Development and validation of a sensitive MEDLINE search strategy to identify literature relevant to limb prostheses.

BACKGROUND

Researchers and clinicians may find it challenging to identify relevant articles about limb prostheses in online databases. Searches may be improved by using standardized strategies, also known as filters or hedges.

OBJECTIVES

To develop and validate a highly sensitive MEDLINE (EBSCOHost) search strategy for limb prostheses.

STUDY DESIGN

Search strategy development/validation.

METHODS

A gold standard (GS) list of peer-reviewed articles on the topic of limb prostheses was created using a relative recall method. This involved identifying and including relevant systematic reviews/meta-analyses and then adding articles that were included in the analysis section of these reviews. Possible terms for the search strategy were identified through brainstorming and exploration of medical subject headings in MEDLINE (MeSH) and standards from the International Organization for Standardization (ISO). Concepts were grouped using relevant Boolean operators (e.g. AND and OR) and database record search fields (i.e. MeSH terms, title, and abstract). Part of the GS was used to refine the search strategy and reduce the number of records retrieved in total. The remaining GS set was used to validate and calculate sensitivity of the search strategy. Performance of the search strategy was compared against searches using only relevant MeSH terms.

RESULTS

After screening, the final GS totaled 853 records. The developed search strategy was highly sensitive (99.80%) and performed with higher relative recall than searches with relevant MeSH terms.

CONCLUSIONS

This study provides a highly sensitive search strategy that can be used by clinicians and researchers when searching for relevant literature on limb prostheses in MEDLINE (EBSCOHost).