I've just commented on Kirstin Hendrickson's article in Scientific American (worth a look), Cesarean Sections in the U.S.: The Trouble with Assembling Evidence from Data, to say:
This article perhaps might have been subtitled, "The Trouble with Assembling Evidence from Mixed Cesarean Data".
One of the biggest problems in the ongoing cesarean debate is the number of research studies that rely on mixed cesarean data (i.e. they do not separate planned and emergency surgeries, or planned surgery 'for a medical/obstetrical reason' and planned surgery 'on maternal request' alone).
Similarly, there are too many research studies that do not properly compare intent to treat (i.e. a comparison of each planned mode of delivery and ALL the actual outcomes of each).
Going forward, we need better data collation, more standardized methods for comparison of different birth plans, and a willingness to include the longer term health and cost outcomes (e.g. incontinence and prolapse) of each birth plan.
There is a great deal of truth in the last paragraph of this article, and while the U.S. does might not yet have achieved the correct balance in terms of wanted and unwanted cesareans, it's worth remembering that in many other countries, the ideological focus on normalizing birth at all cost has resulted in tragic outcomes for some mothers who pleaded for a cesarean birth but were ignored.