Test-retest evaluation of HAT-QoL and SF-36 in an HIV-seropositive sample

Abstract

We sought to correct previously reported psychometric and content problems of the HIV/AIDS-Targeted Quality of Life Instrument (HAT-QoL) and to assess test-retest reliability of this revised HAT-QoL as well as the MOS 36-Item Short-Form Health Survey (SF-36) when used in an HIV-seropositive outpatient sample. A total of 153 participants randomly sampled from an HIV specialty and general medical clinics completed a test booklet that contained HAT-QoL and the SF-36; a random subsample of 60 participants were asked to return in approximately two weeks to complete the instruments again. Using the baseline sample, internal consistency coefficients of all final HAT-QoL dimensions were > or =0.80. The retest subsample, for which there were no significantly different characteristics than the baseline sample, completed the instruments again an average of 14.7 days after baseline completion. For HAT-QoL, only one (Provider Trust) of the nine dimensions revealed a significant test-retest difference (+5.9, p=0.05). All HAT-QoL intraclass correlation coefficients (ICCs) were > or =0.64 (seven values were > or =0.73). All effect sizes were small. For SF-36, two dimensions (Physical Functioning and Social Functioning) revealed significant test-retest differences, and one dimension (Mental Health) revealed a difference approaching significance (p=0.06). The SF-36 ICCs were similar to those for HAT-QoL, with the exception of three dimensions--Vitality, Role-Emotional and Mental Health (all were<0.56). Though most effect sizes were small, Social Functioning and Mental Health estimates were approaching more moderate effect sizes. HAT-QoL reveals considerable short-term stability and has psychometric features in the seropositive population comparable to instruments such as SF-36, while obtaining unique information.