A cost-effectiveness model developed by Public Health England finds that just one year’s worth of pre-exposure prophylaxis (PrEP), given to gay men during a period of moderate to high HIV risk, would recoup its own costs and save the NHS money within 23 years if PrEP effectiveness was as high as that seen in the PROUD study – and if PrEP drug prices remained at the current level.
However, if drug prices fell by 90% as a result of the availability of generics, one year of PrEP would recoup its costs within five years.
A year’s worth of full-cost PrEP given to 5000 gay men at high risk of HIV would cost the NHS £27 million, and a year of generic PrEP, at 10% of the cost, would cost £3.5 million.
The model differs from another one developed by University College London which was published almost simultaneously in The Lancet Infectious Diseases. Whereas that one was a dynamic model which mapped out the savings or cost due to a PrEP programme being implemented over a number of years, this one isolates the savings or costs due to a single year of PrEP being given to gay men who need it and shows the effect of that single year. It takes into account infections directly averted while men are using PrEP, but not the prevention of onward transmission from those men.
The model finds that the cost-effectiveness of PrEP was chiefly influenced by the following factors:
- Its prevention effectiveness. If PrEP only reduced HIV infections by 44%, as in the iPrEx study, then it would only be cost-saving with a 90% fall in PrEP prices or only if given to men who would otherwise have an incidence of HIV of over 9% a year. If 96% effective, it saves money for anyone when HIV incidence is over 3.3% a year, but only if risk behaviour does not change.
- The background incidence. PrEP would not be cost-saving if given to gay men with an annual HIV incidence of 2% a year (i.e. 1 in 50 acquires HIV per year), but it would be cost-saving with a background incidence of at least 3.3% (1 in 30), even if risk behaviour does increase. It would be marginally cost-effective (just over £20,000 per quality-adjusted life year [QALY] saved) if 96% effective and if given to men with 2% background incidence. This is if drug prices do not fall.
- Its cost. PrEP will not be cost-saving, though it would still be cost-effective, meaning it would cost less than £20,000 per QALY saved, at an HIV incidence of 3.3% and an effectiveness of 86%, if drug prices do not fall at all. It would remain cost-saving at 96% effectiveness. However, it would require PrEP prices to fall by 43% for PrEP to be cost-saving at 64% effectiveness, and by 90% for it to be cost-saving at 44% effectiveness.
Other factors had some influence over the outcome. The cases above assume that the cost of antiretroviral therapy does not fall, but it is also likely to do so, if perhaps not as much as PrEP (because it will tend to involve more, and newer drugs). If antiretroviral therapy costs fell by 30% or more, PrEP would still be cost-effective but not cost-saving. However, this is only if PrEP costs do not fall at least as much.
If PrEP is only taken on demand, such that the average dose taken per week is four rather than seven pills, then PrEP becomes cost-saving even at current drug prices if more than two-thirds of PrEP takers use it on this basis.
The model’s base case also assumes that men taking PrEP would have a 20% increased risk of HIV due to reduced condom use, which would of course slightly increase the infection rate unless PrEP was 100% effective. But including this in the analyses did not change the cost-effectiveness as much as the other variables.
Some limiting factors
The model includes some interesting assumptions that are based on data from sexually transmitted infection (STI) clinic attendees in England and from other surveys of HIV incidence in gay men.
Firstly, these data show that the majority of gay men do not stay at high risk of HIV infection for more than a year at a time. After a year only 38% of STI patients who had started in the high risk group at the beginning of the year were still in it. However, other gay men would be joining or rejoining the high-risk group at the same rate, so unless HIV infectiousness fell generally (as seems to be happening) the proportion of gay men who were at high risk of HIV would not fall.
This does assume however that men might need to stay on PrEP for a year or two in general – and that cost-effectiveness would be reduced if they did stay on it.
Secondly, and of more significance for the outcomes of the model, Public Health England points out that just because men join the ‘low risk’ cohort, their risk does not fall to zero, and they may stay at low but significant risk for many years. This means that the year’s worth of PrEP assumed in the model only actually prevents a minority of the HIV infections the men taking it might acquire in a lifetime.
It calculates that without any PrEP, the 5000 men assumed in the study would acquire 165 HIV infections in that first year – but 683 during their subsequent lifetimes.
Nonetheless, despite a deliberately limited set of assumptions about the effect of PrEP, this model finds that under the most likely scenarios, PrEP given to gay men at high risk of HIV would recoup its costs over a relatively short period of time, especially if drug costs fall significantly and it is only taken by men at higher risk.
Ong K-J et al. Economic evaluation of HIV pre-exposure prophylaxis among men who have sex with men in England in 2016. Eurosurveillance 22(42):pii=17-00192. 2017.