Background: Many members of the general public use the internet as a source of health information, including using website calculators to assess their risk of developing serious illnesses. The risk information provided by web risk calculators (WRC) may impact users' worry about their risk, their health behavior and medical decisions. The widespread use of these sites and the potential impact of the information they provide substantiates the importance of sites providing accurate risk estimates.

Purpose: We reviewed WRCs for 3 common cancers – breast, colon, and prostate – to examine what risk factors were used by each to calculate risk, how those risk factors were quantified and whether references for the calculations were provided.

Methods: We conducted a search of “cancer risk” in 5 popular search engines and visited the first 1,000 hits from each engine to create a list of breast, colon and prostate WRCs. For each WRC, we reviewed source information and the risk factors incorporated in the calculation of an individuals' risk for the cancer.

Results: Web sites with risk calculators numbered 11 for breast cancer, 3 for colon cancer, and 4 for prostate cancer. No source for the calculator appeared on the websites for 73% of breast cancer calculators, 67% of colon calculators and 75% of prostate calculators. The calculators varied widely in the factors collected for risk assessment. For instance, 82% of the breast cancer calculators included age, 55% included race, 45% included alcohol consumption and menopausal status, 36% included gender, height, weight, hormone use, 27% included personal cancer history and fruit/vegetable consumption, 18% included Jewish ancestry, 9% include tobacco use history and current use and amount of exercise.

Conclusions: Few WRCs provide source information to substantiate the risk estimates that they provide. The wide variation in the risk factors that WRCs collect to calculate cancer risk raises additional concerns about the accuracy of the risk estimates that they provide. Inaccurate risk estimates can have significant consequences for recipients, including provoking unwarranted distress and encouraging unnecessary preventive measures (which may carry their own risks), or conversely, providing unfounded reassurance and discouraging appropriate preventive measures. Given that websites are likely to continue as a source of public information, improving the accuracy of their content may improve public knowledge and encourage appropriate prevention behavior.