Аңдатпа
Aim: This paper examines the predictors of all-cause mortality among hospitalized HIV-positive patients in Kazakhstan. Material and methods: The study uses baseline data from patient hospital discharge records derived from the Unified Electronic Healthcare System of Kazakhstan (UNEHS) between 2014 and 2019. Artificial intelligence technology was utilized to extract data from the discharge records. Patients were included based on their first hospitalization, and they were subsequently monitored until their discharge or occurrence of death. Results: The study revealed that females had a 2.06-fold higher risk of all-cause mortality compared to males. After adjustments in the Cox proportional hazard model, age, intravenous drug use (IDU), and anemia were observed as independent predictors of mortality within the patient cohort. Conclusions: Findings of this study emphasize the need to enhance efforts to prevent late HIV diagnosis by improving access to testing and treatment for those affected, and to strengthen potential for developing risk factors reduction strategies.