Summary statistics for individual studies

If you manage to collect useful outcome data from the studies you have included in your review, you will have some choices about how to express the outcome of each study.

These are covered in more detail in subsequent modules. For now, think about the types of data you are likely to have:

• are they dichotomous data (such as alive/dead, smoking/ not smoking)?
• are they continuous data (such as weight in kilograms or blood pressure)?

These are the main two data types you can analyse using RevMan.

The other two options in RevMan are "Individual patient data" and "Other data". Individual patient data reviews involve collecting data on each individual participant in the study. This takes a lot of time and effort, but enables more detailed analysis of the relationship between study characteristics (age of the participants, for example) and the effect of the intervention than is possible using summary data for the groups of participants. The "Individual patient data" option allows reviewers who have conducted these types of analyses to put the results into RevMan. More information on the use of individual patient data is given in section 11 of the Reviewers' Handbook. The option for other data is used to present data in a table without doing any statistical analysis on it. This might be used for qualitative data, or quantitative data where it is inappropriate to calculate pooled averages.

Choosing between different ways of presenting dichotomous and continuous outcomes is discussed in the RevMan exercise, under the heading 'Issues in analysis'. If you haven't yet reached that part, now would be a good time to complete the whole exercise.

The RevMan exercise gives a basic introduction to the issues, which will be covered in more depth in future modules. For your protocol, you need to state how you plan to present these summary statistics for studies. Your review group may have a policy on this, so it's worth checking.