Does it work? (2)

More brief comments on how to spin results

Just after posting the first article of this brief series

Does it work?

We received a private email. Something we do not encourage, as we prefer open dialogue.

In this case, the emailer made a point that allows us to explain further why both relative and absolute measures of effect should be used in results reporting.

The emailer argued that the use of risk ratios (RR or Relative Risk Reduction RRR) is widespread, so it’s the right thing to do. We agree with this, but as already stated, only if accompanied by absolute risk reduction. In the case of respiratory viruses, this is even more important if you have large datasets covering multiple “seasons” or periods, as in our Cochrane review example.

The logic of the “everyone does it” argument is strange. It is akin to saying, “eat cow dung as two trillion flies cannot be wrong.”

If the logic is faulty, so is the statement that everyone does it, so it’s the right thing to do. This is not so.

The purpose of a systematic review and a meta-analysis is to infor…