Citation
Allen, Chenoa & McNeely, C. (2015). Self-rated Health in Disparities Research: Construct Validity of SRH Across Racial/Ethnic Groups and Immigrant Generations Among U.S. Adolescents and Young Adults. Society for Adolescent Health and Medicine. Los Angeles, CA.Abstract
Purpose: Single - item self - rated health (SRH) measures, in which respondents rate their overall health on a four - or five - point ordinal scale, are often the only health - related questions in large national surveys. SRH is a good proxy for hysical and mental health, and it strongly predicts future health and mortality. Studies in adults suggest, however, that the construct of SRH is not equivalent across racial/ethnic groups or immigrant generations and thus may not be valid for health disparities research. No studies have examined whether SRH is an equivalent construct across subgroups for adolescents/young adults. This study uses data from waves 3 (ages 18 - 24) and 4 (ages 24 - 32) of the National ongitudinal Study of Adolescent Health (Add Health) to assess the construct validity of a single - item 5 - category SRH measure across six racial/ethnic groups (Latino, non - Hispanic White, non - Hispanic Black, Asian, Native American/Alaska Native (AI/AN), and multiracial) and immigrant generations (first, second, and third - plus) among US adolescents/young adults. Methods: Final sample sizes were 14015 for wave 3 and 14493 for wave 4. We combined health indicators into dichotomous variables indicating the presence of (1) physical health conditions (chronic conditions, activity limitations, and obesity) and (2) depression (score of 11 or greater on a 9 - item CESD). Models treating each physical condition as an independent variable yielded similar results. Control variables included wave 1, 3, and 4 income; age; gender; and national origin. Using linear regression, we examined how well physical health and depression predicted SRH. We used interaction terms (physical or mental health*race/ethnicity or generation) to determine whether physical health and depression are comparably associated with SRH for all subgroups. We then conducted subgroup analyses for Latinos (n=2278) and Asians (n=959) to examine validity of SRH across generations, within race/ethnicity. Finally, to determine whether physical health and depression explain similar amounts of variance in SRH across racial/ethnic and immigrant subgroups, we conducted linear regression stratified by subgroup, using all health variables, and compared R - squared values. Results: All measures of health were significantly associated with SRH (p<.001); SRH decreased with poorer health. There were no significant interaction terms, indicating that physical and mental health conditions equivalently predicted SRH across racial/ethnic and immigrant generation subgroups. For the full sample, number of chronic conditions, activity limitations, obesity, and depression explained 15.1% of the variance in SRH in wave 3 and 19.6% of the variance in SRH in wave 4. Wave 3 R - squared ranged from .087 (for AI/AN) to .22 (for multiracial); wave 4 R - squared ranged from .15 (for Black) to .28 (for multiracial). Ninety - five percent confidence intervals for racial/ethnic groups overlapped. Health predictors explained a smaller proportion of variance for first generation immigrants (5 - 12% in wave 3; 9 - 17% in wave 4). Conclusions: In this nationally representative dataset, SRH demonstrates good construct validity across immigrant generations and racial/ethnic groups. Therefore, SRH can validly be used for health disparities research among adolescent/young adult populations. This study goes beyond previous studies by analyzing SRH validity for subgroups of adolescents and younger adults and by including American Indians/Alaska Natives and multiracial respondents.URL
http://www.sciencedirect.com/science/article/pii/S1054139X14004352Reference Type
Conference proceedingBook Title
Society for Adolescent Health and MedicineAuthor(s)
Allen, ChenoaMcNeely, C.