Skip to content

When you choose to publish with PLOS, your research makes an impact. Make your work accessible to all, without restrictions, and accelerate scientific discovery with options like preprints and published peer review that make your work more Open.

PLOS BLOGS Speaking of Medicine and Health

More people, more time, better data – what we need to ‘treat-all’ with HIV

On World AIDS Day 2015 and as the 18th International Conferences on AIDS and STIs in Africa (ICASA) is in full swing, Helen Bygrave discusses the implications of the recently announced WHO ‘treat-all’ policy.

On the back of the headline studies Temprano, START, and new data from HPTN 052, WHO recently announced the ‘treat-all’ policy: everyone diagnosed with HIV should start antiretroviral therapy (ART), regardless of their immunological status. No more hanging around waiting for people to get sick and transmit the virus, we now crack right on and start the right medication for the right disease. Great news, we all say, as pre-ART follow-up was an undeniable disaster and the new policy may further simplify how we get treatment to those in need, especially in low coverage or unstable settings. But what are the implications of a policy that increases the number of people now eligible for ART from 28 million to 37 million? Can we get an idea of some of the key implementation issues by looking at previous experience with a treat-all approach.

Malawi was the first country to move at a national level to PMTCT B+ (prevention of mother-to-child transmission), which meant starting all pregnant and breastfeeding women with HIV on ART regardless of their immunological status. And, after WHO recommended this approach in 2013, many other countries followed suit. At IAS 2015, the Malawi PMTCT B+ programme reported a meagre 62.6% retention rate after 30 months on ART.[1] Similarly, at the upcoming ICASA conference to be held in Zimbabwe the week of World AIDS day, Médecins Sans Frontières (MSF) will present data from PMTCT programmes in Malawi, Zimbabwe, and Swaziland showing that after 24 months between 20-30% of women are recorded as no longer being in care.[2][3][4] But are these women actually leaving care or are our tracking systems failing under the pressure of increased numbers of patients and care delivery sites? To answer this question, a tracing exercise was carried out in the Zimbabwe site. More than half the women recorded as lost to follow-up were traced, and 33% of these reported that they were receiving ART elsewhere. So why as I write this, are vast amounts of money being invested in monitoring and evaluation systems that simply do not allow for the fact that people move? Technology such as smart cards or retinal scanning, in theory, should be able to help us achieve the elusive unique patient identifier needed to link our records in a shared, confidential online database.

But as we face the need to massively scale up ART provision, should we really continue to move towards a universal accounting system for ART delivery, or should we take a step back? We could instead accept that pharmacy data could allow us to order the drugs our cohorts need and laboratory data tell us how many patients are getting viral load tests and achieving suppression. Would this amount of monitoring be enough? Maybe in addition, once in a while, we could go out and perform a population based survey to get a different look at what is going on. Surveys performed in MSF sites in South Africa, Malawi, and Kenya all showed better treatment outcomes than the monitoring and evaluation systems being used within the clinics themselves.

What can we learn for a treat-all approach from those women who had really stopped their treatment in the PMTCT studies? Firstly, that more resources for patient support and counselling are urgently needed. Unfortunately, in the Zimbabwe tracing study, still by far the most common reason cited for stopping treatment was the barrier of being able to disclose their HIV status to partners and family. In many of the HIV treatment sites I visit with MSF, lack of time spent by health-care workers with women newly diagnosed with HIV is still a challenge as is the too-common absence of any counselling support. Interestingly, several of the women we interviewed in Zimbabwe related that “I’ll wait until my partner can also start”. So maybe the treat-all policy will help with disclosure and retention in care; the alternative situation where the female partner is eligible for ART while the man waits for his immune status to fail is maybe not the best recipe for marital harmony.

Whatever efficiencies can be built into the system in order to treat all of the 37 million people who now should be taking ART, I have no doubt that significant investment will be needed in both providing support so that those in need understand why ART should be started here, now, and forever as well as in data systems that allow us to have any clue as to how many of the “all” we are treating.


Helen Bygrave


Helen Bygrave is a UK-trained General Practitioner. For the past 10 years she has worked for Médecins Sans Frontières (MSF) supporting HIV/TB programmes across Africa and Asia and currently works as one of MSF’s HIV/TB advisors in the Southern Africa Medical Unit.



[1]. Haas AD, Tenthani L, Msukwa MT, et al. Post prevention of mother-to-child-transmission: 30-months outcomes in the Malawian “Option B+ programme”. IAS: Vancouver, 2015. Available at: (accessed 25/11/2015).

[2]. Nundwe S, Jalon O, Metcalf C, et al. Outcomes of the PMTCT Option B+ programme in Thyolo, Malawi. ICASA: Zimbabwe, 2015.

[3]. Bonyo T, Munyaradzi D, Chenjirai G, et al.Retention and virological outcomes from a PMTCT programme in rural Zimbabwe 2013-2015. ICASA: Zimbabwe, 2015. THUAC1206.

[4]. Etoori D, Kerschberger B, Ndlangamandla M, et al. Mother and early infant outcomes of a PMTCT B+ implementation study in southern Swaziland. ICASA: Zimbabwe, 2015.


Cover image credit: SibleyHunter, Flickr

Leave a Reply

Your email address will not be published. Required fields are marked *

Add your ORCID here. (e.g. 0000-0002-7299-680X)

Back to top