Consumers and Health Professionals Find Risk Statistics More Persuasive When Presented in Relative Terms, Study Concludes

BUFFALO, N.Y. Every day, consumers and their healthcare providers make decisions about medications, diagnostic tests and other medical interventions based on risk statistics.

According to the findings of an international study published today, certain methods of presenting these risk statistics are more persuasive than others.

Elie A. Akl, MD, PhD, assistant professor of medicine, family medicine and social and preventive medicine at the University at Buffalo's School of Medicine and Biomedical Sciences and School of Public Health and Health Professions is first author on the paper that was published in the Cochrane Database of Systematic Reviews. Akl also holds a part-time appointment in the Department of Clinical Epidemiology and Biostatistics at McMaster University.

In the following example, Akl illustrates two different statistical presentations of the same information.

A trial compares the incidence of stroke in patients receiving a new drug versus those receiving a placebo; two percent of those taking the placebo suffered a stroke versus one percent of those taking the new drug.

"If you say that the risk of stroke was cut by half or by 50 per cent, this is called a relative risk reduction," Akl explains. "You could also say that the risk was reduced by one percent (2 percent in the placebo group minus 1 per cent in the treated group). This is called an absolute risk reduction."

The findings show that consumers, physicians and other providers were more likely to understand, and be persuaded to take action, when a specific risk about a health outcome was presented in relative, rather than absolute terms.

The research has major implications for the way that medical information is presented to healthcare providers, to patients and consumers and to policymakers.

"It's not about the statistics, it's about the way they are presented," says Akl, who also holds a position at McMaster University. "If we can find the best way to present risks, it can lead to improved health behaviors and outcomes."

The systematic study involved an exhaustive analysis of all relevant peer-reviewed research ever published on the subject covering 23,000 individual studies. Of those, 35 studies, including one published by Akl's group, were found to have direct relevance to the question of how best to present statistics about risk. The current research findings were based on analysis of these studies.

People perceived a larger risk and were therefore more persuaded to adopt an intervention when the statistics were presented in relative, versus absolute terms, Akl says.

There were no differences between healthcare providers and consumers, the study found.

"So physicians, too, are susceptible to being more persuaded by risk statistics that are presented as a relative risk reduction than when presented as absolute risk reduction," says Akl.

However, the researchers note that presenting statistics as relative values by themselves is also not advisable, since that may not provide sufficient bases on which to make a decision. They say there is a need for additional research on how the understanding of risk statistics influences actions taken by consumers and physicians.

"There is a need to educate health professionals about statistics and how to present them," says Akl, "but there is also a need for us to better educate consumers about statistics. People make decisions about risk every day, not just those concerning their health but concerning many areas, such as their finances or the environment. People need to better understand statistics."

Akl co-authored the study with: Holger Schunemann, McMaster University; Andy Oxman and Gunn Vist, Norwegian Knowledge Center for Health Services; Jeph Herrin, Yale University; Irene Terrenato, Italian National Cancer Institute; Cecilia Costiniuk, the Ottawa Hospital and Diana Blank, University of Toronto.

The work was funded by UB, the Italian National Cancer Institute, the Norwegian Research Council and Holger Schunemann.