Validating GAN-BioBERT: A Methodology for Assessing Reporting Trends in Clinical Trials.

Area of Science:

• Biomedical Informatics
• Natural Language Processing
• Clinical Trial Analysis

Background:

• Assessing trends in biomedical literature requires robust tools for analyzing clinical trial abstracts.
• Previous sentiment classification models lacked reproducibility and fine-grained analysis capabilities.

Purpose of the Study:

• To validate a novel three-class sentiment classification model for clinical trial abstracts.
• To leverage adversarial learning and the BioBERT language model for reproducible trend assessment.
• To compare the model's performance against existing methods.

Main Methods:

• Developed a semi-supervised, three-class sentiment classification algorithm for clinical trial abstracts.
• Utilized 108 expert-annotated and 2,000 unlabeled abstracts.
• Employed a Bidirectional Encoder Representation from Transformers (BERT) based model, enhanced with adversarial learning.

Main Results:

• Achieved a classification accuracy of 91.3% and a macro F1-Score of 0.92.
• Significantly outperformed previous models in sentiment classification of clinical trial literature.
• Demonstrated finer-grained sentiment classification with enhanced reproducibility.

Conclusions:

• The validated model provides an easily applicable tool for sentiment classification of clinical trial abstracts.
• The model significantly outperforms previous approaches in terms of accuracy and reproducibility.
• Offers applicability for large-scale studies on reporting trends in biomedical literature.