As a general rule, the credibility of the output can be no better than that of the input data, and in radon even the base figures, for numbers of attributable deaths for example, may still be uncertain by a factor of five. Nevertheless policy must be formulated and the public advised, and it must be left to 'judgement', and to vested interests, to decide priorities.

It is essential to differentiate between public ("public health") and private expenditure. It is also necessary at the outset to recognise what is wrong with some published work on radon.

The first central factor is that life years should be used rather than lives saved because of the relatively small number of years of life extension per lung cancer patient. However, estimates here seem uncertain. Radon mitigation does not prevent damage and injuries amongst young people especially, as do seat belts and smoke alarms, to mention but two of the subject areas from which statistics have been selected.

The second central factor is that marginal cost benefit must be calculated, if only because of the shape of radon distribution curves. In simple terms, there are few houses at extreme radon levels, many at moderate levels that should not cause concern for a few decades of occupancy, and millions at quite normal levels, but which deliver most of the collective dose. Marginal analysis is quite standard in both health and building physics. In principle the worst of the problem should be dealt with, then the residual problem assessed to determine if further action is worthwhile.

Radon is not straightforward to analyse if only because smoking may produce more than a factor of ten increase in risk from any given radon exposure. Within Health Economics it has been acknowledged that it is effective to stop people smoking if it is ignored that they may live longer and draw their pensions so adding to the burden on the State.

There are two key questions, and analysis should underpin policy decisions.

If the aim is to reduce lung cancer, how best to allocate resources?

How best to address cancer in any household?

The usual assertion on radon is that up to 2500 people in the UK may suffer a radon-related death annually. Most of these are attributed to radon in houses. Choosing therefore a base figure of 2000, and considering only alteration of the housing stock, as many as 500 non-smokers may die from domestic radon annually in the UK. If all smokers became non smokers, the 1500 would reduce their risks by (say) a factor of 10, thus becoming 150 to add to 500, but this is valid only over decades.

Treating the housing stock to reduce collective dose.

There are about 22 million dwellings in the UK. Radon treatment in ten million of them, selected to be houses rather than high rise flats, for example, might cost £500 to £1000 each (a few houses in Cornwall cost £2000 to cure) but with uncertain effect on fractional reduction in the lower level houses.

If it is assumed that systems and procedures would work on average to good effect and that half of collective dose could be avoided, 250 out of the 500 lung cancers might be avoided for an initial cost of £5000M to £10,000M. More realistically, treatment might be envisaged for the 80,000 to 100,000 houses that are estimated to have radon levels above 200 Bq/m³, the so-called action level. However, living with this level for 10 years could give a lifetime risk of less than 0.1 to 0.2%, hardly a cause for concern.

The average life of a house is often taken as 100 years but at current replacement rates many may have to last longer. However, NRPB in some of their calculations have taken 50 years as the remaining life of a house, and this will be used here. It is important to consider running and maintenance costs over this period.

Illustrative calculations

It is assumed that the average REDUCTION in radon level would be 200 Bq/m³ in 80,000 houses each with 2.5 occupancy. The collective dose avoided per year is therefore 2000 Sv. For an average population the BEIR IV figure of 0.035 deaths per Sv applies, but for non smokers can be rounded to 0.01 deaths per Sv, perhaps 0.015. However, 80,000 houses successfully treated may be optimistic, and an average 200 Bq/m³ reduction may be too low, but the result is about right: about 3% of collective dose from radon in the UK might be avoided.

So action in 80,000 houses (a part of which would be monitoring probably at public expense in one million homes at a cost of around £30M to £40M) might avoid 2000 x 0.01 or 20 non-smoking cancers per year. The mitigation cost would be 80,000 x £1000 average per house (= £80M) plus running costs over 50 years. Ideally new houses would be 'radon proofed' but that is another interesting story, see Section 32.

Maintenance costs for systems using a 75 watt fan comprise about £50 for electricity per year plus a new fan every 5 to 8 years (the cost of which is assumed to be £120 plus installation). Also, it may be assumed that an average annual maintenance cost of £40 would include some allowance for re-monitoring. Extra energy costs for space heating of between £5 and £50 (say £10) would also be incurred, because many systems draw heat from the house as well as radon from the soil. Overall, £40 is less than the maintenance cost of many domestic burglar alarm systems, so seems reasonable. Thus the costs over 50 years are 80,000 x £100 x 50 = £400M.

However, discounting future expenditure could reduce the total of £480M in net present terms to £300M or less. This calculation is very sensitive to the real interest rate, and it is probably more helpful for illustrative purposes to assume a fixed value in real terms for the running costs each year, and a fixed dose avoided. Over 50 years, 20 deaths might be avoided annually. Each cancer avoided may represent 3 to 5 life-years only (perhaps more), as lung cancer is a disease of older people.

Thus £300M might produce a benefit of only 3000 and 5000 life years, remembering that the calculation might be in error by a factor of five. The costs are between £100,000 to £60,000 per life-year.

However, and as an example of marginal analysis, the 'high level' houses should be remedied first. Many of these houses are easy to find and from the figures below, £6M would save 500 life-years. The marginal figures for the remainder thus become £294M and between 2500 and 4500 life-years.

According to NRPB statistics there are only around 2000 houses in the UK with indoor radon levels over 1000 Bq/m³. Their average level may be around 1300 to 1400 Bq/m³: there are very few houses indeed over 3000 Bq/m³. The reduction possible may be an average of 80%, undertaken in 80% of the houses (some people will not bother, some systems will fall into disuse, some householders will switch off the fans to save money, etc).

However, dose AVOIDED may be estimated well from 1600 houses, 2.5 people per house, and 50 mSv each per year. The annual total is 200 Sv, or 200 x 0.01 = 2 non-smoking deaths per year.

These houses are 'best value' in public health terms: cost would be perhaps £2M initially and another £100 per year over 50 years for upkeep (see above), or 1600 x £100 x 50 = £8M.

Overall, this is a maximum of £10M (discounted to £6M or less) to save 2 or 3 non-smoking lives per year over 50 years. The cost per life-year benefit is therefore only £12,000, assuming 500 life years in total.

The cost per Sv avoided may also be calculated. For the high level houses, it is only £600/Sv. This compares well with the NHS where £500/Sv to £600/Sv is a rule of thumb, when they can find the money.

If only the initial costs are considered, as might be the case for public expenditure, cost per life-year falls to perhaps £4000, which is on a par with breast cancer screening or heart transplants for cost per QALY, according to published figures from Health Economics.

In contrast, NRPB have used £210 million costs and £30,000 per 'life saved' as applied to all the affected houses. However this is skewed by including smokers and by not considering life-years, and inadequate because it does not include marginality.

Despite that high radon exposure combined with smoking may lead to an early death, the way to reduce the risk is to stop smoking, and then to address the residual risks from radon.

For readers who are not convinced by this, consider a situation in which dozens of children are killed each year because they walk across roads whilst blindfolded. The high technology solution to reducing deaths might be to fit anti-collision radar to all cars at a cost of many billions of pounds. It is a solution that would appeal to some people.

The more rational solution would be to ensure that blindfolds were removed before children attempted to cross busy roads. This could be done (one supposes) very easily, and 99% or more of the accidents might be avoided. Reducing lung cancer deaths by phasing out smoking is similarly straightforward.

Much of the above is simplistic, but it is more soundly based than are some of the arguments used in the USA to justify extremes of expenditure on radon. The most common type of error is (for example) to avoid recognising that seat belts not only save lives they save large numbers of injuries and much expense on hospital care.

As an aside, radiologists are sometimes furious about the marginal expenditure at Sellafield: up to £150,000 per Sv and "a tragedy of radiological protection", so some say.

The point about radon is that some at NRPB have viewed with jealousy the budgets of EPA in the USA, and closer to home would wish to see even expense on the Sellafield scale applied in houses. This is not sensible on a public health basis.

Perspectives of radiological protection.

For perspective, many people might reduce their calculated total cancer risks more by eating a few pieces of fruit a day than by worrying themselves about radon exposure over the next five years. The highest level houses merit action, but criteria cannot be those of limiting dose to levels set for work-place exposure: recently advised to be reduced from 50 to 15 mSv per year maximum but with very few workers receiving more than 2 mSv.

Radiological protection must be removed from its position of privilege, and be content to compete with medical budgets where the choices are within the home and community.

The key perspective here can be difficult to rationalise: in terms of collective dose avoidable, radon remediation has probably less to offer than better medical procedures and use of more modern equipment. The potential dose savings in the UK from various patient protection measures in diagnostic radiology have been estimated by NRPB at 5000 Sv annually, expressed as effective dose equivalent.

This estimate assumed a cost per Sv of £500 to £600 - the most the NHS could probably afford whilst not diverting resources from other and more promising areas.

However, much 'preventable' dose in medical radiology occurs at a low dose per patient, in contrast to the severe doses delivered in the highest level radon houses. Calculations using collective dose as the sole criterion are flawed, as the reaction to Chernobyl has illustrated (see Section 45).

One possible approach would be to deal with the highest doses almost irrespective of cost (otherwise identifiable people will remain at high and unacceptable risk) and to question whether low personal doses need to be addressed at all - and irrespective of the magnitude of the collective dose avoidable. Any benefits such as medical diagnosis or keeping a job (for miners) would need also to be considered.

Thus, it may be questioned whether householders living with a domestic radon level above (say) 2000 Bq/m³ should not benefit from public assistance, and whether purchase of new diagnostic equipment in hospitals is justified if the old units never impart more than a few tens of mSv to any individual. It is no doubt an area for future controversy!

Simplified calculations for radon in houses.

These are probably the most realistic representation of radon economics, as it may be assumed that 'someone else' paid for the system in the past. The homeowner (or tenant) has to decide for each period:

"Do I run the system to avoid this annual dose?"

Thus, initial monitoring and installation costs are ignored, and attention focuses on the 1600 high level houses where 200 Sv is avoided annually for £160,000. This gives £800/Sv or £80,000 per non-smoker death avoided, and illustrates the importance of running costs in real terms.

For all 80,000 houses, 2000 Sv/year may be avoidable for £8M/yr. This is £4000/Sv, or £400,000 per non-smoking death avoided. Even applying a fudge factor of two, cost per life-year would still be £40,000, assuming 5 years per avoided premature death.

It is clear therefore, even from this simple analysis, that radon remediation is not a priority for public investment except perhaps in the few very highest level houses. The position might be different if running costs could be markedly reduced by use either of smaller and cheaper fans or passive systems. Capital costs are less of a concern if the money could somehow be found (or 'lost'!) within the public health budgets of either central or local government.

As discussed elsewhere (Section 43) many of the most seriously affected houses might have been found and dealt with years ago were it not for the decision to make radon into an issue affecting tens of thousands of dwellings, and to create a programme lasting decades.

next page

top of radon section