15 November 2019

Gender bias in clinical case reports

Some of the best books are the ones that tilt your perspective completely. I’ve been working my way through Invisible Women: Exposing Data Bias in a World Designed for Men°. I haven’t yet finished but it is a tremendous read and it has the potential, rather remarkably, to change the way you look at the world. The bias towards men in our society, in all the tiny details, is completely baked in, making it astonishingly difficult to see, never mind take action against it.

At the time of writing, I haven’t got to the medical sections so there is the possibility this is covered in the book. My brain was ticking over as I listened to the audiobook and I had the brilliant idea to check if there was a gender bias in clinical case reports in the medical literature. And, of course, because it is a good idea to check that, it has already been done. Not all that long ago either and the full results were published in PLoS ONE in 2017.° They looked at 2742 case reports in the big five’ medical journals. Here is the all important forest plot.

It’s obvious on this. A clear bias towards men.

The authors also ran the numbers again to make sure that this wasn’t, in their words, a NEJM problem”. It wasn’t. The pooled results remained significantly biased towards men. There is no reasonable explanation for this other than bias. Women are not seen less by the medical profession, indeed, without rooting out the evidence, I’d venture to suggest it is likely to be the contrary. Ignoring any stereotypes, there is the simple nature of pregnancy and childbirth that puts women in contact with services. The study results are, as Caroline Criado Perez shows in myriad ways, just another bias towards the male.

I would add that I’ve not found the book completely plain sailing. (I always wonder if this shouldn’t be plane sailing. Turns out it used to be - more on that here°. But I digress…) So far, I have found that some of the points made by Criado Perez have tortured the data beyond its reasonable conclusion. I want to shout: stop, stop, you’ve made the point, now you are actually weakening your argument! I have then, in a liberal angst tailspin, worried this was nothing more than my own implicit bias manifesting itself. Perhaps I just didn’t like the message and it was causing me to find reasons to nitpick. It was a bit of a relief when I was chatting to a female GP about the book, an undoubted expert when it comes to evidence with a hawk-eye for injustice and inequality, and she raised, unprompted, the same concern. Not just me then. There is hope yet that I’m not ageing into a reactionary misogynist.

None of that changes the underlying message - the biases are all still there. As the PLoS ONE study shows the male bias can be found in unpromising and unexpected areas. One just wouldn’t assume there should be a bias there. It is difficult to imagine why men should be more interesting and therefore more likely to appear in case reports. As an editor, it is a good lesson in being vigilant for gender bias.


Previous post
Mobile drug consumption rooms What I love about this research paper is that I simply hadn’t heard of mobile consumptions rooms before I found it.° It cheered me up to find out
Next post
Cormac McCarthy on academic writing [dropcap]I[/dropcap]’m not much of a fan of Cormac McCarthy. Not yet anyway, but I hold out hope. I’ve found it difficult to get into his novels but