In Sub-Saharan African country, particularly in Ethiopia, there is a growing concern of the management of diagnosis and treatment of to diagnose HIV-AIDS patients. These patients later on are presented with advance disease and loss of follow up in not uncommon. There are several factors demotivating these patients to continue their treatment but later on, they will present with advanced disease because they have loss to follow up. The current retrospective cohort study Assessing assesses the predictors of loss to follow up is crucial. OBJECTIVE: To assess predictors of loss to follow up among adults who were enrolled on Anti-Retroviral Therapy from January 1, 2013 to December 30, 2017 in Otona teaching and referral Hospital Wolayita Soddo, Southern Ethiopia. Methods: Institutional-based retrospective cohort study was conducted from January 1, 2013 to December 30, 2017. Simple random sampling technique was used to select 328 medical cards. Data was entered and cleaned by Epi Info version 3.5.3 and analyzed by STATA version 11. Bivariate and Multivariable Cox proportional hazards model was used and Predictors with p-value of <0.05 has been consider as statistically significant. Result: - A total of 328 adults’ medical records were reviewed From multivariable Cox proportional hazards finding, distance from hospital>5 KM (AHR=3.84, 95% CI =1.99-7.38, P=0.0001), no registered phone number on their medical card (AHR=2.47, 95% CI =1.32-4.09, P=0.003), have not taken INH prophylaxis (AHR=2.23, 95% CI=1.3-4.09 P=0.003), alcohol intake (AHR =1.77, 95% CI =1.01-3.12, P-value =0.046) and those who have no care giver (AHR=2.11, 95% CI=1.23-3.6, P-value=0.006) were identified as predictors of loss to follow up. Conclusion This study showed that distance from hospital, registered phone number on patient’s card, alcohol intake, INH prophylaxis and having family support are independent predictors of loss to follow up from ART.

Keywords: predictors, loss to follow up, adults, ART, Otona