ‘moderate’). Item 4 had three disagreements (kappa =0.70), items 3 and 13 each had 2 disagreements (kappa = 0.79 and 0.62 respectively; ‘good’), while item 2 had only one disagreement (kappa = 0.85; ‘very good’). The co-efficient alpha (internal consistency) for the total data was 0.90. A high alpha indicated that there was at least one homogeneous dimension underlying the scores, and that items correlate with one another. A minimum acceptable alpha of 0.70 has been suggested for group data 28. Thus, the indication was that the selected items essentially measured the same construct, and that the construct itself was homogeneous. Co-efficients of Reproducibilty and Scalability A Guttman Scale 29 is one in which a cumulative scale is demonstrated to consist of a hierarchy of items which are homogeneous and relate to one concept. The Co- Unidimensional Scale for Disability Assessment in MS 7 7 efficient of Reproducibility (CR) quantifies the deviation of a scale from perfection29. However, as perfection is rarely achieved, an overall CR of 0.90 has been reported as indicating the existence of a valid, cumulative, and unidimensional Guttman Scale 30. The CR calculated for the primary progressive and the relapsing-remitting patient groups, were 0.93 and 0.95 respectively (Table 2). The CR is considered insufficient on its own for indicating a valid scale 31. It has been shown that a large number of subjects who saturate the scale at the top or the bottom (extreme individuals) could produce a high CR. A similar outcome could be produced if several of the scale items were passed or failed by the majority of subjects (extreme items). Determination of the Co-efficient of Scalability (CS) 31 makes it impossible to erroneously attribute a high scalability to a set of tests in a sample containing many extreme items or many extreme individuals. CS is always lower than CR, but a level of acceptance for CS has been suggested at between 0.60 and 0.65 31. The CS for the two patient groups, by items and by subjects, indicated that an acceptable level of scalability was exhibited to proceed (Table 2). In order to determine if the individual items in the scale were ranked in the most appropriate hierarchy, the frequency profile for scale errors by item (that is, the sum of the frequency of the non-modal group) was inspected for each group of patients. The frequency of scale errors indicated that in both patient groups two test items (2 and 1) had more scale errors than items placed above, indicating that they were inappropriately placed in the hierarchy of items. The items were re-ordered so that 2 and 1 rose in the rank by three places to lie after item 6 and before item 5. The CS and CR were then recalculated. The results showed that the re-ordered scale items produced higher CS and CR than the original rank order (Table 3). In order to determine if all 14 test items were required, inter-item correlation coefficients (ICC) were determined using the total data. Highly significant correlations were found between standing unsupported for 60 seconds and for 30 seconds (ICC Unidimensional Scale for Disability Assessment in MS 10 10 Acknowledgements. The author thanks Dr C J Partridge for helpful advice and discussion, and all the people with multiple sclerosis who participated in the study. Unidimensional Scale for Disability Assessment in MS 1 1 Table 1 The 14 test items and the other assessments which incorporate each ITEM No. TEST ASSESSMENT REFERENCE 14. Stand on L. and R. lower limbs 21, 34, 35, 36 with no upper limb support (5 secs each) 13. 50 metres unaided walk 13 (walk =40 m), 12. Transfer from the floor up to 21, 34, standing with no upper limb aid. 11. Transfer from standing down to 21 the floor with no upper limb aid. 10. Transfer from unsupported sitting 13, 21, 33, 37, 38 to standing with no upper limb aid. 9. 50 m aided walk 39 8. 20 m aided walk 3, 21 (walk=15m), 40 7. Stand no support (60 sec) none 6. Stand no support (30 sec) 21, 41 5. Transfer from bed 21, 33, 42, 43, 44, 45 to chair without upper limb aid 4. Transfer from chair 21, 42 to bed without upper limb aid. 3. Sit no support (60 sec) 21, 33 (time=2 min), 37, 46 2. Roll from supine to L. and 13, 21, 33, 37, 46 then to R. side lying. 1. Roll from supine 13, 21, 33, 37, 46 lying to one side. Unidimensional Scale for Disability Assessment in MS 1 Table 2 Co-efficients of Scalability (CS) and Reproducibility (CR): 14 test items at Stage 1 Patient Group CS Items CS Subjects CR Primary progressive 0.68 0.71 0.93 (N = 30) Relapsing-remitting 0.71 0.62 0.95 (N = 38) 1 Unidimensional Scale for Disability Assessment in MS 4 4 Impairments, Disabilities and Handicaps, WHO, Geneva. 2 Thompson, A.J., Kermode, A.G., MacManus, D.G., et al (1990) Patterns of disease activity in multiple sclerosis: Clinical and magnetic resonance imaging study. Brit. Med. J. 300: 631-4. 3 Kurtzke, J.F. (1983) Rating neurological impairment in multiple sclerosis. An expanded disability scale (EDSS). Neurol. 33: 1444-52. 4 Willoughby, E.W., and Paty, D.W. (1988) Scales for rating impairment in multiple sclerosis: a critique. Neurol. 38: 1793-89. 5 Scheinberg, L. C., and Raymond, C. A. (1989) Quantitative techniques in the assessment of multiple sclerosis: The problems and current solutions. In: Quantification of Neurologic Deficit (Ed. T.L. Munsat) Butterworths, MA., 155-162. 6 Wade, D.T. (1992) Measurement in Neurological Rehabilitation. Oxford University Press, 1992. 7 Grillner, S. (1975) Locomotion in vertebrates: central mechanisms and reflex interaction. Physiol. Rev. 55: 247-304. 8 Hughlings Jackson, J. (1882) On some implications of dissolution of the nervous system. Med. Press Circular 2: 411-413. 9 Levitt, S. (1977) Treatment of cerebral palsy and motor delay. Blackwell Scientific Publ. Ltd., Oxford. 10 Bobath, B. (1971) Abnormal postural reflex activity caused by brain lesions. Unidimensional Scale for Disability Assessment in MS 5 5 William Heinemann Medical Books Ltd., London. 11 Purpura, D.P., Shofer, R.J., Housepain, E.M., and Noback, C.R. (1964) Comparative ontogenesis of structure-function relations in cerebral and cerebellar cortex. Prog. Brain Res. 4: 187-221. 12 Zulch, K.J. (1975) Pyramidal and parapyramidal motor systems in man. In: Cerebral Localization (Eds: K.J. Zulch, O. Creutzfeldt, and G.C. Galbraith) Springer Verlag, Berlin. 13 Lincoln, N.B., and Leadbitter, D. (1979) Assessment of motor function in stroke patients. Physiotherapy 65: 48-51. 14 Denny-Brown, D. (1962) The basal ganglia and their relation to disorders of movement. Oxford University Press, Oxford. 15 Martin, J.P. (1967) The Basal Ganglia and Posture. Pitman Medical, London. 16 Marsden, C.D. (1994) Parkinson's disease. J. Neurol. Neurosurg. Psychiat. 57: 672-681. 17 Brooks, V.B., and Thach, W.T. (1981) Cerebellar control of posture and movement. In: Handbook of Physiology, The Nervous System, Vol. II, Part 2 (Ed. V.B. Brooks). Amer. Physiol. Soc., Bethesda, MA, 877-946. 18 Ito, M. (1984) The Cerebellum and Neural Control. Plenum Press. 19 Matthews, W. B. (1991) In: McAlpine's Multiple Sclerosis. 2nd Ed (Ed: W.B. Unidimensional Scale for Disability Assessment in MS 6 6 Matthews) Churchill Livingstone, Edinburgh and London. 20 Worthington, J.A., and De Souza, L.H. (1989) A simple measurement of speed of index finger movement. Clinical Rehabil. 3: 117-123. 21 Ashburn, A.(1982) Assessment of motor function in stroke patients. Physiotherapy 68: 109-13. 22 Lawrence, D.G., and Kuypers, H.G.J.M. (1968) The functional organisation of the motor system in the monkey: I. The effect of bilateral pyramidal lesions; II. The effects of lesions of the brain-stem pathways. Brain 91: 1-33. 23 Kuypers, H.G.J.M. (1973) The anatomical organization of the descending pathways and their contributions to motor control especially in primates. New Developments E.M.G. and Clin. Neurophysiol. 3: 38-68. 24 Brinkman, J., and Kuypers, H.G.J.M. (1973) Cerebral control of contralateral and ipselateral arm, hand and finger movements in the split-brain rhesus monkey. Brain 96: 653-674. 25 De Souza L H and Ashburn A (1996) Assessment of motor function in multiple sclerosis patients. Physiotherapy Research Internat 1: 98-111. 26 Guttman L (1945) A basis for analyzing test-retest reliability. Psychometrika 10: 255- 282. 27 Cronbach, L. (1951) Coefficient alpha and the internal structure of tests. Psychometrika 16: 297-334. 28 Nunnally, J.C. (1978) Psychometric Theory. McGraw-Hill, New York.