Accurate Evaluation and Forecasting in Chemotherapy-Related Information Needs of People With Breast Cancer: Insights From an Online Medical Consultation Platform.

Background: Chemotherapy-related information needs are crucial for patient-centered care, yet their specific domains remain uncharacterized, particularly lacking large-scale, high-quality evidence from real-world clinical communication contexts. Methods: The current study used the "Haodf.com" Internet-based consultation platform. Consultation records of 3000 breast cancer patients were obtained using province-stratified random sampling with proportional allocation. Using BERTopic modeling and structured data input methods, basic information and chemotherapy-related information needs of patients were identified. Chi-square tests and binary logistic regression were used to examine associations between patients' clinical and demographic characteristics and each information need. Results: BERTopic modeling initially delineated 10 information need topics; manual structured data entry and validation subsequently uncovered an additional 11 latent needs, resulting in a final taxonomy of 21 need categories. The study enrolled 2781 valid consultation records, from which an average of 2.11 chemotherapy-related information needs were extracted per record. "Treatment options available," "benefits and risks of chemotherapy," "problems of side effects of chemotherapy drugs," and "current stage of disease and test results" were the top-ranked information needs. Different information needs were identified and differentiated across specific dimensions, including gender, age, disease metastasis, illness duration, treatment stage, and frequency of interactions. Conclusions: By creating a quantifiable map of information needs related to breast cancer chemotherapy, the current study established an evidence-based framework for the creation of precise, patient-centered, and dynamically adjusted interventional care.