Transforming Patient Feedback Into Actionable Insights Through Natural Language Processing: Knowledge Discovery and...

Area of Science:

• Health Informatics
• Natural Language Processing
• Knowledge Discovery in Databases

Background:

• Patient feedback is crucial for healthcare quality and organizational performance.
• Manual analysis of unstructured feedback is challenging due to data volume.
• Extracting actionable insights from patient comments requires efficient methods.

Purpose of the Study:

• Develop and evaluate a methodology for analyzing patient feedback data.
• Utilize natural language processing (NLP) and Knowledge Discovery in Databases (KDD) approaches.
• Identify patterns, themes, and variations in patient experiences to improve healthcare delivery.

Main Methods:

• Applied an integrated KDD-action research framework to 126,134 patient feedback entries.
• Employed text mining techniques: sentiment analysis, topic modeling, emotion detection, and aspect-based sentiment analysis.
• Ensured validity through multiple analytical techniques and stakeholder engagement.

Main Results:

• Overall sentiment was moderately positive (0.42 average polarity); 68.8% of comments were positive.
• Key topics included staff attitude (10.2%), professionalism (10.1%), environment (10.0%), and waiting time (10.0%).
• Positive aspects: nurse attitude (0.65), staff helpfulness (0.61); Negative aspects: waiting time (-0.42), billing transparency (-0.28). Younger patients prioritized digital efficiency, older patients valued face-to-face interaction.
• Interventions led to improved waiting time satisfaction (+18%), doctor-patient communication (+15%), and reduced billing complaints (-23%).

Conclusions:

• NLP and KDD provide a robust framework for transforming patient feedback into actionable healthcare improvements.
• This approach enables understanding patient experience drivers and identifying targeted improvement opportunities.
• Evidence-based initiatives derived from feedback analysis enhance care quality and patient-centeredness.