Year: 2023 | Month: October | Volume: 13 | Issue: 10 | Pages: 275-288
DOI: https://doi.org/10.52403/ijhsr.20231038
Enhancing Patient Experience by Automating and Transforming Free Text into Actionable Consumer Insights: A Natural Language Processing (NLP) Approach
Ela Vashishtha1, Himanshu Kapoor2
1Master of Health Administration, Designation: Healthcare Planning and Strategy Leader, Texas Health Resources, Texas, USA
2Master of Science in Engineering, Designation: Product Management and Strategy Leader, University of Washington, Seattle, USA
Corresponding Author: Ela Vashishtha
ABSTRACT
Background: The digitalization of healthcare has expanded patient access to care delivery through various virtual platforms. These platforms accumulate a wealth of patient feedback data, offering valuable insights into care perception and experience. Natural Language Processing (NLP) provides a means to efficiently analyze this unstructured patient feedback, enabling healthcare organizations to enhance care quality and patient experience.
Methods: This study collected 19,000 comments from diverse sources, including surveys and social media, spanning from January 1, 2018, to June 30, 2021. The data underwent preprocessing, including lowercase conversion, special character removal, and lemmatization. A "hot word" list was employed to identify critical patient safety concerns. NLP models, including sentiment analysis and text classification, were used to analyze patient feedback.
Results: The sentiment analysis revealed that most patient comments were positive or neutral, with 34.5% positive, 52.6% negative, and 12.9% neutral sentiments. Aggregate analysis identified access and login issues as a primary source of dissatisfaction, with 78% of related comments expressing negativity. The study also utilized Word2Vec to uncover word associations, highlighting positive associations with "staff" and negative associations with "website" and "insurance." Hot word detection identified 580 comments requiring immediate attention for patient safety.
Conclusions: NLP-enabled analysis of patient feedback offers actionable insights for healthcare organizations. It identifies areas for improvement, tracks sentiment trends, and aids in patient engagement. Additionally, NLP benefits research and development by accelerating drug discovery and improving healthcare outcomes. Despite limitations, NLP's role in healthcare data analysis is poised to grow, benefiting both providers and consumers. This study recommends proactive use of NLP tools to enhance care delivery and patient experience.
Key words: Natural Language Processing, Context-aware Algorithms, Quality Improvement Framework, Text Classification, Electronic Health Records, Machine Learning, Word2Vec, Sentiment Analysis, Word2Vec, Hot Word List, Frequency Distribution, Patient Feedback.