Monday, 26 September 2016

When the overdiagnosis is politicized

Rudy Giuliani, the Republican mayor of New York, on that fateful September 11, at an election meeting said: "I had prostate cancer 5 or 6 years ago and I thank God to be an American citizen because here the probability of survival at five years for this cancer is 82%, while in the UK, due to socialized medicine, this value is only 44%." We must clarify that in the US, PSA screening is widespread while in the UK it’s not so. According to Gerd Gigerenzer (I have extracted this case from his book "Risk Savvy"); Giuliani’s words conceal a big mistake, because in reality, although it seems a contradiction, mortality from prostate cancer in the two countries is practically the same. So how is it possible that the survival rates are so different? To explain it, Gigerenzer describes two biases that encourage the intentional error of the conservative politician:

Advance Bias (lead time bias)

Imagine that a British man is diagnosed with prostate cancer at the age of 67 because of the symptoms of the disease and dies three years later at age of 70. His survival after 5 years of diagnosis is 0%. On the other hand, in a similar case, an American man is diagnosed early at the age of 60, thanks to a screening program, but also dies at 70 years of age. His survival after 5 years will have been 100%, but in reality he has lived as many years as the British man, with the difference that he had to endure treatment for 10 years, whereas the British only had to go through treatment for 3 years.

The Overdiagnosis bias

Now imagine 1,000 British men diagnosed with prostate cancer (without screening) with a 5-year survival rate of 44% (the upper part of the chart). In exchange, in the US (bottom of the chart) for the same 1,000 cases of cancer, another 2,000 have emerged these corresponding to the concept of non-progressive cancer, which means that cancer cells have been found, but these do not evolve or will not grow with sufficient speed as to cause the death of the person, and therefore, we will have 2,000 cases that will affect the calculation of the survival rate to up to 81% which Giuliani stated in his political discourse. But note that for both groups, the number of deaths after 5 years of diagnosis is the same: 560. The figures therefore are very clear, the PSA screening does nothing more than to alert a huge number of men who would have lived with their cancer but that would not have died from it.

Gigerenzer’s response against manipulation: icon-boxes

The communicative formula against the manipulation of information proposed by Professor Gigerenzer is very simple: always display the data in absolute terms and represent them with icons. And as shown, notice how useful the icon-box for prostate cancer screening is. In the column on the left of the graph 100 men over 50 years of age without screening are represented. What has happened to them after 10 years? Well, 20 have died due to many causes, including one who died due to prostate cancer. As for those from the screening program, you’ll notice that the number of global deaths will be the same, from which one will also be due to prostate cancer, but in exchange the side effects of the program are as follows: 2 men will go through unnecessary treatment (overdiagnosis) and 18 will be falsely alerted after being subjected to the terrifying biopsy. In short, the equality in the number of deaths proves that prostate cancer screening encourages inappropriate clinical performances.

I invite you to go to the Harding Center for Risk Literacy website and find the icon-boxes or facts-boxes in the right column. You will find several in open and you will see how useful these are to communicate and get people to understand the risks of certain clinical decisions involving them.

Jordi Varela

No comments:

Post a Comment