Obafunsho/LLM_Surgical_RCT_Abstracts: v1.0 — Code for LLM Surgical RCT Abstract Rewriting Study

This release contains the version of the code used for the study: "Improving reporting completeness of randomised trial abstracts with an automated large language model pipeline: development, validation, and large-scale assessment." The repository includes scripts for: • Retrieval of surgical RCT abstracts and full texts from PubMed Central via the NCBI Entrez API • Large language model–based abstract rewriting using GPT • CONSORT-based reporting completeness scoring • Readability analysis (Flesch–Kincaid Grade Level, Flesch Reading Ease) • Statistical analysis and figure generation Key scripts: downloads.py — data retrieval pipeline rewriting.py — LLM-based abstract rewriting scoring.py — CONSORT scoring and readability metrics main.py — orchestration of the full workflow This release represents a frozen snapshot of the repository corresponding to the analyses reported in the manuscript. A Zenodo DOI archive is generated automatically from this release to ensure long-term reproducibility.