Abstract
Introduction: Non-adherence to medication continues to be one of the most pressing global health problems that has implications for treatment outcomes and patient safety. As mobile health (mHealth) technologies have increased, medication reminder applications have been used more widely to assist people with adherence and self-management. This study aimed to investigate user perception, emotion, and satisfaction with the medication reminder application using a combination of sentiment and thematic analysis of real-world user reviews.
Methods: A mixed method was used, and 4,312 user reviews were obtained from the Apple App Store and Google Play Store. Reviews were pre-processed using natural language processing methods. Moreover, sentiment analysis was conducted using a combination of lexicon-based models, VADER, and TextBlob, as well as the supervised machine-learning model and support vector machine, achieving a 91% accuracy. Finally, thematic analysis was performed using NVivo software to identify repeatable emotional and function-controlled themes.
Results: Among the total reviews, 68.6% expressed positive sentiments, while there were 14.7% neutral and 16.7% negative sentiments. The iOS group had 71.3% positive rating that was slightly higher than that of the Android group (67.2% positive rating). Using thematic analysis, five themes were identified: ease of use, reminder reliability, customization, technical performance, and emotional support. The regression analysis revealed that reliability, user-friendliness of the interface, and durability were the most important factors contributing to positive sentiment (R²=0.61).
Conclusion: In general, users over the counter services to the medication reminder apps, predominantly pointing at usability, reliability, and emotional support as the key factors. The research indicates that the merging of sentiment mining and thematic analysis demonstrates a great potential to unravel user-friendly and emotionally supportive digital adherence tools.