Feature story

Modelling the extreme—COVID-19 and AIDS-related deaths

25 May 2020

Kimberly Marsh, a senior adviser on modelling and epidemiology, has worked for UNAIDS for six years. She supports countries in estimating the impact of the HIV epidemic globally and regionally.


Can you tell me more about the latest modelling report that you are a co-author of, which examines the potential for HIV service disruption in times of COVID-19 in sub-Saharan Africa?

This work looks at potential disruptions in sub-Saharan Africa owing to the COVID-19 pandemic on HIV services that might have an impact on HIV incidence—the number of new HIV infections—and on the number of AIDS-related deaths in excess to those we might have observed if we hadn’t had the COVID-19 pandemic.

We are particularly interested in those question because we know that more than two thirds of all people living with HIV worldwide are living in sub-Saharan Africa. That’s 25.7 million people living with HIV, 1.1 million new HIV infections and around 470 000 deaths from AIDS-related causes in 2018. Among all people living with HIV in the region, 64% of people are on life-saving antiretroviral therapy, which also prevents further new HIV infections.

It is really important that we’re able to ensure they will have access to services. In the models, we looked at service disruptions—a complete disruption of any HIV-related services over a three-month and a six-month period of time. And we looked at the impact after one year and five years. Now remember, these are just scenarios, and extreme ones. We don’t expect this to happen, but it helps us to answer two questions: what HIV-related services are most important to prevent additional deaths and new HIV infections and what might happen if we don’t mitigate or address those disruptions.


From this huge amount of work, what are the two key takeaways?

The modelling work predicted that with a six-month disruption in HIV treatment there could be an excess of 500 000 deaths in sub-Saharan Africa. So, when you look at UNAIDS estimates of AIDS-related deaths over time, that would take us back to about 2008, when we had nearly a million deaths.

There is no doubt about it, HIV treatment is critical. Ensuring that HIV treatment is available to people who need it during the three- to six-month periods is the most important thing that countries can do to prevent excess deaths and HIV incidence. All countries should work to ensure that supply chains are providing them with enough medicines to distribute and that people have sufficient medicines so that they can take them over the coming months.

The second thing to say is that these are projections and that there is still time to ensure that people get the HIV treatment services they need.

Let’s prevent what this model potentially predicts and let’s get HIV medicines to the people who are living with HIV.


What about HIV prevention? Does condom availability have an impact?

The models showed that when you look at prevention services, condom availability impacted the results. I think it is important to say that this is a treatment lesson primarily, but things like access to condoms is really important. We saw around a 20–30% relative increase in HIV incidence over one year if condoms were not available for six months. This is definitely something that we should be focusing on.


Can you tell us a little bit more about the impact on mother-to-child transmission of HIV in these scenarios?

In the scenarios, we looked at the potential for HIV testing services to be disrupted as well as for women to not get medicine to prevent transmission of HIV to their children. And what the various models found was that by removing those medicines—which have had an extremely important impact in terms of reducing new child HIV infections over the past five to 10 years—you could see rises in new child HIV infections in selected countries anywhere up to 162%. It really is critical to maintain prevention of mother-to-child transmission of HIV services.


You have said this was an extreme scenario, not a prophecy, but yet you still believe in modelling?

Models are very important for exploring questions that countries routinely pose to UNAIDS and the World Health Organization in terms of thinking of strategic approaches to responding to HIV in their countries. Models aren’t perfect, but they have a lot to tell us and I think in this instance it really highlights some of the strategies that will be important over the coming months as COVID-19 impacts or potentially impacts sub-Saharan Africa.