How to Search for Randomized Controlled Trials

To find randomised controlled trials (RCTs), combine your topic concepts with a validated trial filter or publication-type terms in PubMed, then supplement with trial registries and the Cochrane Central Register of Controlled Trials. Trials are indexed inconsistently, so a topic-only search with no trial terms plus manual screening can be more sensitive.

Options in PubMed

• Use the Randomized Controlled Trial publication type or an article-type filter. This is quick but can miss unindexed or mislabelled trials.
• Use a published sensitive search filter for RCTs (Cochrane Handbook describes recommended filters). If you use one, cite it and give the exact string.
• Search topic concepts only and screen for design yourself, when sensitivity matters.

Other sources

• ClinicalTrials.gov and other registries for ongoing, completed and unpublished trials.
• CENTRAL (Cochrane Central Register of Controlled Trials).
• Reference lists and citation searching of included trials and reviews.
• Conference abstracts for results not yet published.

Illustrative example

Invented for teaching. For "vitamin D and asthma exacerbations in children", the strategy combines topic concepts, then applies a sensitive trial filter. The reviewer also searches ClinicalTrials.gov for completed trials with no publication and finds two. She records them in the search log and contacts the registrants for results.

How sensitive does the search need to be?

For a systematic review of treatment effects, missing trials can change the result, so favour sensitivity and accept more screening. For a quick clinical question, a precise filter that returns fewer, more likely relevant trials may be adequate. Decide before you search and state the choice in the protocol. Also remember that some trials are published only as abstracts or results postings on registries, which database filters may not capture.

Common pitfalls

• Relying on the publication-type filter alone.
• Ignoring registries, which exposes the review to publication bias.
• Using a filter designed for another database without checking.

Key takeaways

• Trial filters increase efficiency but reduce sensitivity; choose by purpose.
• Add registries and other sources.
• Document the filter and its source.

This page is general educational information about research methods. It is not medical, legal or regulatory advice; follow your institution's ethics board and your target journal's instructions.