|
National HIV Prevention Conference
|
[TITLE:] MODELING THE HIV EPIDEMIC
Natl HIV Prev Conf. 2005 Jun 12-15 (abstract no. T1-B1102)
Rhodes, P; Glynn, K
Centers for Disease Control and Prevention, Atlanta, GA
SUMMARY: The talk will look at a statistical model, an expanded form of backcalculation, that estimates the size and composition of the HIV epidemic in the United States and mathematical models that describe the impact that interventions might have on HIV incidence. A crucial and unifying element for both models is the number and composition of undiagnosed HIV+ individuals. The two crucial concepts in both models are a) the inherent difficulty of precise estimates and b) that approximations and qualitative arguments can provide much useful information. The backcalculation method used here uses all known HIV+ cases not just those cases already diagnosed with AIDS. The model avoids the need to incorporate treatment effects by using only the time at which the individual was first known to be HIV+. A similar method has been used to estimate the size of the epidemic Louisiana through 1996 (Posner 2004). I will discuss our efforts to extend these types of models to the entire U.S. and some of the difficulties we have encountered. New ideas about intervention have been developed by considering the size of the undiagnosed population and its contribution to the epidemic. Many feel that undiagnosed cases while constituting only 20-30% of total infections may cause 50% or more of new infections. Thus, reducing the pool of undiagnosed cases is itself now considered an intervention to be evaluated. How much of a reduction in incidence can be achieved in this manner? Can this intervention reduce incidence by 50%? I will examine the factors that make it difficult to calculate a definite answer but show how qualitative analyses can provide reasonable limits about what can be expected.
Download PDF of this abstract.
050612
T1-B1102
Copyright notice: The National HIV Prevention Conference is collaborative effort by the Centers for Disease Control and Prevention, a U.S. Government agency, and other governmental and non-government organizations. All abstracts published by the conference organizers are in the public domain and can be used without permission. Proper citation, however, is required.