Should NICE’s threshold range for cost per QALY be raised? No
BMJ 2009; 338 doi: https://doi.org/10.1136/bmj.b185 (Published 26 January 2009) Cite this as: BMJ 2009;338:b185
- James Raftery, professor of health technology assessment
- raftery{at}soton.ac.uk
The fact that the National Institute for Health and Clinical Excellence (NICE) has not updated its cost effectiveness threshold over the past decade means that the threshold has been falling. This applies whether adjusted for inflation (up 40% from 1999 to 2007) or for the NHS budget (up 90%).1 This decline is appropriate for several reasons. The correct threshold value, which should be set by the value of those technologies displaced by NICE guidance, seems to be lower. NICE has recently increased the threshold for end of life treatments. To offset this, the general threshold should be reduced. Precedent in the form of treatments previously funded has arguably influenced what people think the threshold should be. The most plausible precedent thresholds are no higher than those of NICE.
Displaced services
Opportunity cost, a key concept in economics, expresses cost in terms of the opportunity forgone by buying one thing rather than another. It plays a crucial part in ensuring scarce resources are used efficiently. Since primary care trusts fund the bulk of NHS healthcare, they must deal with the opportunity costs of NICE guidance. To fund NICE guidances primary care trusts must either cut existing services or not implement new services. If displaced services are more cost effective than the NICE threshold, that threshold is too high. NICE guidance would be perverse, reducing not increasing health.
The Commons’ health committee concluded, largely on the basis of evidence from primary care trusts, that NICE’s threshold was too high.1 Two trusts have controversially appealed against NICE guidance in favour of the costly drugs trastuzumab and ranibizumab (both were rejected).2 3 Unfortunately, no systematic data are collected on services displaced by NICE guidance. However, examination of the opportunity costs of NICE guidance on trastuzumab for a Norwich hospital indicated that more cost effective oncology services had to be sacrificed.4 Analysis of the cost per quality adjusted life year (QALY) of annual changes in primary care trust spending shows values well below the NICE threshold.5 This analysis covers three years and may indicate that national service frameworks for particular diseases have been very cost effective. No evidence points the other way—that technologies recommended by NICE are more cost effective than those displaced.
NICE has just raised the threshold for life extending drugs for end of life patients.6 Given the high cost per QALY for the relevant cancer drugs, the new threshold will have to be substantially raised.7 If many cancer drugs qualify, the financial impact on the fixed budgets of primary care trusts will be substantial. Unless the general threshold is lowered to offset these new arrangements, even more services will be displaced.
Precedents
Economists see no theoretical basis for any threshold other than one based on opportunity costs. However, thresholds have plausibly been set as much by precedent as by theory. The first threshold may have been for US government funding of renal dialysis, which was introduced in 1973 and was roughly $50 000 (£30 000) per patient for each extra life year.8 Such benchmarks make it difficult ethically and legally to refuse to fund any treatments of similar cost effectiveness. NICE’s appraisal of dialysis put its cost at around £20 000 a year9 (or around £30 000 per QALY10) Given that NICE’s appeal process involves precedence, this provides some justification for the current thresholds. To avoid exceptional cases setting precedents the case for their funding needs to be explicitly justified and monitored over time. Otherwise benchmarks set in a storm will be repented in leisure.
Economists closely linked to NICE have addressed the threshold problem by characterising NICE as a “threshold searcher.”11 They accept that the threshold should be based on the opportunity cost in terms of displaced interventions, and that it should be lower for recommended interventions which displace more (have a large budget impact). They argue for a single threshold rather than a range and that it should be flexible or dynamic, depending mainly on the size of the NHS budget and efficiency gains. Audaciously they suggest that the case for any rise in the threshold linked to the growth in the NHS budget in the past decade may well have been offset by improvements in NHS efficiency.
The key question is whether NICE can be left to search alone for the threshold or whether it needs help. One suggestion has been for an independent threshold committee, similar to how interest rates are set.1 12 Alternatively, NICE might be made more responsive to primary care trusts by increasing their representation and by improved links with their annual purchasing plans. Since displacement decisions are made by hospitals, they too might be better represented at NICE. Another suggestion is that NICE should develop a disinvestment programme based on identifying treatments with poor cost effectiveness. Several of these options are being researched by NICE.1 Either way, NICE must relate more closely to those who make decisions in the NHS. Improving those links rather than increasing the threshold must be the priority.
Notes
Cite this as: BMJ 2009;338:b185
Footnotes
Competing interests: None declared.