Hybrid Sentiment Analysis Framework for Healthcare Provider Evaluation Using Patient Reviews

Online reviews for doctors and hospitals have become increasingly common, yet most rating platforms continue to reduce patient experience down to a single star score — which, in our view, barely scratches the surface of what patients actually experienced. This paper presents a Hybrid Sentiment Analysis Framework designed to evaluate healthcare providers more meaningfully, by combining the narrative content of patient reviews with their corresponding numerical scores. We gathered 500 patient reviews and ran them through a preprocessing pipeline covering stopword removal, tokenization, and lemmatization, then applied TextBlob and VADER to compute sentiment polarity for each review. The resulting sentiment score is blended with the numeric rating through a weighted formula to produce what we call the Final Provider Score (FPS). Our experiments show this hybrid approach correlates far more strongly with real patient satisfaction (r = 0.82) compared to star ratings alone (r = 0.61). We believe that deploying a framework like this on healthcare platforms could genuinely help patients choose the right provider — and help hospitals pinpoint exactly where they need to improve.