Could the cancer numbers be better explained?

Cristina Roure explained in a post in her blog series "Pantone", an experience that Gerd Gigerenzer, Director of the Max Planck Institute for Human Development in Berlin, wrote in the book "How to know when numbers deceive you", where more than half of the surveyed gynaecologists did not estimate, after seeing published scientific data, what was the probability to have cancer of an asymptomatic woman’s with a positive screening mammogram. In fact the error of most of the colleagues was monumental: they said it was 90% when in fact it was 10%.

In a recent post, I commented on the book of Peter Ubel "Critical Decisions", in which the author showed what the emotional force of numbers was, both for patients and for the doctors. In this certainly worrying line, I want to give you a shred of hope. There are ways to show the numbers in order to reduce the reigning confusion. Many experts are warning us to flee percentages, the ineffable pies and the relative risks, and we have to get used to talking to patients of real numbers, pictograms with people and absolute risks. To make myself understood, I include a couple of graphics from the Harding Center for Risk Literacy, a centre which is in the orbit of Gigerenzer and Max Planck. The first is a pictogram on the evaluation of prevention programs for prostate cancer based on PSA controls and anal touch of the gland.

You will agree with me that the pictogram is expressed clearly enough and that its graphics achieve a very timely population vision. At a glance, or with a little help, anyone can see that screening for prostate cancer leads to more problems than benefits.

The second graph shows the evaluation of breast cancer screening by mammography. Note that, in the table from Harding Center, the preventive program, in 10 years reduces the number of dead women from 5 to 4 per thousand. The relative reduction is of 20%, but this, although true, is misleading. The numbers put in context imply that the population benefiting from the program is one less death for every thousand women, a reduction that is certainly diluted when the total number of women killed by cancer is analyzed. But what surely women from the preventive program are not registering is that there will be 100 (per thousand) who will suffer from a false positive and that 5 shall suffer an unnecessary intervention.

Considering this data, the lack of clarification from the official programs of breast cancer prevention that even today continue to make personalized calls to women aged over 50 without offering any explanation about the real benefits and the side effects of these programs is beyond our understanding.


Jordi Varela

Editor