No suitable precise or optimized epidemiologic search filters were available for bibliographic databases

Area of Science:

• Epidemiology
• Information Science
• Medical Informatics

Background:

• Systematic searches are crucial for identifying epidemiologic publications.
• Optimizing search strategies in bibliographic databases remains a challenge.
• Existing filters may lack the required sensitivity or precision for comprehensive retrieval.

Purpose of the Study:

• To evaluate suitable approaches for systematic searches of epidemiologic literature.
• To identify sensitive, precise, and optimized filters for MEDLINE searches.
• To assess the relevance of different bibliographic databases for epidemiologic research.

Main Methods:

• A test and validation set of epidemiologic publications (since 2007) was generated from systematic reviews and company dossiers.
• Epidemiologic filters were extracted from systematic review search strategies.
• Search syntaxes were validated using the relative recall method.

Main Results:

• The test set included 729 relevant epidemiologic publications; 566 were indexed in MEDLINE.
• One sensitive filter (Larney et al., 2013) demonstrated 95.94% sensitivity.
• No precise or optimized filters could be recommended due to underestimated precision; 77.64% of publications were found in MEDLINE.

Conclusions:

• Currently, no single approach guarantees efficient systematic searches for epidemiologic publications.
• The Larney et al. (2013) filter is suitable for sensitive MEDLINE searches.
• Further investigation into precise/optimized filters and alternative search strategies is warranted.