Morning all, I am curious as to whether the concept of confounding applies in the following situation: Specifically I want to know if "taking antimalarials" is a confounding variable.

The exposure being investigated is "having sprayed insecticide in the house in the last 12 months", and the outcome is "infected with malaria".

Also, would it make a difference if the drugs were taken for preventative purposes rather than curative?

Thanks T

  • $\begingroup$ To determine whether "drug" ("taking antimalarials") is confounding the relationship between "sprayed" and "infected", first you would have to investigate empirical evidence or your intuition regarding the relationship of "drug" with either of the two. Next, consult the back-door criterion. If you, like me, believe that "drug" does not affect "sprayed" but does affect "infected", then it is a collider and you should not control for it. $\endgroup$
    – vkehayas
    Jan 30 at 13:05
  • $\begingroup$ Hi, thanks for your response. I believe "drug" is associated with "sprayed", yes, as both are preventative measures (a prophylactic drug, that is), but not a causal association, no. Would that make it a collider, then? I am still working on the DAG - getting myself a little confused with this back-door criterion. $\endgroup$ Jan 30 at 17:12
  • $\begingroup$ It appears that you have to clarify the relationship between "drug" and "sprayed" in your head. If one affects the other --either directly or indirectly-- and you want to examine the relationship between "sprayed" and "infected", then you need to control for "drug". $\endgroup$
    – vkehayas
    Jan 31 at 8:18
  • $\begingroup$ Hi vkehayas, thanks for coming back to me on this. Yeah, the relationship is somewhat confusing. As a prophylactic drug, such as is used in an MDA campaign, the association is really just that, an association, not one causing the other. It just so happens, that a person who sprays their house is more likely to also participate in the MDA campaign, and so take the drugs. As a curative drug, however, getting malaria might induce persons to 1) take curative drugs, and 2) start spraying their house - but here they are both caused by malaria. $\endgroup$ Jan 31 at 11:00
  • $\begingroup$ But, not spraying could cause malaria, which in turn causes someone to take the curative drugs. But, here, I believe curative drugs are an effect, which rules it out as a confounder. $\endgroup$ Jan 31 at 11:01


You must log in to answer this question.

Browse other questions tagged .