The Development of Highly Sensitive Search Filters for Complementary and Alternative Medicine-Specific, Natural Therapies for Use in Integrative Oncology

Introduction: To develop highly sensitive filters in MEDLINE and Embase for CAM-specific, natural therapies (dietary, supplemental, intravenous, physical and mental/ emotional) relevant to oncologic medicine. Systematic searches using these filters will be run sequentially for cancer topics by tumor type and symptom (e.g., breast cancer, peripheral neuropathy, thoracic cancers), from which we will write structured research summaries on all identified therapies. Methods: We used PubMed's dietary supplement subset as our base. We consulted with naturopathic doctors to identify and add vocabulary for further concepts. Building on previous work wherein a line-by-line version of PubMed's CAM subset was created in OVID MEDLINE, we developed a MEDLINE version of the expanded dietary supplement subset. We translated and adjusted this strategy for use in OVID Embase. Beginning with breast cancer, we ran multi-file searches and removed duplicates in OVID to reduce the deduplication burden. We compiled and circulated screening packages in Reference Manager. After tagging potentially relevant records, we exported them to Mendeley for additional tagging and full-text retrieval. We created a structured template for each relevant trial, including citation, study design, population, intervention, comparator, outcomes, side effects and interactions, to be used to present peer-reviewed data about human trial natural therapies in oncologic medicine. Results: We retrieved 54,061 records on CAM-specific, natural therapies in breast cancer. After deduping in OVID followed by further deduplication in Reference Manager, we reduced the number to 44,383. From this we identified and tagged 827 potentially relevant records and another 370 as "maybe" potentially relevant. We have finalized the tagging template from which we will derive our peerreviewed evidence summaries. Work is ongoing and we will further report developments as they occur. Discussion: Our customized CAM-specific, natural therapies filter has allowed us to efficiently and systematically identify evidence-based CAM material in MEDLINE and Embase for breast cancer. Relevant records have been tagged and will be developed into electronic evidence-based summaries. Our experience with breast cancer will allow us to fine-tune processes for further high-priority topics in integrative oncology. Translating these filters to additional databases (e.g., PsycINFO, CINAHL, AMED) will also be explored.