This website is part of the ECDC (European Centre for Disease Prevention and Control) network

Helping doctors and patients make sense of health statistics

17 Dec 2010

Gerd Gigerenzer, Wolfgang Gaissmaier, Elke Kurz-Milcke, Lisa M. Schwartz, and Steven Woloshin; Max Planck Institute for Human Development, Berlin; Harding Center for Risk Literacy, Berlin; The Dartmouth Institute for Health Policy and Clinical Practice’s Center for Medicine and the Media, Dartmouth Medical School

Because so many of health decisions and behaviour changes are made based on numbers, risks, hazards and probabilities, an informed shared decision making in health communication should rely not only on a transparent doctor-patient relationship, but also on the precondition of a common minimal literacy in health statistics. This monograph by Gerd Gigerenzer, Wolfgang Gaissmaier, Elke Kurz-Milcke, Lisa M Schwartz and Steven Woloshin explores possible causes and proposes solutions to tackle this public health issue. The article advocates that collective statistical illiteracy may affect risk communication and endanger health outcomes.

“Think of a woman who has just received a positive screening mammogram”, the authors propose. She then asks her doctor what are the chances of having breast cancer for certain. “One would assume every physician knows the answer. Is that so?” According to the data retrieved on a 2007 training session with 160 gynaecologists, the majority of them overestimated the probability of cancer between 1% and 90%. The number of physicians that provided the best answer was disconcertingly small (21%). A similar situation occurred when authors visited public health care centres and questioned 20 AIDS Counsellors about the outcomes of an HIV test. Thirteen said they were absolutely sure there are no false positives, although the chance of it happening can go up to 50% among heterosexual men with low-risk behaviour.

Take HIV screening, fingerprints and DNA tests: none of these are 100% certain, even though the majority of citizens think they are. According to the authors, people have to understand that there are no such things as certainty and no zero-risk; only risks that are more or less acceptable. And, together with the benefits, they come with some harms: costs, inconvenience and overdiagnosis, such as the detection of pseudodiseases that would never progress in the patient’s lifetime.

Some public health campaigns, such as pro-screening campaigns that use the term “cancer prevention”, may even mislead people into thinking early detection means prevention. Studies quoted show that women in Switzerland (10%), the United Kingdom (17%), the United States (26%) and Italy (33%) share the illusion of certainty that screening would prevent cancer. What screening actually does, state the authors, it is intended to detect cancer at early stage, , not reduce the risk of developing one.

A transparent representation of information is favoured in the monograph, because numbers offer a more accurate perception of risk than words (such as “unlikely”). The authors also question the absence of probability thinking in mathematics classes taught in school, which encompasses mostly algebra, geometry and trigonometry. “If children learned to deal with an uncertain world in a playful way, much of collective statistical illiteracy would be history”, they put forward.

 Read the article

ECDC comment: The article addresses the widespread inability to understand the meaning of numbers and emphasizes the challenge of learning how to live with uncertainty. What does it mean to have a high survival rate? If someone tests positive on a screening test, what are the possibilities of the person having the disease? These are examples of the health communication challenges faced by patients, physicians, health care providers and others professionals that have to deal with health statistics.

Understanding health statistics is a part of the collection of competences required to navigate through health information and it is particularly important in risk communication and the decision making process. Building statistical literacy should take a more relevant role on health education and it should be taught as a problem-solving technique. It is essential that the statistical knowledge acquired in schools matches the needs encountered in the day to day reality.

Do have any comments? Please email us at

More research on health communication:
Scientific advances in the area of health communication
Public health developments in the area of health communication

© European Centre for Disease Prevention and Control (ECDC) 2005 - 2015