Oral Presentation COSA-IPOS Joint Scientific Meeting 2012

Psychometric properties of the Cancer Health Literacy Test (#66)

Levent Dumenci 1 , Laura A Siminoff 1 , Robin K Matsuyama 1
  1. Virginia Commonwealth University, Richmond, VA, United States

Aims: To develop an instrument to measure cancer health literacy of patients diagnosed with cancer and test its psychometric properties.
Methods: Delphi panel, focus group, and cognitive interviewing techniques are used to define the cancer health literacy construct, develop items with a wide content coverage. Factor analysis and the two-parameter item response theory model (2PL) are used to test internal structure. Cronbach alpha and two-week test retest reliability coefficients are used to estimate scale reliability.
Results: Using data from 895 cancer patients, exploratory factor analyses support a unidimensional measurement structure with standardized factor loadings ranging from .34 to .89. The 65-item test has a Cronbach’s Alpha internal consistency of 0.92 and a two-week test-retest reliability of 0.93. A two-parameter logistic item response model indicates that the test information function is maximized around 1.5 standard deviations below the mean. Correlations with instruments designed to measure general health literacy range from 0.49 to 0.64.
Conclusions: The Cancer Health Literacy Test is a unique instrument with strong psychometric properties specifically designed to measure the cancer health literacy levels of individuals with a cancer diagnosis. Results from the validation studies strongly support the appropriateness of test score interpretations among individuals with cancer diagnoses. Precision of measurement at low levels of cancer health literacy provides a strong argument for clinical utility. Study implications in epidemiologic and clinical studies and future directions are discussed.