Epidemiology

[bwbddb]

---

Members don't see this ad.

While studying, I ran across a practice question that puzzled me because the answer and explanation seemed contradictory to the fundamental teaching I recieved:

From Blueprints Clinical Cases: Obstetrics and Gynecology (2006) pg 132. If a physician were to design a study to determine if labor could be started using accupuncture, would he use a case control or cohort study ?

The "correct" answer was case control. But my understanding of case control is that cases are chosen based on disease/condition (labor) and therefore a rare exposure (accupunture) would be difficult to study. Wouldn't a cohort be better?

Check out the question on Google Books:
http://books.google.com/books?id=HVnyH5QLuBkC&pg=PR2&dq question 23-3 on page 132, answer on page 134

What are your thoughts?

 

[Rendar5]

I'm surprised too honestly. the reason to use case control is for either stuff that's rare (although you could just make that group a cohort) or for stuff that is unethical to randomize (like the SIDS studies). The biggest drawback of a retrospective case control is that it cannot prove cause-effect because there is a high confounding risk and also absolutely no way to establish a time-relation.

Perhaps blueprints is considering the case of a cohort where there is no control or randomization (just watch what people decide to do and record the outcomes), in which case a case control would be superior. But if you're considering an RCT as a type of cohort study, that would be far superior.

btw, the link didn't contain page 132/134