Why Observational Studies shouldn’t be used to assess Respiratory Virus Interventions - Part 3

A response to several helpful comments.

We are responding to several points raised by our previous post on this topic. They are interesting and require a response which is more detailed than commentary space allows and makes readable.

Observational studies do have a role in answering specific questions, and there's a long history of using case-control methods, for instance, to inform decisions. 

Some examples are the nested case controls in the Oxford Childhood Cancer Survey and the National  Encephalopathy Study in the 1970s, testing the association between pertussis vaccine exposure and rare neurological outcomes.

They can be powerful designs for answering appropriate questions. The Oxford CEBM: Levels of Evidence cover the entire range of clinical questions and provide the type of evidence with the lowest risk of confounding you would want to inform decisions. 

Concerning viral respiratory infections and whether interventions help, we must consider the substantial fluctuations in the outcome of interest - transmission - that…