Improving the evaluation of therapeutic interventions in multiple sclerosis: the role of new psychometric methods.
By J. Hobart, S. Cano. Health Technology Assessment, 2009
614 citations recorded by the source corpus.
OBJECTIVES In this monograph we examine the added value of new psychometric methods (Rasch measurement and Item Response Theory) over traditional psychometric approaches by comparing and contrasting their psychometric evaluations of existing sets of rating scale data. We have concentrated on Rasch measurement rather than Item Response Theory because we believe that it is the more advantageous method for health measurement from a conceptual, theoretical and practical perspective. Our intention is to provide an authoritative document that describes the principles of Rasch measurement and the practice of Rasch analysis in a clear, detailed, non-technical form that is accurate and accessible to clinicians and researchers in health measurement. REVIEW METHODS A comparison was undertaken of traditional and new psychometric methods in five large sets of rating scale data: (1) evaluation of the Rivermead Mobility Index (RMI) in data from 666 participants in the Cannabis in Multiple Sclerosis (CAMS) study; (2) evaluation of the Multiple Sclerosis Impact Scale (MSIS-29) in data from 1725 people with multiple sclerosis; (3) evaluation of test-retest reliability of MSIS-29 in data from 150 people with multiple sclerosis; (4) examination of the use of Rasch analysis to equate scales purporting to measure the same health construct in 585 people with multiple sclerosis; and (5) comparison of relative responsiveness of the Barthel Index and Functional Independence Measure in data from 1400 people undergoing neurorehabilitation. RESULTS Both Rasch measurement and Item Response Theory are conceptually and theoretically superior to traditional psychometric methods. Findings from each of the five studies show that Rasch analysis is empirically superior to traditional psychometric methods for evaluating rating scales, developing rating scales, analysing rating scale data, understanding and measuring stability and change, and understanding the health constructs we seek to quantify. CONCLUSIONS There is considerable added value in using Rasch analysis rather than traditional psychometric methods in health measurement. Future research directions include the need to reproduce our findings in a range of clinical populations, detailed head-to-head comparisons of Rasch analysis and Item Response Theory, and the application of Rasch analysis to clinical practice.
Subfields: Education, Occupational Therapy, Pathology and Forensic Medicine
View Paper: https://www.semanticscholar.org/paper/2f17618556f8b0f11d05f13faf9e03eb40d8876a
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Education
In Occupational Therapy
In Pathology and Forensic Medicine
Citations in the corpus
- Testing Rating Scale Unidimensionality Using the Principal Component Analysis (PCA)/t-Test Protocol with the Rasch Model: The Primacy of Theory over Statistics (2014, methodology): “As reviewed above, the most important aspect of the dimensionality issue relates to the central role of variable definition, frame of reference, perspective and context of application, and the fact that unidimensionality is not an “either/or” issue but a relative matter of degree [10] [12] [16].”