Tip #2: Check for erroneous computer indexing in Embase

 I recently sought help from colleagues on the expertsearching listserv regarding puzzling results I was seeing in Embase (in this case, on the Ovid platform).

I was searching for articles about the category of medical trainees known in the USA as residents. Here's the scope note for the EMTREE term "resident":

(For screen reader users: The scope note cites the Dorland's definition, "a graduate and licensed physician receiving training in a specialty, usually in a hospital," and states this subject heading is synonymous with "interns and residents, resident doctor, resident physician, resident surgeon, surgery resident, surgical resident.")

So far, so good, right? Well, here are the first 5 results for exp resident/: 

1. COVID-19 in French Nursing Homes during the Second Pandemic Wave: A Mixed-Methods Cross-Sectional Study
2. COVID-19 vaccine effectiveness against laboratory confirmed symptomatic SARS-CoV-2 infection, COVID-19 related hospitalizations and deaths, among individuals aged 65 years or more in Portugal: A cohort study based on data-linkage of national registries February-September 2021
3. Health impact assessment of salt reduction in bread in Portugal: A pilot study.
4. Coronavirus (COVID-19) Spike in Georgia: An Epidemiologic Study of Data, Modelling, and Policy Implications to Understand the Gender-and Race-Specific Variations.
5. Outbreak of COVID-19 among vaccinated and unvaccinated homeless shelter residents - Sonoma County, California, July 2021.

None of these, as it turns out, are about medical residents; they all refer to residents of a community, such as a nursing home, a geographic place, or a homeless shelter. They all had been incorrectly indexed with the EMTREE term "residents"! So I turned to the expertsearching list for help. A knowledgeable colleague explained to me something that the likely explanation was that unlike in MEDLINE, where in-process records do not have MeSH assigned yet, in Embase newly added records undergo computerized indexing with EMTREE. Eventually, indexing is reviewed and corrected by a human indexer for journal articles, while indexing of conference proceedings never undergoes human review.

Needless to say, this has major implications for Embase searchers. Recent results and conference abstracts will have a high number of false positives due to incorrect computer indexing, and EMTREE will only be reliable for fully indexed journal articles. Likewise, we will likely have to change our approach to searching MEDLINE as well when it transitions to automatic indexing in mid-2022.