Assessment of Oral Health-Related Quality of Life Using Natural Language Understanding (NLU)

Objectives: The objective of this study was to assess the impact of oral health conditions on patients' quality of life using NLU and qualitative research methods.
Methods: Semi-structured interviews were conducted with 10 participants aged 16–25 years suffering from hypodontia. The interview transcripts were analysed using IBM Watson NLU Text Analysis. Automated analysis produced weighted valence sentiment (positive, negative) and emotions (joy, sadness, anger, fear, and disgust) for the specific questions and keywords identified in the transcripts.
Results: NLU analysis of the interview transcripts showed a predominantly negative sentiment towards hypodontia and its management; 93.2% of the entities identified by the NLU in the transcripts had a negative sentiment, while only 6.8 % had a positive sentiment.
NLU analysis revealed that patient sentiment during the interview correlated inversely with age(R=-0.49), treatment waiting time(R=-0.22) and OHIP score(R=-20). Negative sentiment and sadness was strongest when patients were asked about the history of their dental problems and their feelings about their teeth, but they expressed joy and positive sentiment when asked about the successful dental work provided.
NLU keyword analysis revealed that the keywords associated with negative sentiment were mainly those related to treatment length and treatment delays.
Conclusions: NLU could detect patient’s negative sentiments towards oral health conditions, and it could be helpful in qualitative research in dentistry.