Natural Language Understanding to Assess Oral Health-Related Quality of Life: A Cross-Sectional Study Incorporating a Mixed Methods Approach

Abstract

Background: Natural language understanding (NLU), a subfield of artificial intelligence, focuses on the computational understanding of human language. This technology offers an objective and quantitative approach to analysing interviews in qualitative research. This study hypothesises that NLU can assess the impact of oral health on quality of life by analysing semi-structured interviews. Objective: This study aimed to assess the utility of NLU in evaluating oral health-related quality of life by analysing semi-structured interviews with individuals diagnosed with hypodontia. Methods: A cross-sectional qualitative study was conducted on 10 participants (aged 16–25 years) suffering from hypodontia. Semi-structured interviews were transcribed and analysed using IBM Watson NLU text analysis. The analysis identified entities, keywords, sentiments (positive and negative) and emotions (joy, sadness, anger, fear and disgust) expressed in the interviews. Results: NLU analysis revealed a predominantly negative sentiment towards hypodontia and its management, with 93.2% of identified entities presenting a negative sentiment and only 6.8% showing a positive sentiment. Patient sentiment correlated inversely with age (R = −0.49), treatment waiting time (R = −0.22) and OHIP score (R = −20). Negative sentiments and sadness were most prominent when discussing the history of dental problems and feelings about their teeth, whereas joy and positive sentiments were expressed regarding successful dental work. Keywords associated with negative sentiment were primarily related to treatment length and delays. Conclusion: NLU effectively identified patients' negative sentiments and emotional responses to oral health conditions, demonstrating its potential as a valuable tool in qualitative dental research.