| Accession Number | <strong>00002030-200406003-00004</strong>. |
|---|---|
| Author | Nemes, Maria IB; Carvalho, Heraclito B; Souza, Maria FM |
| Institution | From the Department of Preventive Medicine Medical School, University of Sao Paulo, Brazil. |
| Title | |
| Source | AIDS. 18 (Supplement 3):S15-S20, June 2004. |
| Abstract | Objectives: This study evaluated the adherence to antiretroviral therapy (ART) in 322 Brazilian outpatient services located in seven states providing care to 87 000 patients (72%) under ART. Methods: A previous study classified the 322 health services into four levels of quality of care. Sixty of them were randomly chosen on the basis of these levels. A cross-section of 1972 patients under ART visiting these services was interviewed using a structured questionnaire. Patients who reported taking more than 95% of the prescribed antiretroviral pills in the past 3 days were considered adherent. The chi-square test was first used to compare the prevalence of non-adherence among two or more categories of variables. A chi-square test for linear trend was used for ordinal variables. Three multivariate models were applied using health services predictors, treatment predictors, and personal characteristics predictors. The predictors were fitted into logistic regression models using backward elimination procedures. Results: The adherence prevalence was 75% (95% confidence interval 73.08-76.95). The level of quality of care was not associated with non-adherence. The models showed the following predictors of non-adherence: related to health services: services with 100 patients or less and missed appointments; related to treatment: more complex regimens and a high number of pills; related to personal characteristics: under 2 years of formal education. Conclusion: The adherence prevalence was similar to the rates currently obtained in developed countries. However, services with few patients need to be carefully monitored to maintain high rates. Care planning that prioritizes patients at risk also needs to be improved. (C) 2004 Lippincott Williams & Wilkins, Inc. |