Construct Validity of the Late-Life Function and Disability Instrument in African American Breast Cancer Survivors

Limited data exist on the validity of the Late-Life Function and Disability (LLFD) instrument in cancer survivors. We examined the construct validity of the abbreviated LLFD instrument in a sample of African-American breast cancer survivors. African American breast cancer survivors (n = 181) aged 50 years and older completed the abbreviated LLFD instrument and questions about sociodemographic and lifestyle characteristics. Confirmatory factor analysis (CFA), Cronbach alphas, and structural models were used to evaluate the construct validity of these measures. Minor modifications were made to the three-factor functional component portion of the inventory to improve model fit. Cronbach alpha's (range 0.85-0.92) and inter-factor correlations (r = 0.3-0.5, all p < 0.05) were appropriate. The two-factor disability component fit the data and Cronbach alpha's (0.91 and 0.98) were appropriate with a high inter-factor correlation (r = 0.95, p < 0.01). The average variance extracted (range = 0.55-0.93) and composite reliabilities (range = 0.86-0.98) were in acceptable ranges. Floor effects ranged from 7% for advanced lower function to 74% for personal role disability. Education and number of comorbidities were correlated significantly with functional outcomes. The abbreviated LLFD instrument had adequate construct validity in this sample of African American breast cancer survivors. Further studies are needed that examine the stability of the instrument over time.