This proposal is in response to Program Announcement (PA) Number: PA-10-212 issued by the National Insti- tute of Allergy and Infectious Diseases (NIAID), National Institutes of Health (NIH), which solicits grant applications from institutions/organizations that propose to develop improved HIV incidence assays with increased specificity and reliability for distinguishing incident from chronic HIV infections. The stated motivation is to improve the ability to estimate the incidence of HIV infection in order to better control the epidemic. Our proposed method focuses on addressing three important shortcomings associated with current measuring methods: low specificity, low sensitivity and sampling bias. To address these shortcomings, we first introduce a novel entropy measure associated with the viral genome, which will distinguish between recent and long term infections among seropos- itive individuals who test negative on the BED HIV-1 Capture EIA assay, or the best available screening assay. This will increase the overall specificity of the assay. Second, we propose a pooling strategy to lower the cost and increase the accuracy of directly identifying individuals in the critical acute phase of the incidence curve, thus increasing the efficiency and sensitivity of the overall procedure, especially during the period when the anti- body assays have extremely low (or even zero) sensitivity for detecting infection. And third, we propose a way to take advantage of the anonymity afforded by pooled testing to overcome a high non-consent-to-testing rate and its concomitant bias-a non-consent rate that in some studies is sufficiently high to make those survey results suspect. Combining these three proposed advances provides a principled testing method for a cross-sectional estimator of incidence that is cost-effective, timely and reliable. Moreover, in addition to the direct response to the PA we propose a method for estimating the instantaneous incidence curve-rather than a single number-that provides a superior perspective of the progressing epidemic. Because our proposed methods all employ existing technologies in a novel way and do not require any new technological breakthroughs, they offer a high probability of success with only a minimal risk of failure.
Our project takes aim at three serious shortcomings of the existing cross-sectional incidence estimators [1]: 1. the lack of long-term specificity associated with properly classifying seropositive individuals who test negative on the BED HIV-1 Capture EIA assay; 2. the lack of sensitivity and the high costs associated with finding patients in the acute phase of infection; and, 3. the potentially sizable sampling bias introduced by individuals who withhold consent for HIV testing. By basing our cross-sectional method on both a measure of entropy associated with the viral genome so as to age the infection and on pooling methods that target the additional shortcomings, our resultant estimator is not only cost-effective and timely but also principled and reliable, uniquely correcting on multiple levels a number of current deficiencies. Our methods attempt to be cost-effective particularly to be implementable in economically constrained countries.
Showing the most recent 10 out of 12 publications
