profession
Many physicians misled by cancer statistics, study says
Posted March 12, 2012
Most doctors are confused about which statistics provide reliable evidence that cancer screening tests work, says a March 5 Annals of Internal Medicine study.
In an online survey of 412 primary care physicians, researchers analyzed respondents’ reactions to data on mortality rates versus five-year survival rates. They found that physicians were three times more likely to recommend a screening test supported by five-year survival statistics, even though those statistics are an unreliable measure of whether screening benefits patients.
Survival rate statistics are misleading because they are susceptible to lead-time and over-diagnosis biases, the study said (link). For example, if a patient diagnosed with cancer at age 67 dies at 70, his five-year survival rate is 0%. If that patient is diagnosed with cancer by a screening test at age 60, but still dies at 70, his five-year survival rate increases to 100%. Mortality rates are not subject to such biases.
Yet 69% of doctors recommended a test showing increased five-year survival rates, compared with 23% who recommended one that showed decreased mortality rates. About half of survey respondents also incorrectly said finding more cases of cancer due to screening proves that screening saves lives, the study said.
Note: This item originally appeared at http://www.ama-assn.org/amednews/2012/03/12/prbf0312.htm.
