oncologywatch

The blog is on hiatus, while moving to a new Linux-based server.

Waters EA, Sullivan HW, Nelson W, Hesse BW. What is my cancer risk? How internet-based cancer risk assessment tools communicate individualized risk estimates to the public: content analysis. J Med Internet Res. 2009 Jul 31;11(3):e33. 

We have all seen – and most of us have probably used – interactive Web-based tools designed to predict cancer risks. How good are these sites? Not very, if education is the goal, according to a recent survey by staff at the National Cancer Institute.

Risky Business
“Internet-based cancer risk assessment tools can provide cancer risk and prevention information to millions of people worldwide,” writes Erika Waters, a postdoctoral fellow in the NCI’s Cancer Prevention Fellowship Program, and her co-authors in their report, which was published recently in the Journal of Medical Internet Research. “However, poorly communicated risk information could mislead people or frighten them unduly, resulting in maladaptive health behaviors.”

The team of researchers, from the NCI’s Health Communication and Informatics Research and Basic and Biobehavioral Research groups, used popular search engines to locate more than a thousand websites dealing with cancer risk. They identified 44 unique interactive sites providing individualized cancer risk estimates. Nearly half of the sites were run by cancer centers and the health care industry (e.g., Blue Cross/Blue Shield, Merck); others were run by public health agencies (including the NCI) and a variety of private nonprofit groups.

Although they don’t report findings for individual websites, Waters and her co-authors find much room for improvement overall. Of particular concern, they observe that few of the sites communicate basic concepts of probability that are critical for an understanding of risk estimates. Few sites provide risk estimates in both numbers and words, or describe how a user’s cancer risk compares to other hazards.

They offer several recommendations to improve online risk-assessment tools, including the following:

• Describe the risk using both words and numbers. Words can be ambiguous; numbers can lack context.
• Communicate numeric risk as N in 1000 or as a percentage.
• Provide absolute and comparative (i.e., below or above average) risk information.
• Compare cancer risk to the risk of other hazards (e.g., being struck by lightening, being in a car accident) to place the risk in context.
• Frame the risk in positive and negative terms (eg, “Your risk of cancer is 5%”/ “This means you have a 95% chance of not getting cancer)
• Specifying whether the risk estimate is applicable to the next 5 years, 10 years, or over the user’s lifetime.
• Provide safety messages and risk reduction strategies.
• Include a visual display of risk, while avoiding biasing perceptions of risk.
• Acknowledge that the risk estimate contains an element of uncertainty, as it is based on statistical modeling of population-level data.

The study also references several review articles with additional recommendations on communicating risk in ways that help end users become better critical users of health information.

[OncologyWatch directory of open-access oncology journals.]

OncologyWatch: Posts about free-access articles on aspects of oncology theory, practice and policy (about the blogger). This blog is not a source for medical advice.

Technorati tags:
oncology cancer information

Comments and Links Appreciated!