HIV epidemics among injecting drug users (IDUs) are often characterized by rapid spread and high disease burden, which make this group an important priority for HIV prevention and care. Transmission of HIV infection through injecting risk often follows a common and predictable pattern, where a few undetected infections stimulate a rapid rise before reaching some steady state prevalence. This pattern remains quite consistent across most epidemics among injecting drug users, but variation in injecting behaviors and the underlying risk network may influence the speed and spread of HIV infection. Understanding how these important factors interact to alter HIV transmission dynamics can bring about more effective intervention design and more efficient delivery through targeting those components of the network that are key drivers of transmission. One opportunity to elucidate these interactions has arisen in the heterogeneity of emerging HIV epidemics among injecting drug users in Cebu, the Philippines. While HIV prevalence among IDUs has grown rapidly and stabilized at almost 50% in Cebu City, it remains quite low (3.5%) in neighboring Mandaue City. Given their geographic proximity, we would expect HIV infection to quickly saturate the IDU population across both cities, but the data suggest otherwise. To understand the circumstances that precipitated this unusual contrast in HIV prevalence, we propose to study the injecting risk behaviors and IDU network structures in these two cities. We hypothesize that the networks of IDUs in Cebu City and Mandaue City are linked, but HIV transmission dynamics vary because of differences in network structure and injecting behaviors in each city. To assess these effects, we will analyze HIV surveillance data, collected using respondent-driven sampling, to compare the network structures across the two cities (Aim 1). Phylogenetic analyses using HIV nucleotide sequences from cases detected in the same surveillance sample will be analyzed to identify transmission clusters (Aim 2). The clusters identified will be overlaid on the risk network and with behavioral data to describe behavioral and network characteristics of the infections detected in each city. Finally, we will synthesize these data to build a mathematical model and replicate the HIV transmission dynamics observed in Cebu and Mandaue (Aim 3). Using the model, we will conduct simulations to quantify the relative contributions of individual risk behaviors (such as needle sharing prevalence and injecting frequency) and network measures (including size, density, clustering, and distance) on the spread of HIV infection. Understanding how network structures impact HIV transmission, and identifying important individuals, whose network position could initiate a chain of downstream HIV infections, can provide insight on the types of behaviors and individuals we can target to slow or stop the spread of HIV to the broader injecting drug user population.
With a better understanding of which network or local settings slow or facilitate the spread of HIV, we can design and tailor more effective HIV prevention strategies for injecting drug users. This study will utilize existing surveillance data collected using network-based sampling methods to elucidate the extent to which network structures and individual behaviors affect HIV transmission dynamics among injecting drug users in Cebu, the Philippines.