Net Monetary Benefit is becoming an increasingly common way of expressing value in health economic studies. I’m a bit slow in catching up.
This study caught my eye recently.
QALY gain and health care resource impacts of air pollution control: A Markov modelling approach
Environmental Science & Policy
September 2016, Vol.63:35–43, doi:10.1016/j.envsci.2016.05.015
It’s a really neat study, there was an even more interesting study on a similar vein by the same group in BMJ open a month or so later exploring the cost and return of selected air quality policy interventions.
The Schmitt study is sadly paywalled. You can glean a good bit from the abstract
“Reducing mean PM2.5 concentrations by 1 μg/m3 in London and in England and Wales is expected to yield more than 63,000 and 540,000 QALYs respectively, to adults aged 40 and above over their remaining lifetime, discounting at 3.5% p.a.”
So good, reducing air pollution helps people avoid illness and early death. This seems a good thing.
What about the money:-
“At a WTP value for a QALY of £65,000, which is in line with recommendations for the UK, the expected discounted monetary benefit of the intervention amounts to £4 billion in London and £34 billion in England and Wales.”
NB- This seems considerably higher than what is commonly accepted as the benchmark in England and more than four times what prof Claxton and others have suggested as a true willingness to pay in NHS decision making.
This is important stuff. And this is complex. It depends upon what definition of WTP you are using – a demand side definition or a supply side definition. If it is the demand side definition – which implies the budget will increase to cover the cost of any additional intervention that has a cost per QALY < = 65k, then this is correct. If it is the supply side definition of WTP then what this means is that the least efficient intervention currently funded has a cost per QALY of 65k, and that any technology that produces QALYs for less than 65K should be adopted AT THE EXPENSE OF THAT LEAST EFFICIENT TECHNOLOGY and this will lead to an increase in the amount of health produced from the existing budget, by replacing a less efficient technology with a more efficient one.
The willingness to pay (WTP) of £65k / QALY implies that we are happy to buy additional health (here measured using the QALY as a metric) at an additional cost of £65k – ie we are happy to see less valuable things displaced (things that cost more than £65k / additional QALY).
So…… Where on earth does the £65k threshold figure come from!? It’s possible that the 65K threshold is a demand side and hence irrelevant to most real world policy decisions.
If the aim is to pay for this from NHS budget then the relevant WTP is closer to Claxton’s 13k.
Asssume £65k stands.
The assumption is that at this willingness to pay for the threshold the “expected discounted monetary benefit of the intervention amounts to £4 billion in London and £34 billion in England and Wales.”
This is not, I repeat not, a cash saving of £4 of £34Bn.
it is the estimated value of the health benefit.
Figure 2b (sorry about the paywall, not of my doing) indicates this wont save money.
In health service parlance, if the investment is funded from lower value activities in the same sector or others then there will be a likely increase population health, but probably not a cash saving.
I asked my favourite Prof @MccabeCJM whether I’d be right to say that in a zero cash environment a commissioner’s WTP in a NMB sense is ZERO as all investments must be net £ save, otherwise someone else has to bear the opportunity cost?
And he said…….
” not really – because we do want to replace less efficient interventions with more efficient interventions. So operating in the supply side model of the WTP for heatlh, having a positive WTP is rational. The challenge making sure that the savings that will theoretically cover the budget impact of the new technology are actually achieved in cash terms.
I think the future of risk sharing/patient access schemes is in contracts that base payments on the achievement of the promised health gains but also the promised cash savings.
I wouldn’t express it as a WTP =$0 – because we do want to incentives more efficient interventions (including service developments etc..) but the issue of its not cost effective if the savings that make it look cost effective are not achieved, is a real and important point.
So I would take the point you are making and feed it into a consideration of the design of payment by results procurement contracts.”
Sadly the reality is the former of Chris’s statements – most often we can’t release the £ and decreasing budgets overall.
This is geeky stuff
And I’m not necessarily picking on air quality interventions per se, I think it’s mighty important to improve air quality – basically drive less, walk more, plant trees etc.
But I see this methodology of net monetary benefit expressed with a willingness to pay threshold increasingly in health service economic evaluations. Often there seems to be an assumption on the part of authors that the WTP in the real world is in line with where NICE and other bodies seem to settle on a decision making threshold.
Again Prof M had a mighty useful view on this
“This is again complex. When the government sets the NHS budget it should – and historically did – include an expected growth in the drug budget based upon forecasts which included consideration which drugs NICE was likely to approve over the forecast period given NICE’s threshold. So there is some degree – historically – to which NICE’s threshold drove increases in the NHS budget. Obviously this is no longer the case.
The really good thing about the Net Benefit approach it is not open to interpretation in the same way as an ICER. If you have an ICER then you can play around with what the appropriate threshold/WTP should be. If the decision maker commits to a Net Benefit approach they are locked into a WTP and the result simply tells them whether it is good value (NB>=0) or bad value.”
the WTP for that study is way too high.
Seems incredulous that DH has a WTP of £65,000. The amount of money allocated to the NHS definitely does not reflect this.
Sadly, the real world isn’t like that. Nobody can afford anything. And decisions based on this basis only expose more (mostly anonymous) people to the opportunity cost – mostly on the basis of other services they won’t get.
If the WTP is >£0 then the NHS has to stop doing something or someone bears the opportunity cost. The question then becomes what we stop doing… I don’t think anyone would like the answer if I point out some of the least cost-effective most cash releasing treatments in the NHS. I don’t think the problem is the maths – it’s the politics (and ethics, and equality).
Dear economists –
Maybe as a sensitivity analysis – try doing the analysis with a WTP of £0?
If we use a WTP of $0 – only cost saving technologies will have a positive Net Benefit. Less effective but cheaper technologies would have a positive Net Benefit – so there is a risk in going down this line. A two stage process would be required – ‘Is it more effective; if yes does it have a positive net benefit.
And this would not get around the problem that the analysis says it is cost saving but if the cost savings are not actually realized, then it still exposes anonymous people to harm through opportunity cost.
Or do this as base case, and assume different thresholds that reflect willingness to pay on the assumption that there might be some opportunity to disinvest and or new funds to invest.
Thanks to Rachael Hunter, UCL and Prof @mccabecjm