A further Rasch study confirms that ALSFRS-R does not conform to fundamental measurement requirements

Abstract

Our objective was to verify and expand previous evidence of psychometric inadequacies in the ALSFRS-R, in a different sample of subjects suffering from ALS. Since 2009, a prospective registry records all incident cases of ALS in Emilia Romagna Region, Italy (4.4 million inhabitants) referred to its 17 neurological departments. For each patient, demographic and clinical information is collected by the physician in charge, including compilation of the ALSFRS-R at each clinical follow-up. Results showed that a confirmatory factor analysis on the three-factor model previously found (bulbar, motor, respiratory function) showed a good fit. Rasch analysis on the whole scale showed the need to collapse some rating categories, confirmed the multidimensionality of the ALSFRS-R, and demonstrated the presence of differential item functioning between patients with spinal versus bulbar onset. Moreover, some items included in the three ALSFRS-R subscales showed a problematic fit to the respective construct they were intended to measure. In conclusion, the interpretation of a total raw score of ALSFRS-R is hampered by ambiguities due to the different metric properties of the three domains the scale aggregates, and their content and structure. This study confirms that a refinement of ALSFRS-R is warranted, pointing to the need to revise its whole structure, and providing detailed guidelines for its revision.

Key words::

• Amyotrophic lateral sclerosis
• ALSFRS-R
• Rasch analysis
• population based study

Acknowledgements

J. Mandrioli has received research support from Regione EmiIia Romagna (Programma di ricerca Regione-Università 2010–2012, area 2, Ricerca per il Governo Clinico). ALS Registry is supported by a grant from the Emilia Romagna Regional Health Authority.

Declaration of interest: The authors have nothing to disclose.