For Immediate Release: 4 P.M. ET February 5, 2002
Contact: DMS Communications (603) 650-1492

Print Version

Dartmouth Research Examines the Value of Cancer Screening

HANOVER, NH - As people consider the merits or drawbacks of cancer screening, a Dartmouth Medical School study weighs in with some new observations, based on a statistical analysis of past trials, that may help put cancer screening in better perspective.

The conventional way deaths were classified may have caused misclassifications that biased study results in favor of screening, Dartmouth researchers demonstrated. They suggest an additional method of tallying all deaths to help avoid the misinterpretations that can lead investigators to overestimate or underestimate the value of cancer screening.

The findings are reported in the Feb. 6 issue of the Journal of the National Cancer Institute by Dartmouth Medical School professors William C. Black, MD, of radiology and of community and family medicine, and H. Gilbert Welch, MD, of medicine and of community and family medicine, and former medical resident David Haggstrom, MD.

Classifying the cause of death by specific disease is the most widely accepted procedure in randomized trials that assess cancer screening. However, two biases--sticky-diagnosis bias and slippery-linkage bias--affect such classification and can alter the assessment of screening value, the researchers found.

The validity of disease-specific mortality assumes that the cause of death can be accurately determined. An alternative end point, all-cause mortality, depends only on an accurate determination of deaths and when they occur; therefore it is unaffected by misclassifications in the cause of death.

People making decisions about screening want to have pertinent information about what it means for them, explained Black, a member of a national expert panel that assesses cancer evidence. He uses the shark analogy popular among his peers. Instructions and aids to protect yourself from a shark attack are meaningless if you don't go in the water.

Similarly, people have to understand how likely they are to be at risk for certain cancers when they decide to be screened for them. "They should be asking their physicians if this screening intervention is likely to increase their life expectancy," Black says. And their physicians hope screening studies take as much information as possible into account.

He and his colleagues compared the two mortality groups in 12 randomized studies of cancer screening for which both disease-specific and all-cause mortality could be determined. These trials involved screening for cancer of the breast, colon or lung.

In five of the 12 trials, the two mortality end points suggested opposite effects of screening. The researchers attributed these discrepancies to the two forms of bias that affect cause of death classifications.

In one form, called sticky-diagnosis bias, deaths from other causes in the screened group are falsely attributed to cancer because that cancer was detected by screening. This type of misclassification influences the disease-specific mortality results against screening.

In the second form, called slippery-linkage bias, deaths from the screening process or subsequent treatment are falsely attributed to other causes. For example, if an invasive evaluation causes a patient to have a fatal heart attack, the death may be attributed to a heart attack rather than to the disease being tested for. This misclassification tilts the disease-specific mortality results in favor of screening.

Both forms of bias affected the randomized screening trials, according to the analysis, but the Dartmouth researchers argue that slippery-linkage bias had a larger effect. The concept of "slippery linkage" has been hinted at before but never previously defined, notes Black, who says he and his colleagues are among the first to investigate the impact of this bias.

Integrating both types of mortality classification can help avoid flaws in screening assessment, according to the researchers. They conclude that all-cause mortality should always be analyzed and reported along with disease-specific mortality to ensure that major harms or benefits of screening are not missed due to misclassification in the cause of death. "All-cause mortality also puts the magnitude of expected benefit from screening into an appropriate perspective for prospective decision making," they say.

For additional information, contact William Black at (603) 650-5846 or by e-mail: William.Black@dartmouth.edu.

Return to News Releases