Open Collections

Abstract

Health misinformation is an increasingly serious public health concern, particularly in relation to chronic conditions such as diabetes. The rapid spread of unverifiable information through digital platforms, especially on social media, makes it harder for people to find reliable information about diabetes symptoms, care, and treatment options. This study introduces DiaBERT, a transformer-based tool designed to identify misinformation related to diabetes, with the goal of categorizing health claims as True, False, or Partially True. The tool was trained on a varied dataset that merges formal, vetted content and informal online discussions, which presented challenges such as inconsistent terminology, shifts in context, and unclear labels. To address domain differences, DiaBERT was constructed using BioBERT and further refined with a domain adaptation strategy using a Domain Adversarial Neural Network (DANN), enabling the model to adjust to different linguistic styles. In evaluations, DiaBERT outperformed machine learning and deep learning models, achieving an accuracy of 79% and demonstrating the ability to detect nuanced claims labeled as Partially True. A content-filtering feature was also incorporated to help ensure that the system processes only diabetes-related claims. The system provides an explainability feature to enhance transparency and user confidence by highlighting important segments from the input text that influence the classification. Users can better understand the reasoning behind system decisions through this feature, which improves tool accessibility for non-experts. The Chrome extension release of DiaBERT enables users to receive classification results along with confidence scores and brief explanations during their online content exploration. The user study involving 45 participants demonstrated that the tool gained trust from users who found it helpful and simple to understand because of its keyword highlighting and explanation clarity features. This research tackles diabetes health misinformation while supporting better-informed decision-making.