Events & Calendars "Statistical Methods for Improving HIV Treatment Monitoring in Kenya" - Statistics and Data Science Colloquia Featuring Joe Hogan of Brown University

Joe Hogan of Brown University will discuss "Statistical Methods for Improving HIV Treatment Monitoring in Kenya."

In places like the US and western Europe, individuals with HIV who are taking antiretroviral therapy are monitored regularly. This typically involves taking measurements of CD4 cell counts (a measure of immune system function) and HIV viral load (a measure of viral burden) every three to six months.

In low- and middle-income countries, however, this level of monitoring is not typically possible because of limitations in funding, technology and infrastructure. One approach that enables more frequent monitoring is the use of pooling for viral load testing. Individual blood samples are pooled, and a viral load assay is applied to the pooled sample. If viral load exceeds a certain threshold, some or all of the individual samples are tested separately.

We show how statistical methods are used to develop algorithms that lead to the fewest possible tests, thereby saving assays and enabling more testing. Our approach makes use of ideas from machine learning and decision theory.