Tracking patients healthcare experiences during the COVID-19 outbreak: Topic modeling and sentiment analysis of doctor reviews

  • Adnan Shah School of Management, Harbin Institute of Technology, Harbin, China.
  • Xiang Bin Yan Harbin Institute of Technology, Harbin, PR China
  • Syed Asad Shah Harbin Institute of Technology, Harbin, PR China


An emerging voice of patients in the form of opinion and expectation about the quality of care has the potential for improvement in service quality. A large volume of patients’ opinion as online doctor reviews are available online in order to access, analyze, and improve patients’ perception. This paper aims to mine online doctor reviews to explore the different determinants of patient satisfaction and dissatisfaction toward the United Kingdom (U.K.) healthcare services with the two-factor theory. This study is based on the two-factor theory, in light of online doctor reviews from a prominent health rating website in the U.K (, an improved text mining approach is proposed by introducing SenticNet to Latent Dirichlet Allocation (LDA) to analyze patients’ healthcare experience. This method finds latent topics that create either patient satisfaction or dissatisfaction toward a healthcare system. We then calculate the patient satisfaction rate across each disease specialty. Then sentiment analysis was performed using the Hourglass of Emotions model to compute the polarity of each review sentence (positive, or negative). Next, the classification results reveal that the proposed model that analyzes patient opinion toward a different aspect of care outperformed other state-of-the-art models. Finally, a SWOT and root-cause analysis is performed to identify strengths, weaknesses, opportunities, and threats to a medical system. Our findings provide a clue for doctors, hospital, and government officials to enhance patient satisfaction and minimize dissatisfaction by satisfying their needs and improve the quality of care across different diseases. Key words: patient satisfaction; text mining; topic modeling; LDA; sentiment analysis.

Industrial Engineering