Background: Research on search efficiencies (for example, language, methodological and database restrictions) cannot rule out the possibility that applying these limits may result in missed studies. Therefore, uptake has been limited. However, the challenge of keeping reviews up-to-date suggests that some efficiencies may be needed. Objectives: To demonstrate how post hoc analysis of the contribution of sources searched for a review can inform decisions about the update search. Methods: Seventy-two updated systematic reviews of randomized controlled trials of drugs, devices or procedures were analyzed to see how complete identification of new evidence would be with a restricted search. A MEDLINE subject search was paired with a PubMed Related Article search based on the three newest and three largest studies from the original review and limited to randomized controlled trials entering PubMed since the original search date. Results: In 69 reviews, this search protocol retrieved all new eligible studies, either because one strategy or the other was sufficient or because the two together retrieved all studies identified by any means. In the remaining three reviews, a different combination of three methods, including searches beyond MEDLINE, was needed, and the combination was different in each case. Post hoc analysis of the actual indexing status of studies included in the original review will indicate to the review author if additional database or grey literature searches are likely to contribute to the update, once MEDLINE retrieval has been optimized through two independent search approaches, one of which works from key studies found in the original review and does not depend on searching skill. Conclusions: Multi-database, multi-modal searches help assure complete retrieval in original reviews. Yet, when most included studies proved to have been available from MEDLINE, a simple PubMed Related Articles procedure and a structured MEDLINE subject search was often sufficient for updating.