Assessing PubMed metatag usage for plain language summary discoverability [encore]

Objective

Tagging plain language summaries (PLS) on PubMed can maximize discoverability, accessibility and usage of publications by a broader audience; open access (OA) can also enhance these characteristics.1,2,3 We aimed to (1) determine proportions of PubMed records correctly using the PLS tag and reasons for incorrect usage, and (2) establish OA statuses of journals publishing PLS on PubMed.

Research design and methods

The entire PubMed database was downloaded (9 February 2022) and searched for PLS indexed with XML tags in the ‘Other Abstract’ field. Records were de-duplicated; incorrectly tagged PLS were programmatically excluded and confirmed with manual spot checks. Remaining PLS were categorized by journal; OA status was assessed using Journal Selector (Sylogent LLC, Bristol, PA, USA) or information from journal websites. 

Results

Overall, 3217 records were identified with XML tags, of which 1644 (51.1%) were published in 2021 (incidence rate = 929 per million publications). There were 470 records (14.6%) with incorrect usage: non-English scientific abstract (n=137); duplication of, or >90% similarity score with, scientific content in the ‘Abstract’ field (n=32); absence of scientific content in the ‘Abstract’ field (n=99); other non-PLS content (n=197); no content (n=5). Of 105 journals correctly using the tag, 30 (28.6%) were full/gold OA journals and 75 (71.4%) offered OA options. 

Conclusions

Despite increasing use of the tag,1 such records represent a small minority of all PubMed records, and the tag is commonly used incorrectly. So far, PLS on PubMed have been published in journals with OA options. Explicit guidance is needed on indexing processes for improved tagging; greater uptake presents opportunities to increase publication impact.