Delay and bias in PubMed medical subject heading (MeSH) indexing of respiratory journals

IntroductionPubMed is a primary global open-access literature research database. Articles on PubMed are indexed manually with medical subject headings (MeSH(R)) to facilitate more complete literature searches. We aimed to determine the length of delay from publication to MeSH(R) indexing for key respiratory journals and to investigate whether delays are increasing over time and whether there are country or impact-factor specific biases in indexing. MethodsPubMed was searched for the keyword asthma for the 10 year period June 2009 to June 2019. MEDLINE fields including journal title, publication date, PubMed entrez date and MeSH(R) indexing date were extracted and delay in indexing was calculated in days. ResultsTwenty-nine respiratory journals had delay from publication to MeSH(R) indexing ranging from an average of 153.6 days to 409.9 days; 5/29 (17.2%) had never been indexed for PubMed. There was a significantly longer delay in MeSH(R) indexing for UK-based publications (mean delay of 281.7 days) compared to USA-based publications (mean delay of 214.9 days; mean difference 66.8 days, 95% CI 60.8, 72.8, P < 0.0001). Delays in MeSH(R) indexing increased over time. ConclusionThere are long and increasing delays in PubMed MeSH(R) indexing of respiratory journals. All PubMed users should be aware that systematic literature searches that rely on MeSH(R) search terms or utilise PubMed filters are likely to exclude recent research and citations from key journals. Researchers and clinicians need to be aware of these delays and biases to ensure their literature searches are both up-to-date and complete.