Rasch-based scoring offered more precision in differentiating patient groups in measuring upper limb function

Asaduzzaman Khan, Chi-Wen Chien, Sandra G. Brauer

Research output: Journal article publicationJournal articleAcademic researchpeer-review

34 Citations (Scopus)


Objective: To compare the discriminatory ability of Rasch-based and summative scoring in the context of assessing upper limb function of patients with stroke. Study Design and Setting: Data were from a cohort study of 497 adults with stroke undergoing physiotherapy. Upper limb function was assessed at admission and discharge using the upper limb subscale of the Motor Assessment Scale (UL-MAS). Rasch analysis was used to transform raw UL-MAS scores into interval measures. A relative precision (RP) index was used to differentiate patients by discharge destination. Results: The analysis confirmed the unidimensional structure of UL-MAS at both admission and discharge and demonstrated the adequate fit of the items. The RP index favored the Rasch-based scoring over the summative scoring in differentiating between the two patient groups, with significant gains in precision at admission (15%) and discharge (11%). When examining patients in the upper or lower quartile of UL-MAS, the gains in precision were statistically significant in favor of the Rasch-based scoring, with 20% precision at admission and 19% precision at discharge. Conclusion: Rasch-based scoring was more precise in differentiating patient groups by discharge destination than the summative scoring used to measure upper limb function, especially at the extreme range of the scale.
Original languageEnglish
Pages (from-to)681-687
Number of pages7
JournalJournal of Clinical Epidemiology
Issue number6
Publication statusPublished - 1 Jun 2013
Externally publishedYes


  • Item response theory
  • Patients with stroke
  • Rasch analysis
  • Relative precision
  • Summative scoring
  • UL-MAS

ASJC Scopus subject areas

  • Epidemiology


Dive into the research topics of 'Rasch-based scoring offered more precision in differentiating patient groups in measuring upper limb function'. Together they form a unique fingerprint.

Cite this