Introduction Late presentation for HIV care (LP) delays the achievements of the 90-90-90 UNAIDS target, a program in which Ethiopia has subscribed for. However, the prevalence, trend, outcomes and risk factors of LP among children and adults were not assessed very well in the nation.
Methods 12 years retrospective cohort study was conducted using data extracted from an antiretroviral therapy (ART) clinic in Southwest Ethiopia. LP for children and adults was measured using CD4 lymphocyte counts and WHO clinical stages. We described the percentage of LP by mortality, discontinuation from ART and immunological failure to show outcomes of LP. The analysis of descriptive and inferential statistics (logistic regression) was undertaken. Missing data were handled using multiple imputations assuming missing at random (MAR) pattern.
Results Of the 8172 patients enrolled for HIV care between June 2003 and March 2015, 5299 (64.8%) patients were on ART: 4900 (92.5%) were adults and 399 (7.5%) were children. The prevalence of LP was 57% in children and 66.7% in adults with an overall prevalence of 65.5%, and the 11 years analysis of LP showed upwards trends. 74% of died children, 50% of discontinued children, 57% of transferred out children and 45% of children with immunological failure were delayed presenters for HIV care. Similarly, 64.7% of died adults, 65.3% of discontinued adults, 68.1% of transferred out adults and 78.7% of adults who had immunological failure presented late for the care. Factors for LP among adults were: being female, being married, having IF, having Tb/HIV co-infection and having no history of HIV testing. No statistically significant predictor was found for LP among children.
Conclusions The prevalence LP was significant and majority HIV infected children and adults who presented late for HIV care had discontinued, transferred out and immunological failure. To address this, strategies such as unmanned aerial systems for transporting laboratory specimens, programs such as home and community-based HIV testing, ‘opt out’ and self-testing are compulsory.