The theoretical ageism argument and NICE's technical response

The theoretical argument is predicated on the fact that the calculation of incremental cost effectiveness underpins NICE's appraisal decision making.5 This is true, but it misses both the improbability of an ageist calculation and the discretion of the appraisal committees to consider other factors, particularly discrimination, too. The boundaries of this discretion are laid out in the social value judgements document discussed below.

NICE assesses the value of a new intervention by taking into account not only how effective a new treatment is but how much it costs. This is achieved by deriving the incremental cost-effectiveness ratio (ICER). This is calculated by dividing the difference between the cost of the new treatment and the cost of the comparator treatment (ie, the normal treatment currently available to patients), by the difference in the benefits that they offer to patients. These benefits are expressed as QALY. QALY combine a calculation of length of life with quality of life.6

The ICER is expressed as a cost per QALY. NICE's advisory committee considers the ICER in the context of the threshold range that the Institute is advised represents what the health service in the UK, in practice the NHS, can afford, ie, the level of the cost per QALY that would ensure that any new intervention did not displace more health than it added. The value that the Institute uses is a range of £20 000–30 000 per QALY. The Institute has commissioned through the Medical Research Council and National Institute for Health Research a programme of research to test the validity of these values.7 However, advisory committees are expected to use judgement in applying the threshold. Above a most plausible ICER of £20 000 per QALY gained, judgements about the acceptability of the intervention as an effective use of NHS resources should specifically take account of the following factors:

1. The degree of certainty around the ICER. In particular, advisory bodies will be more cautious about recommending a technology when they are less certain about the ICER presented in the cost-effectiveness analysis.

2. The presence of strong reasons indicating that the assessment of the change in the quality of life is inadequately captured, and may therefore misrepresent the health gain.

3. When the intervention is an innovation that adds demonstrable and distinct substantial benefits that may not have been adequately captured in the measurement of health gain.

As the ICER of an intervention increases in the £20 000–30 000 range, an advisory body's judgement about its acceptability as an effective use of NHS resources should make explicit reference to the relevant factors considered above. Above a most plausible ICER of £30 000 per QALY gained, advisory bodies will need to make an increasingly strong case for supporting the intervention as an effective use of NHS resources with respect to the factors considered above.

The variation of the anticipated gain in length of life between people of different ages means that there are greater theoretical gains for a younger person in receipt of a life-saving treatment than for an elderly person. While a 20 year old can expect longevity of approximately a further 55 years, a 45 year old would have longevity of 35 years, and a 70 year old a longevity of 15 years. Saving the life of a 20 year old, therefore, achieves greater QALY gains than saving the life of a 70 year old. However, this theoretical possibility that the QALY is inherently ageist does not stand up when the actual QALY calculations NICE makes are more closely considered.

Table 1 shows what the purely cost-effectiveness-based decisions would be for treatments generating a year of life gain and for life saving treatments for a 20 year old, a 45 year old and a 70 year old at different costs. It illustrates a point argued by Paulden and Culyer's8 economic modelling of NICE's decision making that ageist discrimination as a result of the application of QALY could only occur if the technology in question fits very specific conditions: the ratio of costs to QALY must be lower over the long term than over the short term; the ICER for patients with shorter life expectancy must lie above the cost-effectiveness threshold; and the ICER for patients with longer life expectancy must fall below the cost-effectiveness threshold.

Lines 1–3 of table 1 show that for the 1 year gain decisions, age discrimination cannot apply because the longevity headroom is, by definition, identical for all age groups. The clinical evidence for life extension of a few months to a few years could, of course, show a differential advantage for different age groups, but in practice there are relatively few age-stratified trial results, and it is NICE's practice to assume that what applies to one age group within a particular appraisal will apply inter alia to others.

A more potentially problematical situation arises when normal life expectancy is gained over potential imminent death. Lines 4–6 of table 1 show that when normal life expectancy is attained for £450 000, £700 000 and a million pounds, the ICER vary between a 45 year old and a 70 year old, but the decisions do not discriminate except in the one million pound category. Although life expectancy gained for £700 000 yields an ICER above NICE's normal £30 000 per QALY upper limit in the case of a 70 year old, in fact the resumption of normal life expectancy, when there would otherwise be early death, is covered by NICE's extended end-of-life rules (see below) and an ICER of £47 000 is still acceptable.

So what happens when there is a million pound treatment that is curative in the elderly? NICE has never faced this situation but it is highly unlikely that a knee-jerk QALY-based decision would be made. This hypothetical situation is of questionable significance, not only because of its rarity (it has not yet occurred), but also because NICE appraisal committees are not obliged to accept blindly the outcome of this selective part of its methodology guidance. Indeed, NICE's track record to date of assuming an identical (trial-based or averaged) benefit across all age groups reflects the fact that it is incumbent on them to consider the possibility of ageism. The appraisal committees can go further than standardising QALY gain assumptions across age groups. They retain authority to exercise discretion up to the point at which to do so would be perverse.

NICE's procedural protections against ageist decisions

Arguments claiming that NICE must be ageist because it uses QALY have also not considered the context in which QALY are applied: NICE has implemented processes, built around principles of transparency and stakeholder involvement, to safeguard against ageist decisions.9 These include its equality and social value programmes, and recent additions to rules regarding end-of-life treatments. They ensure that all stages of the appraisal process are open and visible and the evidence and reasons behind recommendations accessible, that there is systematic examination of implications for equality, and that public and patient input is actively sought.

Actual NICE decisions and age

Arguably, the success of the two layers of protection NICE has implemented to avoid ageism is best judged by a review of NICE's track record. Table 2 shows the actual decisions NICE has taken for the years 2008 and 2009. It shows both the decision taken and whether there was any age restriction. It is clear that not all decisions were ‘yes’ decisions, although they are in the clear majority. However, of the ‘yes’ decisions, it can be seen that only four had restrictions according to age. These concerned osteoporosis (two decisions), insulin pump therapy for diabetes and neuraminidase inhibitors for influenza. In the neuraminidase case, the discrimination was in favour of the elderly (over 65 years) because of the clear relationship of increasing age with influenza morbidity and mortality risk. In this case, NICE was only ageist in the opposite of the usual sense, and only in a case in which the epidemiology was very clear.

A similar situation arose in the post-menopausal osteoporosis decision. The epidemiology of fracture risk in osteoporosis rises steeply with age after the menopause and the cost-benefits of treating relatively young women (those in their 50s) are much less favourable than the cost-benefits of treating much older women. So there were restrictions in the decision. However, these were rather limited. In the case of secondary prevention, the principal drugs for osteoporosis (the bisphosphonates) were made available to all women over 50 years of age irrespective of cost-benefit, on the grounds that the minority of younger women at risk would have constituted an overzealous restriction. The only other example of age restriction was in the case of insulin pump therapy, when children and adults were treated differentially. The child/adult issue is a sort of age discrimination, but again, one that is not what NICE's antagonists have in mind when making an ageism attack on NICE. It is, however, one that NICE will continue to consider as it explores further the impact and development of its methodology.