Development of Search Filters for Clinical Studies of Herbs in PubMed

Abstract

Clinical studies of herbs are crucial to practice in alternative medicine. However, finding those articles from databases is difficult due to the complexity of terminology. While clinical search filters are helpful, to date there are no search filters specific to clinical studies of herbs. This study aimed to develop search filters to retrieve clinical studies of herbs from PubMed. Relevant clinical study of herb articles from 2019 to 2020 was manually identified from ten selected journals. Search terms were extracted from the articles using PubReMiner, and search strategies composed of one to four search terms were formulated and evaluated. Out of 8,169 articles, 211 were clinical studies of herbs. The two most effective search filters showed 100% sensitivity with 85% specificity and 99% specificity with 55% sensitivity. These queries, respectively, reached 99% sensitivity with 85% specificity and 99% specificity with 60% sensitivity in the validation set. The developed search filters yielded more relevant articles than PubMed’s clinical study filter when using the same herb name. Search filters for clinical studies of herbs were developed successfully. The best search strategy is to use these filters with the name of the herb that the researchers desire.