PrEP will need high adherence, high effectiveness and high coverage in specific populations to be affordable in the US, New York study finds

This article is more than 10 years old. Click here for more recent articles on this topic

A study based on New York City that modelled pre-exposure prophylaxis (PrEP) uptake there has found that in order to be affordable, PrEP would need to be tightly targeted at gay men at higher risk of HIV infection. Within this target population, it would need high levels of usage.

A reduction in the price of Truvada – the tenofovir/emtricitabine combination pill which is currently the only one used for PrEP – would also help. The model ran through a number of scenarios and found that while PrEP could be potentially cost-effective at Truvada’s present price, it would only become cost-saving if its current US price was halved. Even then, it would require near-universal uptake in higher-risk gay men.

Like all mathematical models, this one has to start with assumptions about PrEP’s likely uptake and effectiveness. Drug-level monitoring studies show that PrEP is highly effective (over 96%) in people who are even moderately adherent. This model uses a base-case scenario of PrEP having 44% effectiveness, which was what was seen in the original iPrEx study in gay men (effectiveness was 50% in the open-label extension of this study).



Cost-effectiveness analyses compare the financial cost of providing health interventions with their health benefit in order to assess whether interventions provide value for money. As well as the cost of providing medical care now, analyses may take into account savings on future health spending (because a person’s health has improved) and the economic contribution a healthy person could make to society.


How well something works (in a research study). See also ‘effectiveness’.

linkage to care

Refers to an individual’s entry into specialist HIV care after being diagnosed with HIV. 

mathematical models

A range of complex mathematical techniques which aim to simulate a sequence of likely future events, in order to estimate the impact of a health intervention or the spread of an infection.


A clinical trial where both the researcher and participants know who is taking the experimental treatment. 

An updated analysis of adherence by site, however, presented at the recent third IAPAC summit on treatment as prevention, shows that trial participants in the US sites managed adherence of approximately 80% at any one time or 67% over time, suggesting that effectiveness rates in US populations could be more like the 75-80% seen in the Partners PrEP study in heterosexual couples. If effectiveness was 75%, then PrEP would become cost-saving with somewhat lower coverage and would be potentially cost-effective, even at Truvada’s current price, in all gay men.

There is no scenario in which PrEP was cost-effective if offered to high-risk heterosexuals, and almost none if it were offered to people who inject drugs. In the former case, this is because HIV incidence in the US is not sufficient even among high-risk heterosexuals to justify the cost of PrEP – you would need to give a lot of people PrEP to avert one infection – and in the case of people who inject drugs, while they are at very high risk of HIV, there are not enough of them in New York City to make a PrEP programme cost-effective in terms of the number of infections it would prevent.

As the researchers say, this does not preclude offering PrEP to members of these groups who are at exceptionally high risk of acquiring HIV, for instance heterosexuals in a sero-different relationship with a partner who is off treatment or has adherence difficulties, or men who inject drugs who also have sex with men.

The model used

The model’s base-case scenario relies on figures already gathered from New York City surveillance data and from scientific study results. It assumes that, as described above, PrEP stops 44% of HIV infections that would otherwise have happened. It assumes that PrEP uptake among the various populations specified (see below) is 50%; it inputs an annual cost of $9672.00 for Truvada (this includes staff and monitoring costs); it assumes that 6% of men in New York have sex with men; and it assumes that a maximum of 44% of sexually active men who have sex with men and 21% of heterosexuals have “multiple, concurrent partnerships” – their definition of 'high risk'. It assumes that 1.4% of the population are people who inject drugs and approximately a third of these sometimes share equipment.

It assumes that 31% of all adults test for HIV annually and of those diagnosed HIV-positive, 61% are on ART and adherent to it (and therefore non-infectious); both of these figures are New York-specific and do not reflect the rest of the United States.

The model calculated the absolute cost of PrEP and its incremental cost (i.e. the cost per HIV infection averted) if it was offered to various different groups: to all adults: to all high-risk heterosexuals, to men who have sex with men (MSM) and people who inject drugs; to MSM only; to high-risk MSM only; to people who inject drugs only; to high-risk heterosexuals only; and to combinations of these groups.

The cost of PrEP

The sheer cost of PrEP offered to the general population would be vast, and completely uneconomical. If PrEP was offered to the entire HIV-negative population, it would (under the model inputs above) prevent 29% of all HIV infections at a cost of $52 billion a year. If given to just MSM, people who inject drugs and high-risk heterosexuals it would prevent 24% of HIV infections at a cost of $7.6 billion annually. Even if one then subtracts the $360,000 that is the estimated cost of a lifetime’s HIV treatment, this would still cost $10.64 million for each HIV infection averted.

If given solely to MSM, that would stop 19% of HIV infections overall and would cost $1.24 million per infection averted. But if targeted specifically at those MSM having “multiple concurrent partnerships” this would cost $740,000 per lifetime infection averted at an annual net cost of $467 million a year. This is still not cost effective but at least within touching distance of it.

High adherence needed

Things change if there are different input values for PrEP efficacy, adherence and drug costs. If PrEP is 75% effective rather than 44%, then you could prevent 22% of HIV infections for the same cost and this would put PrEP within the realm of cost-effectiveness.

Under 75% efficacy, if the drug price was halved and coverage in high-risk gay men was as high as 70%, then PrEP would actually start saving money – the PrEP programme would cost less than a lifetime’s antiretroviral therapy for the HIV infections that would otherwise have happened. But this does assume high levels of usage among MSM at high risk of HIV – and no usage amongst others who need it less.

Seventy-five per cent efficacy is probably achievable in the US. In a separate article, researchers from the original iPrEx randomised controlled study of PrEP update data on the adherence to PrEP seen in participants.

While adherence was overall 55%, it was the influence of the largest site, Lima in Peru, which had adherence of only 35%, that brought it down to this level. In the US, adherence was 72% in Boston and 90% in San Francisco: it was also above 70% in both sites in Brazil, the one site in Thailand and only just under this (68%) at the one site in South Africa (Cape Town).

This is adherence as measured by one single measure of drug level in the body at or around the eighth week in the trial. Over time, average adherence in iPrEx fell from 59% to 44%. But in San Francisco 67% of participants showed consistent adherence over time – even though in a PrEP trial in a limited group of people, one would expect some participants to decide they no longer needed PrEP.

A UK scenario

What about the UK? The same modelling has not been done here, but figures obtained by show that annual HIV incidence in gay men testing at sexually transmitted infection (STI) clinics – who are likely to have higher infection rates than other gay men – is 2.5% and in men diagnosed with a bacterial STI in the last year, 3.7%. If PrEP was taken by 30% of the men in these categories, then 133 gay men would need to take PrEP in order to prevent one HIV infection, or 90 who had had an STI. If coverage was 60%, then the figures would be 67 and 45 respectively to prevent one infection.

Whereas the US price for Truvada used in the model is about £6045 in UK currency, the list price for Truvada in the UK is £5022 per year (17% lower)  and most UK health authorities in fact negotiate prices some 20-30% lower than the list price.Using the UK figures for drug cost, this would lead to a PrEP cost of £38,000 per HIV infection averted for gay men diagnosed with bacterial STIs in the UK, even if the US lifetime treatment cost is used. This is not within the UK cost-effectiveness limit of £30,000 but is not far off it.


So PrEP could be theoretically cost-effective but for it to be cost saving in high-income and high-price countries, its use needs to be restricted to those most in need, adherence needs to be maintained and drug costs need to fall, probably considerably – which may be unlikely until 2018 at the earliest, when tenofovir comes off patent.

It also needs to be integrated into combined programmes of condom provision, efficient testing and linkage to care for people with HIV, STI screening, support for alcohol and drug dependency, partner notification, and continued sexual health education and support programmes in the community for it to reach its full potential.

A previous model by the New York researchers found that a comprehensive prevention programme featuring all these elements apart from PrEP could save around $250,000 per infection averted if lifetime HIV treatment costs are taken into account. While the portfolio of services suggested may be difficult to provide in practice, this is a reminder that PrEP may have a long way to go to become a generally affordable intervention.


Kessler J et al. Evaluating the impact of prioritization of antiretroviral pre-exposure prophylaxis in New York City. AIDS, early online publication. doi: 10.1097/QAD.0000000000000460. September 2014.

Liu A et al. Patterns and Correlates of PrEP Drug Detection among MSM and Transgender Women in the Global iPrEx Study. J Acqui Immune Defic Syndr, early online publication. DOI: 10.1097/QAI.0000000000000351. September 2014.

Kessler J et al. Averting HIV infections in New York City: a modelling approach estimating the future impact of additional behavioural and biomedical HIV prevention strategies. PLOS ONE 8(9): e73269. doi: 10.1371/journal.pone.0073269. September 2013.