Abstract
Background: As the healthcare landscape progressively adopts a patient-centered paradigm, the imperative to enhance patient experience has become more pronounced. Efforts to improve patient experience have yielded modest results, partly due to limited understanding of the key factors influencing patient expectations. Objective: To explore the determinants of patient experiences through analyzing patient feedbacks, assisting healthcare institutions in prioritizing service improvements. Methods: A digital topic modeling approach was employed. Data were derived from a secondary analysis of the National Patient Experience Base, incorporating patient feedback from 226 hospitals. Initially, the feedback text data underwent a cleansing process, and the sentiment intensity within the text was quantified using the SnowNLP algorithm, and XGBoost classifier was utilized to categorize sentiments as positive or negative. Subsequently, the feedbacks were subjected to topic clustering using the BERT model and X-means clustering algorithm. Third, TextRank was applied to extract significant keywords from each cluster, and these keywords were analyzed to identify the determinants that impact patient experience. Results: A total of 4689 patients' feedbacks were collected, comprising 2918 outpatients and 1771 inpatients from 165 tertiary and 61 secondary hospitals across 24 provinces. Through cluster analysis, 10 main clusters emerged (two of which were positive response and eight were negative response). By qualitatively synthesizing, patient experiences were distilled into five determinants: treatment, service, environment, economic, and process. Conclusions: The findings underscore the importance of a holistic approach to patient experience enhancement, where healthcare providers must address not only the clinical aspects of care but also the service delivery, environmental conditions, economic considerations, and procedural efficiency. By identifying and prioritizing the improvement of these determinants, healthcare organizations can tailor their services to better meet patient expectations and enhance overall satisfaction.