Paper
Saturday, July 16, 2005
Testing a Quality of Life Model in Latinos Living With HIV/AIDS
Sheila S. Tann, RN, ND1, Karen H. Sousa, RN, PhD1, and Oi-man Kwok, MS2. (1) College of Nursing, Arizona State University, Tempe, AZ, USA, (2) Department of Psychology, Arizona State University, Tempe, AZ, USA
| Learning Objective #1: Understand the use of Structural Equation Modeling for the purposes of model testing and validating outcome measures |
| Learning Objective #2: Recognize the appropriate use of a quality of life assessment model in both Latino and Caucasian populations |
BACKGROUND: Latinos living in the United States experience infection with Human Immunodeficiency Virus (HIV) at a disproportionately higher rate than the general population. In addition to disparities noted in the epidemiology of HIV infection and AIDS-related deaths in Latinos, disparities are also noted in the manner by which the psychological aspects of living with a chronic, potentially fatal disease are assessed in populations other than Caucasian Americans. Often cultural bias is present in psychometric measures of quality of life (QOL) as a result of methods of survey construction and validity testing. This disjuncture in survey development brings into question the appropriateness of administering them to then broader, ethnically heterogeneous samples. AIMS: This study examines the cross-cultural equivalence of concepts and measurement properties of a QOL model by ethnicity, specifically, Latino and Caucasian men living with HIV/AIDS using Structural Equation Modeling. METHODS: The sample for this secondary analysis consisted of two cohorts (n=118 Latinos, n=658 Caucasians) from the AIDS Time-Oriented Health Outcome Study (ATHOS) databank. Analyses were conducted in two major stages using the MPLUS program and maximum likelihood estimation procedures. RESULTS: Invariance analysis supported the fit of the QOL model in both Latino and Caucasian samples, however the significant chi-square test between weak and strong invariance (p<.01) in both cohorts suggests that the QOL questionnaire should not be used when comparing outcomes between the Latino and Caucasian groups. CONCLUSIONS: This investigation into the appropriateness of a QOL model supports its use within Latino populations. The analysis, however, did not support the use of the model for use in cross-cultural comparisons. Decisions related to treatment and health resource allocation are often based on QOL assessment outcomes. It is vital that these decisions be made with confidence in the fidelity of the measure across cultural groups.