How Money Affects Results in Breast Cancer Studies

Breast cancer studies funded by pharmaceutical companies are more likely to report positive, or beneficial, results than those funded by government or non-profit organizations, according to a new report in the Journal Cancer. That doesn't necessarily mean that the results are wrong, as Amanda Gardner points out in the Washington Post. You would expect drug companies to have a better track record in researching new drugs.

But the Cancer study also fits with previous research that shows that scientists in general are less likely to report (and journals are less likely to publish) negative results--in which a drug or other therapy didn't work--than positive ones. Now add the typical reluctance of a for-profit company to announce that a drug it's developing doesn't work, and you have a bias against studies that show negative results.

That doesn't mean that for-profit drug companies are bad. It just goes to show that you knowing who funded a particular study can be just as important to evaluating a study as understanding the science behind it.

More on Vitamins and Pregnancy

One of the hardest parts about writing health stories is figuring out what to leave out. Last Friday I wrote a post that focused on prenatal vitamins and the risk of pediatric cancer. Although the vitamins reduced the risk of some cancers by 50 percent, I calculated that that was not a huge change because the absolute risk of getting cancer in childhood is so low.

"But isn't taking vitamins during pregnancy a good thing?" a reader asked. Yes, absolutely. Taking folic acid, for example, which is found in prenatal vitamins and fortified bread and other cereals, has dramatically reduced the numbers of neural tube defects in newborn babies.

I should have acknowledged that fact in the last paragraph of Friday's post. My oversight serves as a reminder that whenever you read or hear health news, you should ask yourself what has been left out. Even the best health writers can't cover everything.

Vitamins Reduce Cancer Risk--Relatively Speaking

A new study shows that women who take multivitamins and folic acid during pregnancy reduce the chance that their baby will develop some childhood cancers by almost 50%. Your first question on reading this should be "50% of what?"

In other words, you want to know what the absolute risk of developing pediatric cancer is. Then you can figure out whether a 50% reduction (the so-called relative risk) is a lot or a little.

Fortunately, the overall risk of childhood cancer--such as leukemia or neuroblastoma--is very low. Joseph Hall of the Toronto Star, for example, characterizes the absolute risk of neuroblastoma this way: "the disease affects about one in every 6,500 children under five years in North America."

Do the math and you realize that he's talking about a 0.015% chance of developing neuroblastoma. A 50% reduction in relative risk means that the real-world chance of developing a neuroblastoma has dropped from 0.015% to less than 0.008%.

Such a slight decrease is not a bad thing--especially when you consider how easy it is to take vitamin supplements during pregnancy and how devastating neuroblastoma can be. But most health stories in the popular media focus on relative risk and ignore absolute risk, giving you an incomplete picture.

Treating Herpes May Prevent HIV Transmission

Everyone takes shortcuts--even medical researchers. That's not necessarily a bad thing. After all, where would we be if Christopher Columbus hadn't decided to take a shortcut to Asia? Hmm, maybe that's not the best example.

The point is that the best health-and-science journalists, like Alice Chang of the Associated Press, tell you right away when scientists are taking shortcuts.

Chang reported on a study that concluded that treating genital herpes might prevent transmission of the AIDS virus in people who are infected with both herpes and HIV. This makes sense since doctors already know that folks who are infected with herpes, or another sexually transmitted disease, are at greater risk of getting infected with HIV as well if they are exposed to the AIDS virus.

Near the very top of her article, Chang makes this important point:

In the latest study, conducted in Africa and published in Thursday's New England Journal of Medicine, women who took the herpes drug valacyclovir had less HIV in their blood and in their genital secretions.

The study did not look at whether the drug, sold as Valtrex by GlaxoSmithKline PLC, actually reduces transmission of the AIDS virus. However, scientists generally have found that the more virus someone has, the greater the risk of transmission.

In other words, the study investigators didn't measure whether the women in the study group actually passed on the virus less frequently than their untreated counterparts--that would have taken far too long and would have exposed more people to HIV. Rather the researchers measured the amount of HIV in the women's bloodstream and inferred that because the HIV levels dropped in the treated women, their risk of passing on HIV to someone else had fallen as well.

The researchers' assumption is reasonable but there is a chance it is mistaken. And that's why calling it to your attention, as Chang did, is a sign of good journalism.

Premature Babies and Happy Endings

Readers sometimes wonder whether journalists allow their politics or their advertisers to influence how the news is covered. But I've noticed a different bias over the years in coverage about extremely premature infants. Call it the bias in favor of a happy ending.

Take, for example, the opening sentences from Monday's article by Erika Beras in the Miami Herald under the headline "Miracle Baby Going Home to Homestead with Parents":

When she was born, she was 9 ½ inches long and weighed about 10 ounces. The doctors didn't give her much of a chance.

But Amillia Taylor is a fighter.

This week, Amillia, now 17 weeks old and weighing 4 pounds, drank from a baby bottle for the first time.

And today, she'll go home.

But in fact, the odds are stacked against a baby born this small. And whether we like it or not, little Amilla Taylor, whose conception occurred as a result of in vitro fertilization, has many medical problems that will likely plague her the rest of her life.

Who doesn't want a happy ending when it comes to little babies? But I've seen it happen so many times before. Initial enthusiasm over a very premature baby who weighed less than a pound at birth gives way to grim updates--if any appear at all--about the baby's health.

And indeed, I now see that doctors are not releasing Amilla from the hospital after all. She has, according to the (shorter) followup story in the Miami Herald, "a number of ailments" and "needs extra care."

Women and Heart Disease

You've got to understand risk if you want to stay healthy. Why do women fear breast cancer more than heart disease? As a group, women have a 4% chance of dying from breast cancer vs. a 38% chance of dying from heart disease. But because women are less likely than men to develop heart disease in their 40s, they (and their doctors) often assume they don't need to worry too much about cardiac problems.

So, the American Heart Association has come up with new guidelines on preventing heart disease that emphasize a woman's lifetime risk. The idea is for women to understand their real risks better and, hopefully, to motivate them to exercise more and adopt a healthier diet. Indeed, there's probably more you can do to prevent heart disease than to prevent breast cancer anyway.

Autism Gene Is Found--Or Is It?

Headlines are often major sources of error. The Canadian Press just ran a story under the banner "Autism Gene Found." Trouble is, if you read the CP story, or the scientific study in Nature Genetics on which it was based, it's clear that no single gene--or even collection of genes--for autism has been found. Instead, researchers scanned genetic samples from 1,400 carefully selected families with autism and discovered suspicious patterns in their DNA that suggested where they might look for genes that could potentially predispose a person to autism.

In other words, if families with austistic members share similar DNA patterns, then it's possible those patterns might tell you something about a genetic predisposition for autism.

Notice all the caveats and hedges in these statements? Researchers use associations--like the ones found in this autism study--as ways of figuring out where to look for interesting results. That doesn't guarantee they will find anything.

Just as there's no guarantee that just because you found a few football players who wear red jackets that all football players wear red jackets.

Or, as the Canadian Press article by Sheryl Ubelacker (2/19/2007) puts it:

What’s important about the discovery, said [autism researcher Stephen] Scherer, is that it has given scientists in the field a new foundation for further research, which eventually could lead them to specific mutations on specific genes that cause — likely in combination — the spectrum of disorders that fall under the autism umbrella.

Welcome!

We learn a lot about our health from the media--books, magazines, television, radio and the Internet. But you can't always trust everything you read, hear or see. I should know. I've written about health and medicine for more than 20 years--mostly at TIME magazine. This blog is aimed at evaluating health stories in the media, identifying the good, the bad and the ugly and helping you to be a better consumer of health information.