Toward More Informed Choices: Analysis of the Side Effects of Female Contraception Options Using Natural Language Processing of Online Reviews

Abstract

Abstract Purpose This study aims to explore women's online descriptions and evaluations of their experiences with birth control products, utilizing natural language processing (NLP). Focusing on non-favorable reviews, the objective is to shed light on the issues and side effects discussed by women, providing additional information that could aid women and their health care providers in making informed contraception choices. Methods Employing topic modeling and descriptive statistics, this study analyzes 19,524 unique patient reviews of female contraceptive products posted on Drugs.com. The study also explores patterns in drug ratings depending on the side effects described and contraception products reviewed. Results Topic analysis identifies eight main areas of side effects: weight gain, skin problems, loss of libido, mental health issues, menstrual irregularities, cramps and pain, continuous bleeding, and multiple cause dissatisfaction. Descriptive analyses reveal that female contraceptive products vary greatly in how frequently and firmly their side effects are described by women. Drug ratings, indicative of a product’s impact on reviewers’ well-being, vary substantially with contraception type and prevalent complaints voiced in the reviews. Conclusion Although exploratory in nature, the study underscores the value of using NLP to analyze large volumes of online reviews for obtaining qualitative insights into women's experiences with contraceptive products. This method can be a useful tool for helping women and doctors make more informed decisions, despite the inherent risk of bias in online reviews. These findings serve as a preliminary guide, suggesting the need for further research to confirm the links between specific side effects and women’s well-being.