When the Search Returns Almost Nothing

In practice

A clinician asks whether a particular hand-off script reduces omitted information at shift change on a paediatric oncology unit. The first search returns four records, none of them close. The rebuild comes first, because too little is more often a search fault than a literature fault. Subject headings are checked against keywords. The population is broadened from paediatric oncology to inpatient nursing, since transferability is judged after finding rather than before. OR is used across handover, hand-off and handoff. Limits are stripped back. The yield rises to thirty-one, of which three are relevant and none is a synthesis. The grey literature is searched, and a healthcare librarian is asked, which is the step most often skipped. What remains is a small body: one quasi-experimental study, one case series, and a quality improvement report. The response is neither to abandon the question nor to overstate what was found. Lower-level designs are read for what they can offer, which is description and hypothesis rather than cause, and their high risk of bias is stated rather than discounted. The recommendation records that the evidence is thin, names what would strengthen it, and proceeds on clinician expertise and patient preference with the internal evidence watched closely.