The Kirby Institute case study

The dynamics of STI epidemiology are complex, multi-factorial, and are often best studied by computer simulations calibrated to real-world data. We have built an agent-based model to simulate STI epidemiology in Sydney’s GBM population. Such models are well-suited to incorporate the large variation in relevant parameters at the individual level. Currently our model only considers the transmission of gonorrhoea and accounts for sexual behaviour, HIV status, PrEP use, and STI testing frequency. The model has been constructed, however, to facilitate the inclusion of other STIs (e.g. chlamydia, syphilis and Mycoplasma genitalium) in due course. The model is also anatomical site-specific, meaning that the infection status of individuals also includes the relevant anatomical sites of infection (rectum, pharynx, urethra). This is important because the various modes of transmission (anal sex, oral sex, etc.) have different transmission probabilities and their involvement in sexual contacts occur with differing frequencies. Furthermore, the use of condoms is only applicable to sexual contacts in which the penis is involved. The current gonorrhoea model is calibrated to trends in gonorrhoea notifications both prior to and following the introduction of PrEP, and makes use of data on testing rates, gonorrhoea diagnoses and sexual behaviour in the PrEP era from the EPIC-NSW PrEP trial⁵ that was funded by the NSW government. The model will be used to predict the trajectory of gonorrhoea incidence (and later, other STIs) with the expanding uptake of PrEP, and will evaluate a range of scenarios around testing and changing sexual practices.

A working group has been formed including stakeholders from GBM advocacy groups such as the Australian Federation of AIDS Organisations, as well as representatives from the NSW Ministry of Health and experts from within the Kirby Institute. Our model has the capacity to examine counterfactual scenarios and test hypotheses for which clinical trials are not suitable. Hence questions of relevance to government policy, such as the frequency with which GBM should be tested for STIs, may be addressed and guidelines proposed. It has also been agreed that capability development in modelling, including the interpretation and use of models, will be provided to staff at NSW Health through annual workshops. 

Translation into policy and practice will be through the submission of reports to relevant stakeholders, the publication of academic papers, as well as presentations at seminars and conferences. To further enable translation into action, our model is open-source with code readily available for download from a public repository (GitHub⁶), in accordance with emerging publication standards. The work has already leveraged additional research funding from the SPHERE (Sydney Partnership for Health, Education, Research and Enterprise) seed grant scheme⁷ for a project in which modelling is being used to evaluate the utility of a range of surveillance and contact tracing strategies in mitigating the emergence and spread of extensively drug resistant gonorrhoea in NSW.