Google Versus PubMed: Comparison of Google and PubMed's Search Tools for Answering Clinical Questions in the Emergency Department.

STUDY OBJECTIVE

We assess which search tool returns the highest-quality, most relevant citations for standardized clinical questions arising at the point of care in the emergency department (ED).

METHODS

Search terms related to 3 clinical questions commonly encountered in the ED were entered into 5 search tools. The following search engines and tools were assessed: Google Web, Google Scholar, PubMed, PubMed Clinical Queries set to narrow search, and PubMed Clinical Queries set to broad search. The first 60 hits, in the order of most recent first, were reviewed and assessed for quality of evidence, relevance versus irrelevance, and number of high-quality hits, and each search instrument was graded for overall readability with a visual analog scale. Quality relevance ratio, defined as the ratio of high-quality relevant hits to low-quality irrelevant hits, was calculated for each tool according to these searches.

RESULTS

Overall, PubMed Clinical Queries narrow search had the highest quality relevance ratio, averaging 0.85. PubMed Clinical Queries narrow search also returned high-quality relevant hits without the need to filter out as many low-quality irrelevant hits. Google Scholar retrieved the highest number of systematic reviews and randomized controlled trials, and returned the most complete search results, finding relevant citations other search engines did not. Google Web consistently had the lowest quality relevance ratio and contained many duplicate hits.

CONCLUSION

For the common clinical questions assessed in this study, PubMed Clinical Queries narrow search had the highest-quality, most relevant, and most readable hits. Google Scholar performed well, in some cases retrieving citations that other search engines did not. PubMed and Google Web were not as efficient.