Quantifying research on anatomical structures, a potential new metric for assessing clinical relevance

Abstract

Human anatomy remains an integral part of medical education, and recent studies have documented an emerging consensus on the key anatomical learning objectives for physicians and other health professionals in training, both at the graduate and postgraduate levels. Despite this progress, less attention has been given to assessing the clinical relevance of individual anatomical structures, and which structures students should master to achieve these learning objectives. In this study we hypothesized that published research involving individual anatomical structures is largely driven by the clinical relevance of these structures, and that tabulating the number of such publications can provide an up-to-date, evolving metric of clinical relevance. To test this hypothesis, we developed a semi-automated search routine that uses the PubMed database to quantify the publication frequency of anatomical structures and compared that to a previous study that assessed the importance of structures of the head and neck using the Delphi method, a formal procedure of generating expert consensus. Using our new approach, we were able to rank the research intensity of 2182 anatomical structures included in Grant's Dissector, a widely used textbook for

Further assessment of the clinical relevance of anatomical structures through PubMed and other measures

“…In previous research, 2181 unique bolded anatomical structures from the 15th edition of Grant's Dissector (Tank, 2012) were extracted (Dahl et al, 2023). To determine the level of research conducted on each anatomical structure, the PubMed database was utilized, which quantifies the number of citations for each structure (National Center for Biotechnology Information, 2022).…”

“…Detailed explorations of different search parameters for individual terms led us to conclude that when the search with quotes and the search without quotes yielded a similar number of citations, that number was a reliable estimate of the research intensity for that particular anatomical structure. This inference was supported by the fact that these structures had a stronger correlation with the ranks derived from Tubbs et al (2014) (Dahl et al, 2023). This discovery led to the creation of High Confidence and Low Confidence lists, and unfortunately, 34.6% of the bolded structures were on the Low Confidence list.…”

“…Our approach to testing whether our PubMed-derived ranks reflect clinical significance is to assess whether there is a statistically significant correlation with expert derived ranks. Although we have found evidence to continue exploring this approach (Dahl et al, 2023), there are good reasons for continued caution. The concept of "clinical relevance" is hard to measure, and ideally, we would further test PubMed-derived ranks with other types of data.…”

“…In the previous PubMed study (Dahl et al, 2023), it was observed that search results were often combined with Medical Subheading (MeSH) terms. MeSH is a controlled vocabulary of standardized and specific descriptors maintained by the National Library of Medicine to index and categorize articles in medical and scientific literature for efficient retrieval of information (NIH, 2020).…”

“…It was noticed that some anatomical terms were retrieving irrelevant citations in PubMed due to the automatic addition of MeSH terms, resulting in inaccurate outcomes for certain searches. Therefore, a second search was conducted with quotes to eliminate MeSH terms from the results (Dahl et al, 2023). Although this removed irrelevant citations from search results, the use of quotes caused its own problems; many relevant papers were excluded because of slight differences in the ordering of multi-word names for anatomical structures.…”