Factor analysis techniques for assessing sufficient unidimensionality of cancer related fatigue

Jin-Shei Lai*, Paul K. Crane, David Cella

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

80 Scopus citations

Abstract

Background: Fatigue is the most common unrelieved symptom experienced by people with cancer. The purpose of this study was to examine whether cancer-related fatigue (CRF) can be summarized using a single score, that is, whether CRF is sufficiently unidimensional for measurement approaches that require or assume unidimensionality. We evaluated this question using factor analysis techniques including the theory-driven bi-factor model. Methods: Five hundred and fifty five cancer patients from the Chicago metropolitan area completed a 72-item fatigue item bank, covering a range of fatigue-related concerns including intensity, frequency and interference with physical, mental, and social activities. Dimensionality was assessed using exploratory and confirmatory factor analysis (CFA) techniques. Results: Exploratory factor analysis (EFA) techniques identified from 1 to 17 factors. The bi-factor model suggested that CRF was sufficiently unidimensional. Conclusions: CRF can be considered sufficiently unidimensional for applications that require unidimensionality. One such application, item response theory (IRT), will facilitate the development of short-form and computer-adaptive testing. This may further enable practical and accurate clinical assessment of CRF.

Original languageEnglish (US)
Pages (from-to)1179-1190
Number of pages12
JournalQuality of Life Research
Volume15
Issue number7
DOIs
StatePublished - Jan 1 2006

Keywords

  • Bi-factor analysis
  • Cancer-related fatigue
  • Dimensionality
  • Factor analysis

ASJC Scopus subject areas

  • Public Health, Environmental and Occupational Health

Fingerprint Dive into the research topics of 'Factor analysis techniques for assessing sufficient unidimensionality of cancer related fatigue'. Together they form a unique fingerprint.

Cite this