A cost-effectiveness model of HIV pre-exposure prophylaxis (PrEP) for gay men based on the data from the PROUD study, and adapted to the UK epidemic, finds that providing PrEP from 2016 will be cost-effective if gay men continue to test for HIV at the current rate, are referred for PrEP using the same criteria as PROUD, and do not substantially reduce their existing rates of condom use.
In this base-case scenario, PrEP would cost £9466 for each additional healthy life year gained (quality-adjusted life year or QALY), compared with not introducing PrEP.
This is the case if current list prices of drugs remain as they are. However, the two drugs in Truvada (tenofovir and emtricitabine), the pill used in PROUD, come off patent from 2017. If these drugs and other drugs used to treat HIV are halved in price when their patents expire then PrEP could even be cost-saving, resulting in the NHS paying less than it would have done had PrEP not been introduced.
This assumes, however, that there is no increase in HIV testing rates because gay men hear about PrEP and want to get it, or that men who currently use condoms 100% choose to start using PrEP instead of condoms.
What would happen if large-scale interest in PrEP led to an upsurge in HIV testing in gay men, especially if PrEP’s popularity also led to fewer men using condoms? In this case, PrEP would not be cost-effective unless drug prices fall substantially – by 50 to 80% – after drug patents expire.
The model finds that PrEP would be cost-effective if different definitions of ‘higher risk’ are used too. If PrEP is only offered to gay men who have had condomless sex with at least five partners in any three-month period in the last year, then PrEP would be cost-saving even at current drug prices.
A second cost-effectiveness model published in September 2015 has since come to substantially the same conclusions using different methods, also emphasising that PrEP may not be cost-effective if too many lower-risk STI clinic attendees start receiving it.
The model - more details
The model, developed by Valentina Cambiano of University College London, computes the cost-effectiveness of PrEP for five levels of HIV risk in HIV-negative gay men, from lowest to highest.
- Any gay man who had had condomless anal sex in the last three months (the risk criterion for joining PROUD).
- Any gay man who had had condomless sex with at least one short-term or casual partner in the last three months, rather than any partner.
- Any gay man diagnosed with a bacterial sexually transmitted infection (STI – primarily gonorrhoea, chlamydia or syphilis) in the previous three months.
- Any gay man who had had condomless sex with at least five short-term partners in any three-month period during the last year.
In the case of the first group, the model then poses two scenarios.
1a. HIV testing rates in gay men remain at the current level, meaning that 94,900 would get tested in the model’s first year (2016). Furthermore, there is no fall in gay men’s condom use.
1b. Increased awareness of and interest in PrEP leads to a big upsurge in gay men coming forward for an HIV test, in order to get PrEP, and 165,800 get tested in 2016. Furthermore, 25% of gay men who were using condoms start using PrEP instead. This drop in condom use is modelled as a ‘one-off’ event.
The model assumes that not everyone offered PrEP and who could benefit from it will want it; in all scenarios, the rate of acceptance when PrEP is offered is 50%.
It uses the current UK ‘list prices’ of HIV treatment (Atripla) and PrEP (Truvada) – respectively £6488 and £4331 – and finds that, including other healthcare costs, the cost of a year’s HIV treatment is about £11,000 and of PrEP £5000.
In the base-case scenario (1a), the model predicts that out of all the gay men in England and Wales, 11,800 – only about two per cent – would start PrEP in 2016.
In scenario 1b, an immediate increase in interest in PrEP would mean that many more – 81,400 or 14% of the gay men in England and Wales – would start PrEP. This is partly because the increase in testing that interest in PrEP prompted would lead to an increase in the number of people who are offered PrEP.
However even under scenario 1a – no immediate upsurge in HIV testing – the number of men on PrEP would increase substantially over time. By 2031 under scenario 1a, 46,400 gay men (7%) would be on PrEP. Under scenario 1b it would be 105,500 (17%).
It is also assumed that, while on PrEP, men continue using it even if they go through low-risk periods, as long as they continue to have any condomless sex. PrEP, it is assumed, would be stopped only if they resume 100% condom use or if their only partner is an HIV-positive primary partner who is virologically suppressed on antiretroviral therapy.
The model finds that by 2031, the number of men who have sex with men seen for HIV care would be slowly declining whereas with no PrEP, it would continue to increase. Stretching out the model to its maximum length – 80 years – the model finds that over that time, under scenarios 1b, 2, 3 and 4 there would be 84% to 86% fewer HIV infections than there otherwise would have been without PrEP. Because scenario 1a assumes no annual increase in HIV testing, then the reduction in HIV infections is somewhat smaller, at 72%. Deaths due to HIV would decrease by 10% to 13% under all scenarios.
The model finds that even without PrEP, the cost of HIV care peaks at £800 million by 2036 and thereafter declines to half that amount by about 2070, because of the fact that the great majority of people with HIV are on ART and virally suppressed.
With widespread PrEP provision, the cost of HIV treatment would peak at about £600 million in the early 2020s and then dwindle away to £100 million or less by 2070. However, under scenario 1b, unless drug prices fall, the cost of PrEP would nearly double the peak HIV budget to over £1000 million in the late 2020s. It would stabilise thereafter at about £700 million. If there was a 50% reduction in the cost of ARVs this would come down to about £400 million.
Cost-effectiveness – the figures
The model finds that for risk group 1a and with no drug price decrease, the cost of PrEP would be £9466 per quality-adjusted life-year (QALY) gained: this is below the UK threshold for cost-effectiveness, which ranges from £13,000 to £30,000.
In scenario 1b, many more gay men would start PrEP – including many who would not have acquired HIV – and also assumes a fall in condom use. Because of this, scenario 1b is well outside the cost-effectiveness threshold: PrEP at 2016 prices would cost £57,145 per QALY gained.
For scenario 2 (PrEP for men who have had condomless sex with a casual partner, rather than just any partner, in the last three months) the cost per QALY gained is £39,314, which is still not cost-effective.
But for scenario 3 (PrEP for those diagnosed with an STI in the last three months), PrEP costs £9290 per QALY gained, which is cost-effective: and for scenario 4 (PrEP only for those who have had at least five condomless sex partners in any three-month period last year), PrEP would actually save money, saving £1522 million over the model’s entire 80-year course.
If drug costs fell by 50% then PrEP in scenarios 1a and 3 would also become cost-saving (£471m and £660m saved respectively over 80 years) and scenario 2 would become marginally cost-effective (£13,010 per QALY saved). If drug costs fell by 80%, then PrEP in scenario 2 would become cost-saving too and for scenario 1b would cost £3934 per QALY gained, which is cost-effective.
In scenario 1b, if people who start PrEP actually use it only half of the time rather than take it full time then, even if some of the time they are not on PrEP they have condomless sex, this brings the cost per QALY down to £22,000 which is marginally cost-effective, or £4000 if drug prices are also reduced by 50%.
In summary this analysis finds that, if the effectiveness levels seen in the PROUD study are achieved in a general PrEP provision programme, PrEP in the UK will be cost-effective when:
- HIV testing rates do not increase, PrEP is only started in men referred by clinicians when they take an HIV test, rather than ‘en masse’, and a substantial drop in condom use does not occur; or
- PrEP is quite highly targeted at only those who have condomless sex with five or more short-term partners, or present with a recent bacterial STI; or
If, in response to the availability of PrEP, testing rates do rise and condom use falls:
- The cost of drugs fall by at least 50%, and PrEP is only offered to people who have condomless sex with non-primary partners; or
- People take 50% or less of the Truvada they would take if they took it daily; or
- The cost of drugs falls by 80%.
Further sensitivity analyses will be done in future to test what would happen under other scenarios, Dr Cambiano told Aidsmap.
Cambiano V et al. Is pre-exposure prophylaxis for HIV prevention cost-effective in men who have sex with men who engage in condomless sex in the UK? BASHH Spring conference, Glasgow. Abstract #01. 2015.