Sunday, November 27, 2011

A lesson in being wary of health articles in the news

I recently read a news article about a new study on the safety of home birth. Later, I wanted to find the article again. All I remembered was that it was from the Huffington Post, so I navigated back to that site, and what did I see in the pregnancy section? I saw this unintentionally hilarious juxtaposition of articles:

They really are right next to each other like that. See for yourself here (at least until new stories push it off the front page).

One of these articles is about how home birth is basically safe. The other is about how it's really risky for first-time mothers.

The best part? Both articles are based on the exact same study - this one that was recently published in the British Journal of Medicine.

And it's not just that each article is highlighting a different aspect of the study. They each pull out different figures to talk about the risk to first-time mothers to make it sound more or less safe, according to the article's mission.

The pro-home birth article says: "... researchers found a higher risk for first-time mothers planning a home birth. Among those women, there were 9.3 adverse events per 1,000 births, including babies with brain damage due to labor problems and stillbirth. That compared to 5.3 adverse events per 1,000 births for those planning a hospital birth."

The second article, the one highlighting risk, says: "Serious adverse outcomes for the baby are rare - occurring just 3.5 times for every 1,000 babies whose birth was planned in an obstetric unit. But the research, carried out at Oxford University, shows this figure rises to 9.5 per 1,000 babies if the mother chooses a home birth."

These articles are actually using two different figures for the risk of adverse outcomes for babies born in a hospital! So what's the truth? Is a home birth, according to this study, less than twice as dangerous to the baby, or almost three times as dangerous? Because one article says one thing, and the other article says another.

Fortunately, the BMJ study is available online, so we can find out. First of all, a summary website for the study uses the 5.3 vs. 9.5 per 1000 figure. Doing a quick skim and a text search of the article itself, I couldn't even find that 3.5 adverse outcomes per 1000 births figure, but I did find a statement that the study overall had 4.3 adverse outcomes per 1000 births. That number includes both first-time mothers and mothers who have had children before. Knowing that, I'm guessing that the 3.5 per 1000 figure must be multiparas (women who have had at least one child before).

Thus, it seems clear that the second article cherry-picked its numbers, and compares multiparas giving birth in the hospital to nulliparas giving birth at home in order to make the risk sound greater than it is. On the other hand, the first article definitely downplays the risk and glosses over it.

This highlights the importance of always being wary when we read health news in popular media. Authors often misinterpret, misrepresent, and misconstrue the data, and almost everyone has some bias toward one side or the other. This may be deliberate, unconscious, or accidental, but it's still very common.

Sorry for the funky formatting in this post. I messed it up and can't figure out how to fix it.

1 comment: