Optimizing HIV Treatment Monitoring under Resource Constraints

点击翻译按钮获取中文摘要

 DESCRIPTION (provided by applicant): With escalating global access to HIV antiretroviral (ARV) therapy, treatment failure is inevitable and must be anticipated. Correct and early diagnosis of treatment failure is essential for cost savings, durable response to therapy and prevention of morbidity and mortality. Most HIV treatment programs in the developing world either do not have access to viral load (VL) testing, the gold-standard treatment monitoring modality, or can apply it only on a limited basis. Other monitoring technologies such as drug resistance testing are even less common for financial and infrastructure constraints. In this proposal we will develop, evaluate and implement methods to optimize monitoring of ARV therapy in resource limited settings (RLS) that have diverse VL availability. Though several recent studies have proposed and evaluated lower-cost markers as VL surrogates and strategies for selective VL use, they generally are not based on a formal, decision theoretic framework that allows discovery of strategies with optimality properties that can be expressed in terms of misclassification rate, cost, and other clinically relevant parameters. We propose to develop the statistical framework, theory and methods required to discover optimal diagnostic algorithms for monitoring treatment failure with limited or no VL availability; to use cohort data from both the US and Kenya to derive, calibrate and cross-validate the algorithms; to use extant plasma samples from patients in a PEPFAR-funded HIV care program to design and cross-validate a new diagnostic algorithm that includes implementation of pooled assays; and to develop usable software that will enable programs to design their own protocols based on the characteristics of their patient population and local capacity for viral load testing.