I have just received notice of the newest campaign in the Detect Cancer Early strategy here in the UK. As an experienced GP with a strong interest in evidence based medicine I am looking for more information that demonstrates clearly that detecting cancer early actually significantly reduces death rates from these cancers as that will be the best way I can persuade my patients to adopt this strategy.

So, for example, evidence from the USA for prostate cancer shows that the number of diagnoses rose exponentially between 1975 and 1995 and although the numbers have fallen since then they are still 50% more than in 1975. Yet the death rate from prostate cancer has remained essentially static with only a slight fall in the last decade (well after the rise in diagnoses began - indeed after it had started to decline again - and probably due to newer treatments).

Similarly the evidence for lung cancer in the USA shows higher death rates in those in whom we went looking for cancer as almost all with "suspicious" lesion underwent very risky surgery from which a number died. This seems crazy when we have no idea how any of these newly discovered lesions would have behaved. The famous Mayo Lung Study showed a persisting long-term increase in deaths in the screened group.

You may well be aware of the breast screening controversy currently agitating people in the UK medical press but, again, it's fair to say that mammography screening per se has failed to demonstrate a significant impact on reducing deaths from breast cancer (as opposed to the benefits of newer treatments) - the number of new diagnosis between 1975 and 1995 in the USA rose by nearly 70% but the death rate was largely unaltered.

Similar cases can be made for melanoma (three fold increase in diagnoses, death rate unaltered) and thyroid cancer (two fold rise, death rate unaltered).

One of the stated aims of the programme is "to work with GPs to promote referral or investigation at the earliest reasonable opportunity for patients who may be showing a suspicion of cancer whilst making the most efficient use of NHS resources and avoiding adverse impact on access". But how, in practice, can these effectively mutually exclusive ideals be married?

If we look harder for cancer we will, of course, find more cancer but we have no way of knowing which lesions would have caused harm if left undiscovered. But having found it we are compelled to act thus exposing our patients to all the physical and psychological traumas of unnecessary investigations and treatment.

It has been argued that this programme isn't just about mortality, it's about improving length of survival. But, of course, if you find something earlier which is still going to kill you in the end your perceived "survival with cancer" will be longer.

I am not against screening as such and I fully support it where there is at least some reasonable evidence of benefit such as with colon or cervical cancer. But I need to see much more solid evidence before I can recommend that my patients sign up to "discovering cancer early".