Sentiment Analysis of Patient Feedback for Strategic Business Decisions in Hospitals

Abstract: Background. Patient narratives are increasingly recognized as a valuable source of intelligence for healthcare quality improvement, yet their systematic integration into performance management remains limited. Sentiment analysis, particularly aspect-based models, offers opportunities to align patient feedback with strategic metrics. This study investigated the role of sentiment analysis in patient experience, operational performance, leadership decision-making, and organizational sustainability. Methods. A multi-method analytic design was applied to a sample of 50 hospitals, yielding 600 hospital-month observations across one calendar year. Patient feedback was collected from internal hospital surveys, complaint logs, and narrative comments submitted through electronic portals. Data were de-identified in accordance with HIPAA standards, with duplicate entries removed and missing cases handled via listwise deletion (<1% loss). Aspect-based sentiment analysis was conducted using a validated LLM-assisted model, producing domain-level polarity scores (communication, access, billing, discharge). These sentiment scores were linked to Centers for Medicare & Medicaid Services (CMS)-reported HCAHPS domain scores and hospital operational metrics, including no-show rates, call-center wait times, and length of stay. For leadership decision-making (RQ3), a survey of 120 hospital leaders assessed adoption, perceived usefulness, and decision-making effectiveness using validated scales consistent with the Technology Acceptance Model. RQ4 compared hospitals with governed, validated sentiment pipelines (n = 25) and those without governance (n = 25) using quarterly aggregated KPI improvements. Results. Aspect-based sentiment scores were significantly associated with HCAHPS domain scores (average r = .42, p < .001), supporting measurement validity. Regression models demonstrated that sentiment scores predicted operational outcomes, including reduced no-show rates (β = −3.21, p < .001) and shorter call-center wait times (β = −2.3 minutes, p < .001). Mediation analysis confirmed that perceived usefulness mediated the effect of adoption on decision-making effectiveness (indirect effect = 0.21, 95% CI [0.09, 0.37]). Finally, hospitals with governed pipelines achieved significantly greater and more sustainable KPI improvements, including a −4.7% reduction in no-shows versus −1.9% in non-governed hospitals, t(48) = 7.28, p < .001. Conclusions. Sentiment analysis operates at multiple levels of healthcare performance: validating patient experiences, predicting operational efficiency, enhancing leadership decision-making through perceived usefulness, and sustaining improvements through governance. The findings advance theory by extending the Technology Acceptance Model and advance practice by offering a roadmap for embedding sentiment analysis as a strategic tool for patient-centered, efficient, and sustainable healthcare delivery. Keywords: sentiment analysis; patient experience; operational performance; Technology Acceptance Model; healthcare leadership; governance. Title: Sentiment Analysis of Patient Feedback for Strategic Business Decisions in Hospitals Author: Dr. David Augustine Bull International Journal of Healthcare Sciences ISSN 2348-5728 (Online) Vol. 13, Issue 1, April 2025 - September 2025 Page No: 610-631 Research Publish Journals Website: www.researchpublish.com Published Date: 27-August-2025 DOI: https://doi.org/10.5281/zenodo.16964956 Paper Download Link (Source) https://www.researchpublish.com/papers/sentiment-analysis-of-patient-feedback-for-strategic-business-decisions-in-hospitals