Purpose
Estimate current pain intensity from personal narratives of the chronic pain experience and explore the linguistic differences that allow for it.
Methods
Chronic pain patients were interviewed, and demographic and clinical data were collected. Patients reported their current pain intensity on a Visual Analogue Scale (VAS), which was discretized into 3 classes: mild, moderate, and severe pain. Language features were extracted from the transcribed interview of each patient and used to classify their pain intensity category in a Leave One Out Validation setting. Performance was measured using the weighted F1 score. Possibly confounding variables were analyzed for internal validity.
Results
65 patients (40 females), averaging 56.4 ± 12.7 years of age, participated in the study. The best performing model was the Support Vector Machine with an Early Fusion of select language features, with an F1 of .60, improving 39.5% upon the baseline. Patients with mild pain focus more on the use of verbs, whilst moderate and severe pain patients focus on adverbs, and nouns and adjectives, respectively.
Conclusion
Pain intensity estimation is commonly based on facial expressions and various bio-signals, but language does not seem to have been previously used. We demonstrated a proof-of-concept for the analysis of the language of chronic pain in that context, and, importantly, that focus on specific words/themes is especially correlated with specific pain intensity categories. We show that language features from patient narratives indeed convey information relevant for pain intensity estimation, and that our computational models can take advantage of that.