Media Over-Reports Cancer Studies

Alex B. Berezow

The media notoriously over-report epidemiology studies. One day, coffee is bad for you; the next day, it's the fountain of youth. This isn't entirely the media's fault; epidemiology studies, by their very nature, can be tricky and contradictory.

A new study has now possibly linked sexual orientation with cancer. The study was conducted by survey, which is not the strongest design. Participants were asked to report their sexual orientation, whether or not they ever had cancer, and the current state of their health. The study concluded NO difference in cancer prevalence between straight or gay women. However, gay men were 1.9 times as likely to have received a cancer diagnosis as straight men.

According to Reuters, the authors suspect that women also have a higher risk of cancer (even though their own data shows the exact opposite). The authors also appropriately warn us not to generalize the results of the study.

That's good, because studies like this should not be reported by the mainstream media. Why? Because it's really hard to draw firm health conclusions from studies based on surveys.

For instance, what if a particular group of people is more likely to go to the doctor than another group? That would skew the results. One of the authors contends that gay people are LESS likely to go to the doctor. But, that would seem to contradict the data. In this study, 8% of gay men had received a cancer diagnosis, but only 5% of straight men had. If gay men were indeed less likely to go to the doctor, then wouldn't one expect to see fewer cancer diagnoses among gay men, not more?

And, if sexual orientation is linked to cancer, then why is there no difference in the prevalence of cancer between straight and gay women?

Studies like this are very necessary, but are only preliminary. Until follow-up studies can confirm (or refute) the results, the media would be well-advised to avoid them. Alex B. Berezow is the editor of RealClearScience. He holds a Ph.D. in microbiology.

  • Share