Controlled Vocabulary Services Program

The work of the Controlled Vocabulary Services Program directly supports the “Curate” component of what the User Services and Collection Division does.

NLM Indexers support the biomedical research community by making it easier for users to find relevant journal articles in MEDLINE, the National Library of Medicine’s biomedical and life sciences database which comprises the majority of content in PubMed, through the application of the MeSH (Medical Subject Heading) vocabulary. MEDLINE articles are automatically indexed with MeSH via the MTIX (Medical Text Indexer-NeXt Generation) algorithm, which uses a machine-learning model known as a neural network. NLM Indexers with biomedical domain expertise provide quality assurance of automated MeSH indexing, with curation focused on specific sets of articles (e.g., those involving genes/proteins or involving known instances of term ambiguity). These domain experts contribute to the continuous refinement of the MTIX algorithm and the expansion of the MeSH vocabulary; they also create links between the biomedical literature in MEDLINE and the NCBI Gene database to describe new gene information.

NLM Indexers also provide descriptive data for article citations that documents research funding and databank relationships. These staff also create links between the literature for comments, errata, retractions, and expressions of concern.

In addition to their fundamental role in supporting MEDLINE indexing, NLM Indexers have applied their knowledge to many special projects through collaboration with NLM and external researchers by providing gold standard data sets used in AI approaches to knowledge discovery.

For more information about MEDLINE data review and indexing, see: