Specifying your analysis in advance ... as far as possible

The final part of the Methods section in a protocol for a systematic review covers the plan of analysis. Ideally, you would tell the readers exactly what you are planning to do when you have collected all the study results together. This would mean that your analysis could not be influenced by the results you have looked at. In practice, it is not always possible to specify everything you will do in advance.

The compromise is to specify as much as possible in the protocol. Where you make decisions after seeing the study results, you should report exactly what you've done in the review, and usually present some analysis to show what would happen if you made a different decision.

Now would be a good time to read Section 8.2 of the Reviewers' Handbook, which covers some issues in planning the analysis for your review

Here's a checklist to use when writing the Methods section, followed by some practical tips

• what are the main comparisons in your review?
• for the outcomes you specified, how will you summarise the result for each study?
• how will you decide whether to combine the results of the separate studies?
• do you plan any subgroup or sensitivity analyses?

Specifying the comparisons

This should be fairly easy, as you should have thought it through carefully when you were specifying the types of intervention you were interested in.

As an example, imagine you are working on a review of whether corticosteroids (in either oral or intravenous form) were better than non-steroidal anti-inflammatory drugs (NSAIDS) or placebo for early rheumatoid arthritis.

Try writing a list of the possible combinations

A versus B
A versus C
B versus C ……and so on

Now set up a list of comparisons for your review. Draw on the Table you completed in Module 5 to help you do this.