Two approaches to measuring quality of life in the HIV/AIDS population: HAT-QoL and MOS-HIV

Objectives: Reduce the number of HIV/AIDS-Targeted Quality of Life (HAT-QoL) items, assess psychometric performance of the shortened HAT-QoL, and compare psychometric performance of HAT-QoL to that of Medical Outcomes study HIV Health Survey (MOS-HIV). Design/Subjects: Convenience sample of 215 cross-sectionally studied seropositive individuals. Methods: Subjects completed the HAT-QoL, MOS-HIV, and sociodemographic and disease-specific questions. HAT-QoL and MOS-HIV responses were entered, separately, into principal components analysis (PCA). Results from PCA, internal consistency and correlation assessments were used to aid the item removal process. Dimension characteristics (e.g., score distributions, internal consistency, scaling success rates, intercorrelations, construct validity) were evaluated. Results: PCA of subjects' (80% male; 45% white; 62% gay/bisexual) responses revealed nine previously identified HAT-QoL dimensions. The measure was shortened by removing 12 items. Two HAT-QoL dimensions (HIV mastery and provider trust) had ceiling effects. All internal consistency coefficients were > 0.80, except those for sexual function (0.57) and medication concerns (0.57). HAT-QoL scaling success rates were > 90% for 7 of 9 dimensions, and a majority of dimensions showed minimal to low intercorrelations. Validity assessments indicated consistent, expected relationships (p less than or equal to 0.05) for all dimensions except the medication concerns and provider trust dimensions. PCA indicated five MOS-HIV factors. Six of the 11 previously defined MOS-HIV dimensions - physical, role, social, and cognitive function, pain, and health transition - had substantial ceiling effects. MOS-HIV scaling success rates were > 90% for only 2 out of 8 evaluable dimensions; three dimensions had very low (40-73%) scaling success rates. Most MOS-HIV dimensions were moderately-to-highly intercorrelated. Validity assessments indicated consistent, expected relationships for all MOS-HIV dimensions. Conclusions: Six dimensions of the shortened HAT-QoL instrument (overall function, disclosure worries, health worries, financial worries, HIV mastery, and life satisfaction) exhibited good psychometric properties, including limited ceiling effects, low dimension intercorrelations, high internal consistency, and evidence for construct validity. All multi-item MOS-HIV dimensions had high internal consistency and all dimensions revealed consistent evidence for construct validity.