Objective: To establish Rasch validation of the Lucerne International Classification of Functioning, Disability and Health (ICF)-based Multidisciplinary Observation Scale (LIMOS) in stroke. LIMOS measures the level of assistance in daily life activities related to motor function, communication, cognition, and domestic life. Rasch Measurement Theory (RMT) transforms an ordinal scale into an interval scale and thus the Rasch-based LIMOS scale captures a more accurate improvement of functional outcomes via Rasch-transformed scores.
Design: Cross-sectional study design.
Setting: Neurorehabilitation center, Luzerne, Switzerland.
Participants: We recruited participants with stroke (N=407; age=63.2±16.0y; n=157 women) through consecutive sampling.
Interventions: Not applicable.
Main Outcome Measures: Participants were assessed with LIMOS at admission and discharge of rehabilitation. We used RMT to evaluate overall model fit, response dependency, floor and ceiling effect, reliability, and differential item functioning (DIF) for sex, age, type, and time of stroke on the 4 LIMOS subscales using the Rasch Unidimensional Measurement Model (RUMM) 2030 program.
Results: The Rasch-based LIMOS subscales fit the Rasch model after reducing and rescoring items: motor (from 20 to 18 items), communication (5 items), cognition (from 15 to 13 items), and domestic life (5 items). There was no floor or ceiling effect. Some artificial DIF was identified. Scoring at discharge was dependent on the scoring responses at admission, which means that without applying a correction factor to the discharge scores, there was an underestimation of change in scores between admission and discharge, ranging from 0.24 to 0.97 logits (10.79%-49.24%) on the different subscales.
Conclusions: The Rasch-based LIMOS scale is recommended to measure functional outcome in people with acute or chronic stages of ischemic or hemorrhagic stroke.
Download full-text PDF |
Source |
---|---|
http://dx.doi.org/10.1016/j.apmr.2019.02.012 | DOI Listing |
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!