A New Method for Estimating the Number of Undiagnosed HIV Infected Based on HIV Testing History, with an Application to Men Who Have Sex with Men in Seattle/King County, WA.

Abstract

We develop a new approach for estimating the undiagnosed fraction of HIV cases, the first step in the HIV Care Cascade. The goal is to address a critical blindspot in HIV prevention and treatment planning, with an approach that simplifies data requirements and can be implemented with open-source software. The primary data required is HIV testing history information on newly diagnosed cases. Two methods are presented and compared. The first is a general methodology based on simplified back-calculation that can be used to assess changes in the undiagnosed fraction over time. The second makes an assumption of constant incidence, allowing the estimate to be expressed as a simple closed formula calculation. We demonstrate the methods with an application to HIV diagnoses among men who have sex with men (MSM) from Seattle/King County. The estimates suggest that 6% of HIV-infected MSM in King County are undiagnosed, about one-third of the comparable national estimate. A sensitivity analysis on the key distributional assumption gives an upper bound of 11%. The undiagnosed fraction varies by race/ethnicity, with estimates of 4.9% among white, 8.6% of African American, and 9.3% of Hispanic HIV-infected MSM being undiagnosed.