Abstract Background: This research examined the psychometric properties of the newly revised Cervical Cancer Literacy Assessment Tool (C-CLAT) in a diverse sample (N = 543) of Black, Latina and Arab women in the United States. This study was in response to the Institute of Medicine's report Health Literacy A Prescription to End Confusion (2004) that indicated a gap in health literacy measurements. Methods: Data to validate the C-CLAT were collected in the context of a community based research intervention, Kin KeeperSM Cancer Prevention Intervention. Twentyone community health workers (CHWs) who reflected the ethnicity of the research participants, recruited women from their respective public health programs. The CHWs administered the C-CLAT in the homes of their clients. The C-CLAT is a multidimensional scale consisting or 21 items in three health literacy domains, Awareness, Knowledge & Screening, Prevention & Control. Modern psychometric methods were used including multigroup confirmatory factor analysis and analysis of differential item function. Practical aspects include interpretation of scores from a criterion-referenced framework. Results: The overall scale difficulty was .33 or 67.24% (SD = 18.78) correct. This score represents the percent of cervical cancer literacy knowledge attained at the time of testing. This score was sufficiently low so that knowledge improvements can be monitored without ceiling effects. Individual scales, were similar in difficulty level, with 63.12%, 62.15%, and 65.23% correct, for Awareness, Knowledge & Screening, Prevention & Control, respectively. The overall scale reliability was .65 in the total sample and .68, .67, and .55 for Black/African American, Latina/Hispanic, and Arab American, respectively. The subscales were moderately correlated (r's =, .26, .38, .49), each contributed significantly to the overall functional literacy score, and factor to scale relations did not differ across ethnic groups. In addition, the rank order correlations of item difficulties across ethnic groups were high (r's = .97, .98, .98). One subscale, Awareness, appeared to be easier for Hispanics. The C-CLAT total score did not differ by age, education, income, self-rated health, or insurance coverage. Divorced or separated women had higher scores (p < .05), and unemployed or self employed women had the lowest scores (p < .05). Two trends in the data were only marginally significant. There was a tendency for women with a history of cervical cancer in their family (n = 27) to have lower C-CLAT scores (p = .098) and for women with any cancer history in their family (n = 210) to have higher C-CLAT scores (p = .073). Conclusions: The psychometrics analysis supports the use of the revised C-CLAT for use in Black/African American, Latina/Hispanic, and Arab American populations. Additionally this study demonstrated that psychometric evaluation can be conducted within the context of community based methods. This study was supported by the NIH/NINR 1R21NR010366. Citation Information: Cancer Epidemiol Biomarkers Prev 2011;20(10 Suppl):A5.
Read full abstract