Table 5

Reasons why studies were missed by electronic search and handsearch

Category

Reason

Studies citing reason (n = 42)


Electronic Resources

Indexing

Inadequate indexing in general

16[3,27,29,32–44]

Journal not indexed in resource

8[3,13,20,27,35,40,45,46]

Not indexed with proper methodologic terms

9[21,26,28,29,32,36,42,47,48]

Not indexed as RCT/CCT

7[35,37,49–53]

Not indexed for any of the methodologic terms used by searchers

2[21,54]

Database does not contain relevant index terms

2[17,50]

Not indexed with any of the subheadings used in searches

1[21]

Search Strategy

Insufficient or restricted search strategy

6[27,32,35,55–57]

Reporting

Keywords or methodology not reported by author

9[13,19,29,34,41,48,53,58,59]

Database

Issue(s) omitted from resource

4[3,26,56,60]

Time lag in updating resource

2[60,61]

Article(s) omitted or missing from resource

6[3,29,36,50,51,56]

Journal

Published before database was created or beyond coverage years of database

2[56,60]

Other

Unknown reason

6[26,38,38,54,60,62]

Published as abstracts, books, book reviews, brief reports, letters, proceedings, supplements, etc.

9[20,26,38,38,49,51,56,63]

• Wrong abstract assigned to reference[36]

• Journals do not encourage authors to explicitly report methodology[37]

• Length and complexity of search is limited using search engines[64]

• Ongoing or recently finished studies[64]

• Found through references[20]

• Missed most Japanese reports[65]

• Some personal communication did not get responses; organizations and people must be willing to supply trial information[18]

• Not all abstracts are published in full[50]

• Authors misspelled gingko[39]

• Many RCTs are not identifiable in database[66]

Handsearch

Handsearchers not trained properly

2[57,67]

Methodologic terms "hidden" in article

1[26]

Searcher fatigue, boredom

1[26]

Journal not handsearched

2[35,42]

Article not an RCT or CCT

1[54]

Other

Different topic than what handsearchers were looking for

1[42]


Crumley et al. BMC Medical Research Methodology 2005 5:24   doi:10.1186/1471-2288-5-24

Open Data