Accuracy of PubMed -based author lists of publications and use of author identifiers to address author name ambiguity: a cross-sectional study

Objective: To assess the accuracy of PubMed-based author lists of publications and use of author identifiers to address author name ambiguity. Methods: In this Swiss study conducted in 2019, 300 hospital-based senior physicians were asked to generate a list of their publications in PubMed and complete a questionnaire (type of query used, number of errors in their list of publications, knowledge and use of ORCID and ResearcherID). Results: 156 physicians (52%) agreed to participate, 145 of whom published at least one article (mean number of publications: 60 (SD 73)). Only 17% used the advanced search option. On average, there were 5 articles in the lists that were not co-authored by participants (advanced search: 1.0 (SD 2.6) vs. 5.9 (SD 13.9), p value 0.02) and 3 articles co-authored by participants that did not appear in the lists (advanced search: 1.5 (SD 2.0) vs. 3.6 (SD 8.4), p-value 0.05). Although 82% were aware of ORCID, only 16% added all their articles (39% and 6% respectively for ResearcherID). Conclusions: When used by senior physicians, the advanced search in PubMed is accurate for retrieving authors’ publications. Author identifiers are only used by a minority of physicians and are therefore not recommended in this context, as they would lead to inaccurate results.