OBJECTIVES: In December 2022, a systematic literature review (SLR) was conducted to identify clinical trials on adults with human epidermal growth factor receptor (HER2) negative, hormone receptor (HR) positive, inoperable or metastatic breast cancer. The search strategy resulted in 22,218 hits. This project explores ways to make the search strategy more efficient, in terms of reducing the screening burden, without excluding relevant studies.

METHODS: The original search strategy was developed to maximise sensitivity (i.e. studies identified/all relevant studies). The population terms comprised terms for breast cancer combined with terms for advanced, metastatic, and non-resectable disease. We applied stepwise alterations to the existing population search terms, measuring sensitivity and precision (studies identified/number screened) after each change. Of particular interest was how the search strategy performed when limited by terms for HER2 and/or HR status. The final set of included studies from the SLR was used as the reference set (n=26 papers [14 unique trials]; n=24 indexed in Embase at time of search).

RESULTS: The precision of the original search was 0.22%, 0.35% and 0.39% for Embase, Medline and CENTRAL. The total number of hits were 10,880, 5,385 and 5,767. In Embase, the highest precision was seen combining HR/HER2 terms with the population terms using the AND Boolean operator (0.42%; 2,124 hits); however, this resulted in a sensitivity of only 37.5% (9/24 of the relevant studies identified). The best compromise between sensitivity and precision was found in a CENTRAL search that combined the original terms with HR/HER2 terms (65.3% and 0.54%; 3,146 hits); however, 9 papers (5 unique trials) were excluded.

CONCLUSIONS: Limiting searches using disease subtype-specific population filters may reduce the screening burden in large topics, albeit at the expense of reductions in sensitivity. Depending on the research objective, the review could be limited to CENTRAL to reduce screening burden.