Drug Review Sentiment Analysis: Applying Transformer-Based Models for Enhanced Healthcare

Abstract

Analyzing patient feedback on drug reviews is crucial in the healthcare sector as it determines the efficacy of treatment and patient experiences. Amidst the exponential growth in patient-generated data, the method of sentiment analysis has emerged as a key means of interpreting text-based reviews. In this research, the use of various machine learning and transformer-based approaches to analyze sentiments in drug reviews and gain meaningful insights from patient reviews or opinions is outlined. It juxtaposes traditional machine learning models such as Logistic Regression, Random Forest, and Support Vector Machines with deep neural networks such as Long Short-Term Memory and transformer-based models such as Bidirectional Encoder Representations from Transformers (BERT). Various models' performance is tested using the UC Irvine drug review dataset, and data preprocessing, feature extraction, and cross-validation are used in the study. Transformers, more precisely BERT, perform better than conventional approaches at 0.96 accuracy based on findings, as they can read into intricate patterns of language and contextual hints undetectable by basic models. The research reveals how transformer-based sentiment analysis can enhance healthcare decision-making through better and context-based information.

Received: 30 September 2024 | Revised: 14 January 2025 | Accepted: 4 September 2025

Conflicts of Interest

The author declares that he has no conflicts of interest to this work. 

Data Availability Statement

The data that support the findings of this study are openly available in UC Irvine Machine Learning Repository at https://archive.ics.uci.edu/dataset/461/drug+review+dataset+druglib+com. The data that support the findings of this study are openly available in GitHub at https://github.com/imabhi01/sentiment-analysis-dissertation.

Author Contribution Statement

Abhishek Chaudhary: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing—original draft, Writing—review & editing, Visualization, Project administration. Sangita Pokhrel: Supervision, Writing—review & editing. Swathi Ganesan: Writing—review & editing. Prashant Bikram Shah: Investigation, Writing—review & editing. Nalinda Somasiri: Writing—review & editing.