Determinants of Non-Adherence to Antiretroviral Therapy and Modeling Progression of Adherence Level: At Yirgalem Regional Hospital, Ethiopia; January, 2012

Non-adherence is a global problem and has been seen in all diseases. According to WHO, adherence to long-term therapies in chronic illnesses averages 50% in developed countries? Adherence to HAART is challenging, because patients need almost perfect adherence of at least 95 percent to keep viral load at undetectable levels as long as possible and maintain the functionality of the immune system. Poor adherence to antiretroviral therapy increases the risk of incomplete viral suppression, disease progression and death. The purpose of this study was to assess the level and factors associated with non-adherence to antiretroviral therapy. A random sample of 368 patients were taken from the ART clinic at Yirgalem Regional Hospital from September 2006 to July 2011.Adherence levels were measured using methods pill count, pill identification test, and self-report. Descriptive statistics and both Bivariate and multivariate Logistic Regression were used to analyzed the data. Homogeneous Semi-Markov Stochastic Model was used to reveal the conditional probability of staying in non-adherence. The Multivariate Logistic Regression Method showed that the gender (sex), family size, social support available to the patients, ART knowledge of the patients, time on ART in month, and baseline CD4 counts/mm3 were significant factors associated with non-adherence to ART medications. The conditional probability of staying in non-adherence, until a given time in month was higher than staying in adherence. Thus, it is suggested that health-care providers need to be watchful about adherence levels of patients. Specifically males, who have been on ART for several months, living in large family members, lack of social support, having high baseline CD4 counts/mm3, and having poor knowledge of ART.