Author(s): Pia Lpez JA, Corrales Rascn AE, Mungaray Padilla K, Valencia Vidrio MA, Pia Lpez JA, Corrales Rascn AE, Mungaray Padilla K, Valencia Vidrio MA
Abstract Share this page
Abstract OBJECTIVE: To test the psychometric properties of an instrument for measuring psychological variables and adherence to treatment behaviors in persons with HIV infection or AIDS. METHODS: A cross-sectional study was performed with 68 subjects who were seropositive for the human immunodeficiency virus (HIV) and who were being treated in a health facility in the city of Hermosillo (Sonora), Mexico, at the time the study was carried out. Five consecutive statistical testing procedures were applied for data analysis: the Student t test for comparison of means, for assessing the comprehensibility of the questions; an exploratory factorial analysis with varimax rotation for determining construct validity; Cronbach's alpha coefficient for determining reliability; Pearson's correlation coefficient (r) for assessing concurrent validity; and multiple linear regression for determining predictive value. RESULTS: In 25 out of a total of 30 questions, the Student t test for comparison of means gave statistical significance values of P < or = 0.05, P < or = 0.005, P < or = 0.01, and P < or = 0.001. A four-factor structure was found with the factorial analysis with varimax rotation: (1) current adherence behavior and treatment; (2) underlying reasons for adherence or lack of adherence; (3) psychological well-being and social competencies; (4) past adherence behavior with values above 1 that, taken as a whole, explain 65.50\% of the variance. An overall alpha coefficient of 0.852 was obtained with the test for reliability. The test for concurrent validity showed correlations between factors 1 and 2 (r = 0.615; P < or = 0.01), between factor 2 and factors 3 and 4 (r = -0.178; P < 0.05 and r = 0.265; P < 0,05, respectively), and between factor 3 and duration of infection in months (r = 0.265; P < 0.05). Finally, multiple linear regression showed that psychological factors 2 and 4, which together explain 48.6\% of the variance, predicted current adherence behavior (adjusted determination coefficient [R2] = 0.486). CONCLUSION: According to the results obtained, the instrument has optimal reliability as well as construct validity, concurrent validity, and predictive validity.
This article was published in Rev Panam Salud Publica
and referenced in Journal of AIDS & Clinical Research