Dataset for: Trends in Diagnostic Terminology and Consensus Adoption in Fracture-Related Infection Literature: A Contextual Text Mining Analysis
Description
This dataset contains the results of a context-aware large language model (LLM) pipeline applied to full-text PDFs of fracture-related infection (FRI) literature. The pipeline was designed to identify and count FRI-related diagnostic terminology while accounting for contextual usage, reducing false positives from non-diagnostic mentions of keywords. The dataset includes article-level counts of terminology associated with diagnostic criteria and consensus definitions, generated through structured, context-sensitive text analysis. These data were used to evaluate trends in diagnostic terminology and consensus adoption over time, as described in the associated manuscript: “Trends in Diagnostic Terminology and Consensus Adoption in Fracture-Related Infection Literature: A Contextual Text Mining Analysis.” Code used to generate these results is available separately. (https://data.mendeley.com/datasets/bc5vsrfrv6/1)
Files
Steps to reproduce
The results in this dataset were generated using a context-aware large language model (LLM) pipeline applied to full-text PDFs of fracture-related infection (FRI) literature. To reproduce these results, users should run the analysis pipeline using the code repository associated with this project (https://data.mendeley.com/datasets/bc5vsrfrv6/1), following the instructions in the repository README. This dataset represents the processed outputs used in the manuscript and can be used directly to reproduce reported analyses and figures.
Institutions
- University of Colorado Anschutz Medical CampusColorado, Aurora