Searching the Question You Actually Asked

In practice

The question from video 3601 supplies the terms, and the search starts there rather than from a phrase typed whole. Databases are chosen before searching: CINAHL and MEDLINE through PubMed for the clinical literature, and the Cochrane Database of Systematic Reviews first, because a synthesis would answer the question outright. Each concept goes in alone. Pressure injury returns several thousand hits, so the subject heading is checked, because keyword and heading find different things and the heading catches records that never use the searcher's wording. Repositioning goes in alone. Patient positioning as a heading is added with OR, which broadens, since the two terms name one concept. The two concept sets are combined with AND, which narrows. Truncation on reposition catches repositioned and repositioning. Limits come last and cautiously: English language and adult are defensible, and a full-text limit is not, because it silently discards what the library could obtain by interlibrary loan. The yield is fifty-one records. The search history is saved, because the search is dated and will need re-running before anything is published. Knowing when you have found it is not a feeling: the same studies begin recurring across databases, and the reference lists stop offering new names.