Thursday, October 8, 2009

Unreliable-reliable information: case

The other day I was in the teaching room and I was putting out some of the new brochures I ordered on the topic of contraception. We have a little display in there so the residents can grab something quickly when I'm not at the clinic.

Information about choosing a contraceptive method is one of the most commonly requested handouts and we keep running out. I found what I thought was a really good one produced by the Society of Obstetricians and Gynecologists of Canada (SOGC). You can't get more reliable and authoritative than that, I figured. I still think that... but....

One of the supervisors opened one up- they look like little brochures but actually they are little posters folded up nicely. When you unfold them you get a comparison chart at the top that shows the number of unintended pregnancies for all the different methods, from Intrauterine system (IUS), through Oral contraception, Condoms, Diaphragm etc., all the way to No contraception (no surprise, pretty high failure rate here: 850 out of 1000, or 85%).

It struck the doctor as odd that Intrauterine device (IUD) had such a high failure rate compared to the others (9 out of 1000 with perfect use vs 3 out of 1000 for the pill for example). It did seem a bit odd. Also for typical use there was a little "-" instead of a number, implying what I'm not sure. No data? Zero failure rate? An asterisk next to the 9 brings you to a little footnote: "This perfect use failure rate corresponds to the Nova-T 200. The typical use failure rate for the Nova-T 200 is likely slightly superior." So then shouldn't there be a number there instead of "-", maybe a higher number than 9?

I decided to check out the references listed right below the chart: Trussell 2007, Trussell 2004, Black 2004* and Andersson 1994.

The paper that seems to have provided the data for the chart is Trussell 2007, or complete reference:

Reducing unintended pregnancy in the United States
Contraception, Volume 77, Issue 1, Page 1
J. Trussell, L. Wynn doi:10.1016/j.contraception.2007.09.001

I am still a bit confused. The paper has a table that shows numbers that correspond exactly to all the other data in the chart except for under IUD we have:

Contraceptive method Typical use Perfect use
ParaGard (copper T) 0.8 0.6
Mirena (LNG-IUS) 0.2 0.2

So where is the Nova-T 200? And how come the above numbers aren't used? I did not find Nova-T 200 data in the other papers either. Provided I found the right ones.**

I still think the handout is great. It provides much more than just data. Aside from the chart there is information about all the contraceptive methods, what they are, how they work, advantages, disadvantages, stuff that will certainly help people make informed choices. And if they have any questions they can discuss with their GP.

However it is interesting to see that even information provided by the most authoritative sources can be flawed. Considering all the shocking debacles we have seen recently, Elsevier's fake journals, authors ghostwriting for pharmas, this is just one more reason to start assessing the contents of individual resources, and not rely so much on authorship.

* These are the Canadian contraception guidelines:

** Once again, is it too much to ask that people provide the full reference?
Some authors are prolific. It took me a while to figure out which Trussell 2007, and which Trussell 2004 etc. Still not sure about all of them. It shouldn't be so hard.

No comments: