Roanoke Times
Copyright (c) 1995, Landmark Communications, Inc.
DATE: TUESDAY, March 2, 1993 TAG: 9303020117
SECTION: EXTRA PAGE: 1 EDITION: METRO
SOURCE: Jane Brody
DATELINE: LENGTH: Long
DON'T PANIC; STUDY THOSE SCARE STUDIES
The steady stream of well-publicized reports ascribing horrendous health risks to nearly every aspect of modern life, from electric blankets and cellular phones to post-menopausal hormones and now vasectomy, can fill otherwise rational people with feelings of panic and paranoia.Such reports can leave people longing for the "simple" life of bygone days, when - keep in mind - life ended decades earlier and polio, rheumatic fever and tuberculosis claimed more lives than cancer.
Most people see or hear only the headline news when a purported new health threat looms, like the recent report of two large new studies linking vasectomy to an increased risk of developing cancer of the prostate or the report several weeks ago linking the use of cellular telephones to brain cancer.
In years past, panic-button-pushing reports have included a link of hair dyes to cancer; coffee to JANE BRODY heart disease, and menopausal hormones and alcohol consumption to breast cancer, among many others.
In the weeks after such reports, there is typically a dip in sales or a change in habits, eventually succeeded by a return to normal when memory of the purported danger is obscured by practical considerations or personal preferences.
Untrained in the scientific method, let alone statistics, the average person understandably has great difficulty interpreting these findings, assessing their validity, recognizing their potential flaws and determining whether a behavioral change is warranted by the evidence.
Even the researchers themselves sometimes fail to see or mention caveats that may diminish the significance of their observations.
Still, there are some guidelines that can help everyone better understand the significance of reported findings and avoid frenetic and pointless changes in habits.
A link between two phenomena does not imply cause and effect. Correlations and relative risks are merely clues to possible cause-and-effect relationships.
For example, there is a well-known association between the number of television aerials in a region and death rates from heart attacks. This does not mean television causes heart attacks, but it might mean that people who watch a lot of television get little or no exercise and eat fattier foods, both of which can increase their coronary risk.
And just because consumers of artificial sweeteners tend to be heavier than users of sugar does not mean low-calorie sweeteners cause obesity. More likely it means that overweight people try to trim calories wherever they can.
A very large early study found that users of the birth control pill were three times as likely to get cervical cancer as were women using the diaphragm. But this association did not prove that the pill causes cancer.
It more likely showed that the diaphragm helped to prevent this cancer, which has long been known to spread like a sexually transmitted disease and to involve a virus.
With the new study linking vasectomy to an increased risk of developing prostate cancer years later, possible factors that could "explain away" the association might include differences in the men's sexual habits before or after their vasectomies.
Other factors could include differences in the amount of medical attention the men receive, which could result in a higher cancer detection rate among vasectomized men or other behavioral factors no one has yet thought of as possibly associated with prostate cancer.
Since the biological causes of this cancer are not yet known, it is not possible for researchers to take them into account when they do studies like those reported last week.
Establishing cause and effect is often a scientific challenge when studying people, since it is unethical to expose a group of people to a suspected hazard deliberately, to see if it indeed harms them.
Also, it can take decades of study for the hazard to be unequivocally established. This is why the tobacco companies can argue that science has yet to prove that cigarette smoking causes cancer in people.
Short of iron-clad proof, then, what can you look for in deciding how much influence a study finding should have on your life?
The source of the report matters. Whether the study was done at Harvard or Podunk University, the most reliable reports are those that are published in peer-reviewed journals, like The New England Journal of Medicine or The Journal of the American Medical Association. If the study is of great import, an editorial should accompany it, discussing its strengths and weaknesses. Unpublished reports that come out of medical meetings or news conferences are generally least reliable. Also less reliable are studies done at private research institutes that do not receive government grants, since there may be good reason why taxpayer money is not awarded to them.
Find out if more than one study reached the same conclusion. It is rarely wise to make radical changes based on a single study. Scientific research demands replication by independent researchers, since findings, even when "statistically significant," can arise by chance alone. When an association is repeatedly and consistently found in well-designed studies, like the link between smoking and cancer, then there is reason to believe it.
Consider the size of the risk involved as well as benefits that may result from the suggested cause. A tripling (300 percent increase) of a very small risk, say from 1 percent to 3 percent, is less important than a 50 percent increase in an already large risk. But even in the latter case, the risk may be worth taking if the benefits outweigh them. Even if, for example, hormone replacement after menopause increases a woman's risk of breast cancer by 50 percent, the therapy also reduces her risk of osteoporosis and heart disease, which is still by far the leading killer of older women, and may significantly improve her quality of life.
Weigh the degree of danger and whether there are safer and equally acceptable alternatives. For example, if the new vasectomy studies are correct, the risk of developing prostate cancer is less than doubled by the procedure. The overwhelming majority of men who get this slow-growing cancer have not had a vasectomy, so clearly other factors are more important. Also, the cancer is now readily detected in its early and highly curable stages by a $50 blood test, the PSA.
Also note whether the finding is supported by animal studies and if there is any biological explanation for the observed association. If not, a chance discovery of a potentially important health risk, like prostate cancer after vasectomy, should send researchers scurrying to the laboratory to look for supporting or explanatory evidence.
Jane Brody writes about health issues for The New York Times.