ShareThisAs the old truism holds, "There are three kinds of lies: Lies, damned lies and statistics."
Never has that been more accurate, it seems, than when it comes to health statistics. Risk percentages, drug-benefit numbers and survival rates can be manipulated as deftly as a chiropractor cracking a back.
The multitude of studies people read about in the newspaper, see on TV commercials or receive via e-mail from the pharmaceutical industry sometimes can tell only half the story or, worst case, distort the results to such an extent that it might cause harm to patients.
Being fully educated about medical statistics can be lifesaving. What exactly is the survival rate for that new breast-cancer drug? Should men over a certain age undergo prostate surgery?
Much of it depends on the type of study, how the study was vetted and the population targeted in the study.
A cautionary tale would be the case of the heart-disease drug BiDil, the first medication given FDA approval exclusively for African-Americans in the United States. Researchers originally only included African-Americans in a small study and found their blood-pressure risk higher than whites'.
But a subsequent study of 85,000 subjects in 11 racially diverse countries found that diet, not race, was the main factor in high blood pressure. Richard Cooper of Loyola Chicago University found that the top five countries with elevated blood pressure were in Europe, led by Germany. The lowest: Nigeria.
To help consumers review health information more critically, three physicians and researchers (Steven Woolshin, Lisa Schwartz and H. Gilbert Welch) have written "Know Your Chances: Understanding Health Statistics" ($16.95, University of California Press, 158 pages). Here's a sampling of information in the book, a handy reference for patients faced with choices in medication or treatment.
KNOW THE CODE WORDS
Risk: "The chance that something will happen: for example, 'In this group of patients, the risk of heart attack is 10 percent.' "
Outcome: "The 'something' that may happen -- maybe death or maybe a medical event such as a heart attack or cancer diagnosis."
Statistics: "Numbers that summarize information -- based on observations of large groups of people and are useful in predicting what is likely to happen."
HOW RESEARCHERS DISTORT STATISTICS
"Highlighting the number of cases without mentioning the number of people at risk, to make the risk sound big."
Time frame: Be sure to look for fine print showing the chances of getting cancer over a "lifetime" compared with a year. Example: If a public-service ad says, "Colon cancer will strike 53 out 1,000 people," that is over a lifetime. In reality, colon cancer will strike 5 out of 10,000 people in one year and 5 out of 1,000 people in 10 years -- much less scary numbers.
"Getting cancer does not necessarily mean dying from cancer."
For perspective, try to reframe the statistic: When a news story says "Jones' chances of dying of colon cancer are 2 out of 1,000," turn it around and say, "Jones' chance of not dying is 998 out of 1,000."
Failing to tell how much a certain drug changes the risk of dying if you are among the people in the study who did not take the drug (the placebo control group).
KEY FACTS FOR ANALYZING STUDIES
"Be sure the study involved people similar to you -- not just in terms of age and sex but also in terms of their starting risk."
"Unless you know the starting risk (the 'lower than what' part), the message really tells you nothing."
Ask your doctor for complete information of a drug's effectiveness: the outcome over a time frame, the starting risk (the untreated group) and the modifying risk (the treated group).
TYPES OF OUTCOMES
Surrogate: These results mean lab measurements such as lower cholesterol levels or tumor shrinkage on an X-ray. "Judging benefit based solely on a surrogate outcome requires a big leap of faith."
Patient outcomes: "Fewer symptoms of disease, fewer deaths from disease ... (the) impact becomes increasingly direct. ... You clearly feel symptoms."
TYPES OF STUDIES
Animal or lab studies may not be relevant in people.
Observational uncontrolled studies, in which all patients get a drug, and then it's determined how many get better, are the weakest type of research. "You can't know what would have happened without the drug."
Controlled studies: Researchers watch what happens to different groups of people. (Example: the link between cigarette smoking and lung cancer.) "They can show that an intervention is associated with a particular outcome; they cannot by themselves prove the intervention causes the outcome."
Randomized controlled trials: "Researchers construct two groups similar in every way except whether or not they get the intervention being studied. You can have the most faith in statistics resulting from large, randomized, controlled trials."
SIDE EFFECTS
"Find out how often people taking the drug experience the side effects (listed in an ad). These numbers aren't as accessible as they should be."
"See how often (side effects) occur in the placebo group (the starting risk)."
SURVIVAL STATISTICS EXAGGERATION
"Earlier diagnosis (of cancer) always increases survival rates, but it doesn't necessarily mean that death is postponed." Example: Prostate cancer. Saying there's a 100 percent 10-year survival rate for men diagnosed with prostate cancer at age 57 due to screening is no different than the patient diagnosed at 67 because of symptoms. Both patients die at age 70.
Death rates are a better gauge. "If screening is finding only more non-progressive cancer, the death rate doesn't change."
(Sam McManis can be reached at smcmanis(at)sacbee.com.)
(Distributed by Scripps Howard News Service, www.scrippsnews.com.)
Must credit Sacramento Bee
