Skip to main content

Table 7 Some starting principles for qualitative searching

From: Searching for qualitative research for inclusion in systematic reviews: a structured methodological review

Component

Starting principles

Sampling

Where approaches other than comprehensive sampling are used, reviewers must justify their sampling strategy, match it to their synthesis method and describe fully how it was implemented.

Sources

For health topics, MEDLINE and CINAHL are considered a minimum, augmented by topic-specific and setting-specific sources. Reviewers should devise specific strategies to find specific types of grey literature, where included.

Structured questions

In the absence of empirical data on effectiveness of structured approaches, the question structure should be selected to match the purpose and focus on the review. When accompanying a review of clinical trials, the two review questions may or may not be co-terminous.

Search procedures

Given the comparatively low yield of qualitative topic-based searches, reviewers should privilege specificity (retrieval of relevant items). Retrieved relevant items can then be used as a starting point for developing supplementary search techniques. Reviewers should compensate for reported deficiencies in indexing by using a broad range of supplementary strategies.

Search strategies and filters

Filters should be commensurate with the intended purpose of the review. When extensive supplementary strategies are being employed to offer improved sensitivity, the topic-based searches may use a simple filter (using terms such as qualitative OR findings OR interview).

Supplementary strategies

Reference checking must be a default for every review. For diffuse topics, or those with significant variation in terminology, hand searching, citation searching or contact with authors/experts may be relatively productive. Where context or theory is particularly important, the CLUSTER method [110] may be appropriate. Trial identifiers (ISRCTN or trial name) may be useful for sibling or kinship studies for trials.

Standards

In the absence of a consensual standard for reporting, ENTREQ [88], supplemented by PRISMA [128] and STARLITE [12] where necessary, should be used when reporting a search.