'The cancer industry provides much advocacy and little evidence of cost-effectiveness...

The government's two-week target for cancer referrals risks distorting resource allocation in an NHS already contorted by the waiting list initiative'

There are three certainties in life: death, taxes and the scarcity of resources. Most people are afraid of death, and they know that one of the major causes of death is cancer.

Health messages continually tell the populace that much cancer is avoidable: stop smoking, drink moderately and eat sensibly. These messages tell us that cancer mortality and morbidity can be curbed. Then there is a trick. The next assertion is that most cancers can be cured. Some may be cured or delayed if caught early. But some cannot be delayed significantly despite severe bombardment with high technology.

The policy problem is that the optimism portrayed by anti-tobacco messages may often be translated into advocacy of unproven techniques. This transfer is assisted by our fear: we really want to believe doctors can reverse this dread disease.

The cancer industry is formidable. The patients' lobby and the medical profession tell us that survival rates for common cancer in Britain are lower than on the continent. They argue that more investment in their industry will bring health gains. So it may, but do these groups ever demonstrate which cancer investments give the best rate of return - health improvements - for what cost, or whether the resources used would produce greater gains elsewhere?

Parliamentarians are easily captured by such lobbies. Thus an all-party group on cancer has been established.1 Its first 'statistical bulletin' presents international data on survival rates -but it is not clear what the percentages mean: are they five-year survival rates or what?

It also presents data on cancer drug expenditure (we spend less, but the authors offer no insight into which cancer sites will yield the best health gain for additional investments). And there is data on waiting times for chemotherapy (which cites clinical standards whose evidence base and cost-effectiveness should be better demonstrated).

The parliamentary group is also looking at the staffing requirements necessary to meet the government's white paper commitment that all patients referred by GPs with a possible cancer must be seen within 14 days. The royal colleges are quite rightly opposed to this target. It is ludicrous. If trusts are required to meet it by 2000 much essential routine work will be crowded out.

For instance, a referral for 'query prostate cancer' to a trust will direct the urologist from worthy inpatient procedures to outpatient procedures where the diagnostic test is of dubious reliability (many false positives and false negatives occur) and where surgical interventions are of debatable benefit.

The staffing requirements of these targets will be significant. The royal colleges do not like the targets but they do like to adopt the role of advocate for more members. Nowhere in this advocacy do the colleges or their associated lobbies consider the issue of opportunity cost. Such behaviour, if it leads to inefficiency, is unethical.

An extra surgeon to treat cancer patients means fewer orthopaedic surgeons to repair hips. Hip replacement is very cost-effective. Which cancer surgical interventions are most cost-effective? We need answers so that scarce resources are targeted where they produce most health gain per unit of cost.

The evidence base in the cancer area is impressively bad. The pharmaceutical industry and charities fund many cancer trials and their common characteristic is bad design. Many cancer trials in recent decades have measured the end point (the outcome) in terms of additional months or years of life. Trialists have been very reluctant to measure the quality of the additional time.

The authors of a recent article searched the Cochrane controlled trial register for 1980 to 1997 to identify trials which reported, however well or badly, quality of life.2

In this period, reporting of the inclusion of quality of life measures rose from 0.63 per cent to 4.2 per cent for all trials, and from 1.5 per cent to 8.2 per cent for cancer trials.

This is a depressing finding. Purchasers want to know what the health gain is, and what gain is achieved per unit of cost. The funders of more than 90 per cent of randomised controlled trials in the period 1980-97 were not apparently interested in quality of life issues. We do not need more research and development funding. We need properly designed trials to inform purchasers.

The cancer industry provides much advocacy and little evidence of cost- effectiveness. Yet trusts are required to introduce the Calman-Hine proposals. These are based on evidence of 'under-treatment', inequalities in access to treatment (further variation by postcode) and some evidence of effectiveness.

However, purchasers and providers also want evidence of cost-effectiveness.

Until that evidence is available, the siren songs of inadequately informed clinicians have to be resisted.

There is a risk that government, with its two-week target for cancer referrals, will further distort resource allocation in an NHS already contorted by the waiting list initiative.

Clinicians, patient lobbies and government should resist the temptation to con the public with unproven cancer technologies. They must recognise that the resources they consume deprive people of mobility (from hip replacements), sight (from cataract removals) and life (from delayed cardiac surgery).

It is madness to prioritise services in isolation from evidence and without such recognition, even if this buys favour from a public so anxious to delay death by cancer that they clutch at straws offered by false prophets of salvation. REFERENCES

1 All-Party Group on Cancer, Statistics Bulletin September 1998. Available from the office of Ian Gibson MP.

2 Sanders C et al. Reporting on quality of life in randomised controlled trials: bibliographical study. BMJ 1998; 317(7168): 1191-94.