Rasch‐ionale for neurologists

Abstract

Outcome measures are considered the most important tools to monitor patients' outcome in both clinical and research settings. Measuring the clinical state of patients is a fundamental part of our daily clinical practice and research that sometimes is taken for granted. In peripheral neuropathies, there are many scales available, but most of these are at the ordinal level. This paper will systematically address the types of scales available (being nominal, ordinal, interval, or ratio data-based) in terms of their strengths and weaknesses. The differences between classical test theory-based and modern test method-based outcome measures will be addressed with emphasis on Rasch methodology. Various steps will be highlighted as part of the evaluation and construction of outcome measures using the Rasch method, with the aim to increase the knowledge and utility of this technique. We argue that Rasch-built outcome measures should be used for future studies in neuromuscular disorders and t... Abstract Outcome measures are considered the most important tools to monitor patients' outcome in both clinical and research settings. Measuring the clinical state of patients is a fundamental part of our daily clinical practice and research that sometimes is taken for granted. In peripheral neuropathies, there are many scales available, but most of these are at the ordinal level. This paper will systematically address the types of scales available (being nominal, ordinal, interval, or ratio data-based) in terms of their strengths and weaknesses. The differences between classical test theory-based and modern test method-based outcome measures will be addressed with emphasis on Rasch methodology. Various steps will be highlighted as part of the evaluation and construction of outcome measures using the Rasch method, with the aim to increase the knowledge and utility of this technique. We argue that Rasch-built outcome measures should be used for future studies in neuromuscular disorders and their method of construction could be easily extrapolated to other neurological illnesses.

Impairment measures versus inflammatory RODS in GBS and CIDP: a responsiveness comparison

“…The Rasch methodology has been described in detail in this journal and elsewhere (Rasch, 1960;Tennant and Conaghan, 2007;Merkies et al, 2010;van Nes et al, 2011;Vanhoutte et al, 2015). The data for all selected outcome measures were subjected to Rasch analysis (software Rasch unidimensional measurement model, RUMM2030) using the following person factors (age categories: <50/50-64/more than 65 years; gender: males vs. females, illness: GBS vs. CIDP, visits: T0 vs. T1 vs. T3 vs. T6 vs. T12 months) (Andrich et al, 2010).…”

Future needs in peripheral neuropathy outcome measures

“…Linear Rasch scales are suitable for conventional statistical analysis and allow for the creation of sum scores. Using sum scores based on ordinal scales is not legitimate as Rasch model application reveals that the sum of the parts does not equal the whole (Stucki et al, 1996;Vanhoutte et al, 2015a;2015c). They may give a false sense of precision and lead to the impression of a clinically meaningful improvement, when one does not really exist.…”

“…Using the Rasch model, it was demonstrated that across the range of an interval/linear outcome measure, the standard error (SE) varies (U-shape). This has led to the MCID-SE concept taking into account the dynamic pattern of responsiveness of individual patients (Hobart and Cano, 2009;Vanhoutte et al, 2015a;2015c).…”

“…Because the clinical features of MMN are different from other immune-mediated neuropathies, a Rasch-built Overall Disability Scale specific for MMN (MMN-RODS) was created (Vanhoutte et al, 2015a;2015c). The scale was validated in 26 patients from seven European and North American countries by comparison with the ODSS.…”

Minimum clinically important difference analysis confirms the efficacy of IgPro10 in CIDP: the PRIMA trial

“…While grip strength evaluates focal impairment, in the current study of the dominant hand, the INCAT and MRC scores provide a more overall dysfunction of the patients examined (Léger et al, 2013). In addition, the stringent cut-off used for grip strength could have led to a lower MCID response when compared with the ordinal-based INCAT and MRC measures, the scores of which might be inflated (Tennant and Conaghan, 2007;Marais and Andrich, 2008;Léger et al, 2013;Vanhoutte et al, 2015). The small sample size and possibly non-uniform improvement across muscle groups might have also contributed to the differences.…”