4.3.7 Subgroups in meta-analysis

Subgroups refer to diverse grouping of studies based on specific characteristics of the studies such as study design. These characteristics may include the types of participants, types of comparators, and the outcomes. For example, it is possible to group all randomized experimental studies in one subgroup and all observational studies in another group; similarly reviewers may wish to group all studies with young participants in one subgroup and all studies with older participants in another subgroup. For these subgroups, it is possible to perform meta-analysis and to report the summary effects computed within subgroups. Also, it is possible to compare the summary effects computed in diverse subgroups. It is recommended that if subgroup analyses are performed these should be limited in number, should be pre-planned in the review protocol, and explanation and justification should be explicitly provided. These analyses should be carefully interpreted.