Standardization of the Cancer Health Literacy Test – 6 in a non-cancer populations

Abstract

Abstract Background The Cancer Health Literacy Test-6 (CHLT-6) was originally developed in a cancer population to identify persons with limited cancer health literacy (LCHL). There is no instrument to identify individuals with LCHL in non-cancer populations such as caregivers and epidemiologic samples. This study aims to test if the CHLT-6 can be used to identify persons with LCHL in non-cancer patient populations. The CHLT-6 was administered using a hand-held touch-screen computer. Methods A patient sample of 512 persons was used to test the CHLT-6 measurement structure and externally validate the test. English speaking, aged 18 or higher, and no history of cancer diagnosis were used as eligibility criteria. Participants were recruited from health fair grounds, primary care clinics, and community centers in the Mid-Atlantic region of the U.S.A. Consistent with the analysis of the developmental sample, discrete latent variable modeling was used as the underlying measurement structure of the instrument. Results Two classes were clearly separated by the CHLT-6 items. Measurement invariance tests showed equivalence of conditional probabilities between gender, race/ethnicity, and educational groups. There were significant group differences in prevalence of LCHL. The CHLT-6 was externally validated using race/ethnicity, education, and income. Overall results were very similar to those reported in the cancer population. Conclusions The CHLT-6 is the first and only instrument to identify individuals with LCHL using a discrete latent variable framework. Measurement invariance tests provided empirical evidence that labels used to describe latent classes (i.e., limited and adequate cancer health literacy) have the same meaning between gender, race/ethnicity, and educational groups. Persons self-described as African Americans and low educational attainments have a high probability of belonging to the LCHL class. Key messages There is a strong support for psychometric properties of the CHLT-6 in a cross-validation sample. This study provides empirical basis for using the CHLT-6 in non-cancer populations.