Canada's newspapers get failing grade for coverage of new prescription drugs

August 1, 2003

It's hard to miss the explosion of media stories regarding miraculous new pharmaceuticals. But given that medicine can both help and harm, how reliable is the information that Canadians -- including patients and health care professionals -- receive from our country's mainstream newspapers?

Nobody in Canada had ever researched this question before, so no one really knew how well the media -- one of the most trusted and valued sources of information about health care -- was doing its job.

To remedy this situation, I set out with a team of researchers to systematically examine the accuracy, completeness and balance of reports in major Canadian newspapers on a sample of five new drugs. But before I tell you what we found, take this simple skill-testing question to measure your drug-reporting savvy:

* Would you take a cholesterol-lowering drug, that has minimal side effects, every day for five years in order to reduce your risk of having a heart attack or stroke by 42% -- Yes or No?

* Would you take a cholesterol-lowering drug, that has minimal side effects, every day for five years in order to reduce your risk of having a heart attack or stroke by 3.5% -- Yes or No?

Hmm. The "42%" sounds impressive. If I thought I was at risk for those diseases I would probably take the first drug. But the second? A "3.5 %" reduction in risk seems pretty insignificant to me for the trouble of swallowing a pill every day for five years.

If you answered 'yes' to the first drug and 'no' to the second -- as I did when I first saw those numbers -- then you've been bamboozled by statistics. That's because both questions refer to the same study of the same cholesterol-lowering drug simvastatin (trade name: Zocor). The results were just reported to you in two different ways.

The trick comes in understanding the distinction between a relative difference and an absolute difference. You see, in the clinical trial of this drug, 8.5%of people taking a placebo had a heart attack or stroke, compared to 5% of those taking Zocor. That's an absolute difference of 3.5% but a relative difference of 42%. (5 is 42% LESS than 8.5) Guess which number grabs the spotlight? Voila, welcome to the world of drug reporting.

I wanted to know how often this kind of statistical manipulation happened. Our team looked at all articles published in 24 major English and French newspapers during the year 2000 on five recently-launched drugs. We extracted 193 articles that mentioned one or more health effects relating to Celebrex for symptoms of arthritis; Lipitor, a cholesterol-lowering drug; Evista, for post-menopausal osteoporosis; Tamiflu, for the flu; and Aricept, to treat Alzheimer's disease.

We found, for starters, that only one quarter of the articles even mentioned how much a drug worked -- meaning, the articles didn't even tell us if the drug helped 3% of the people who used it, or 60%. Knowing how much better the drug works than a placebo is a pretty important piece of information. If a drug supposedly prevents heart attacks, strokes, or broken hips, how many of these events did it prevent?

One in three articles presented benefits using relative statistics -- the misleading 42% -- rather than more accurate absolute numbers. Overall, 68% of the articles did not mention a single side effect or harm related to the drug and when harms were mentioned they were usually placed towards the end the article.

Contraindications -- information on who shouldn't take a drug, such as pregnant women or the elderly -- were mentioned in only seven (4%) of the articles in our sample.

Most troubling, perhaps, was the overwhelming silence about funding links. Although pharmaceutical manufacturers fund most of the pre-marketing studies of new drugs, our study found that over 90% of the time, this link was not revealed. Paid spokespeople were often quoted, but financial or other ties to manufacturers were mentioned only 3% of the time.

Why does completeness and accuracy in drug reporting matter? Because drugs are an increasingly important part of health care. Collectively, Canadians now spend $15 billion per year on pharmaceuticals. A side effect of misleading media reporting on new drugs is that it can lead to much wider use of a drug than is warranted. Overly rosy reporting may lead to demands on doctors to prescribe the latest new drugs, or pressure on policy makers to cover them--even if they aren't any better than last year's batch of "wonder drugs."

The information people get from the media on new drugs should be balanced and accurate, and it should contribute to informed health care decisions. Journalists have a duty not to raise unrealistic expectations or to participate in disguised advertising. Because the news is such a trusted source of health information, journalists need to inject a higher dose of 'healthy skepticism' into their reporting.