(916) 482-5200

Lies, Dammed Lies, And Statistics

Published: March 16, 2011

So goes the saying of what types of lies there are. While stats may not always represent falsehoods, they are always capable of being misleading. This was pointed out with great clarity by New Scientist magazine in an article worthy of examination from their 2/12/11 issue.

Rudy Giuliani, while running for president said that the chance of a man surviving prostate cancer in the US was 82% versus 44% in the UK, with its National Health Service. He meant to show that this proved the superiority of the American system of health care delivery. He meant to show it was twice as good. He was in fact citing a misleading stat.

Giuliani cited five-year survival rates in countries with different diagnosis methods. In the USA we screen for prostate cancer, in the UK they act on the basis of symptoms. Screening inevitably picks up earlier cases, but that may not translate into as big an advantage as we’d hope. Why?

Well, suppose a group of men with the disease all die at 70. If the men do not get prostate cancer until they are over 65 their 5 year survival rate will be 0%. If screening picks up cancer in all of these men at age 64 they will have a five-year survival rate of 100%. In this instance all the men have the same mortality despite different “survival rates”.

We presume that earlier detection via screening will improve the men’s odds since corrective measures can be taken. Some positives will be false, however. Worse, screening also picks up non-progressive forms of the cancer that will not lead to actual symptoms, let alone death from the disease. The real world extent of this over diagnosis is still not clear, but current estimates are that 48% of men diagnosed via screening do not have a progressive form of the disease!

Since a false diagnosis and an over diagnosis of cancer type both cause treatments that are unnecessary (and sometime cause significant harm) screening for prostate cancer is not producing the positive effect on mortality doctors had hoped.

This over diagnosis of more benign forms of disease also has the effect of inflating the five-year survival rates. It undoubtedly has done so in the US whether Rudi Giuliani is aware of the fact or not.
To accurately compare prostate cancer stats between the USA and the UK one must look at mortality rates.
According to the Cancer institutes of the two nations the respective rates for age-adjusted mortality are 24.7 per 100,000 in the USA vs 23.9 per 100,000 in the UK, which is a statistical dead-heat!

Giuliani’s stats may be correct, but obviously they do not tell the real story. Mortality rates of the two countries do NOT support the notion that thanks to the UK’s National Health Service and a reluctance to spend money on screening for prostate cancer that British citizens are dying on droves compared to Americans.

One take-away lesson: be wary of “survival rate” comparisons between nations as compared to mortality statistics. Such claims are currently being used to influence the health care debate here in the USA. They need to be examined closely.