To mark the launch of the PLoS Medicine collection “Investigating the Impact of Treatment on New HIV Infections”, I caught up with Dr. Timothy Hallett, who is Reader in Global Health at Imperial College London, and Principle Investigator of the HIV Modelling Consortium which sponsored this collection. He described some of the issues the ten articles in the collection address, and gave me insight into the role of mathematical modelling in HIV research. PLoS Medicine and Dr. Hallett will be at the AIDS 2012 conference in Washington D.C.
What role can mathematical modelling play in evidence-based decision making about HIV?
Many of these really big questions in public health can only be answered by drawing together data from lots of different places – including studies in behaviour, economics, virology, and operations research. Mathematical models provide a way for us to bring all those data together, and that can help us to look at the questions in a careful way. In particular they can help show us what we do and do not know and can tell us if our ideas about how to intervene might work in the way we would hope. I think this all helps inform decision making at lots of different levels.
How does this collection of articles fit with the HIV Modelling Consortium’s mission?
The HIV Modelling Consortium aims to strengthen the support that modelling can give to decision making in the response to the worldwide HIV epidemics. In preparing for this collection, we brought a wide range of modellers together with statisticians, economists, and public health policy leaders. We collectively reviewed a lot of work together and this led to some good conclusions and raised plenty of important new questions. For example, by comparing the model results, we identified what the models agreed upon – which gives us more confidence in the results – and what they could not – which begins to tell us where we need to do more work. Through examining the data about treatment in industrialised countries, we raise the questions about whether this is consistent with current models. And, by bringing out a conversation between modellers and those actually facing the questions who have to rely on modelling research, we think we can begin to get an understanding of how we can improve communication and collaboration between these groups.
Why is the discovery that antiretroviral treatment (ART) can be used for the prevention of HIV transmission so exciting?
I think it comes down to two things. First, the magnitude of the effect on transmission – it was 96% in a trial and about that in observational studies too – and we have not seen such a large effect size in HIV prevention before. Second, is the fact that the effect comes from something that, in some sense, we are already doing – treating HIV-infected patients. But, while it is exciting, I think what’s also coming out of this work is a sobering message about why that 96% reduction in individual-level trials is not going to immediately translate into a 96% reduction in HIV incidence in populations.
What will be the biggest obstacle to using ART as a preventative measure?
The interesting thing is that, without really realising it, ART programs may already have been having an impact on reducing HIV incidence, even though that was not their aim. But, the impact will certainly not have been as big as we would want it to be, and we will face plenty of obstacles as we try to expand and strengthen programs to maximise that prevention impact. I would say that there might have initially been a feeling that this very biomedical model of intervening (i.e. distributing pills) would be a bit easier to do than trying to shift sexual behaviour – for example, encouraging more people to use condoms. But, for the kind of treatment programs we would need to reduce incidence substantially, we would still have the challenge of asking people to change their behaviour – to test for HIV regularly, to take pills every day, to return to the clinic regularly for check-ups for the rest of their lives. I think that maintaining that kind of behaviour over the long-term could be a challenge.
What will be the next steps in this area of research?
It is a very active area of research. I think modellers will look for forms of intervention that maximise the preventive impact of treatment and try to evaluate the impact that treatment could already have had. There are lots of projects in the field now that are describing and trying to improve on different parts of ART programs. Very interestingly, there are also trials that are going to directly test the hypothesis that very ambitious treatment programs can reduce HIV incidence. In about 4 years’ time, cumulatively all of those results could put us at a real turning point in the history of the epidemic.
Dr. Hallett works on the development and application of mathematical models for interpreting HIV data and anticipating the impact of interventions. He is on the PLoS Medicine editorial board.
The HIV Modelling Consortium aims to strengthen the support that mathematical modelling and related quantitative disciplines can provide to global decision-making in HIV. The “Investigating the Impact of Treatment on New HIV Infections” collection comprises nine reviews and one research article and provides insights into the factors which will support evidence-based decision-making in HIV prevention, with a focus on the use of antiretroviral treatment to prevent HIV transmission. Comment on the articles, http://www.ploscollections.org/TasP2012, and join us on Twitter #plosAIDS2012.
The collection is produced with support from the HIV Modelling Consortium, which was funded by a grant from the Bill & Melinda Gates Foundation.