Subjective experiences of post-sleep activity and vigilance were the strongest predictors of clinical pain severity, rather than physiologically measured sleep quality, with sleep profiles increasing explained variance in pain from 32% to 35%.
Sociodemographic, clinical, and anthropometric factors alone explained 32% of the variance in clinical pain severity using Gradient Boosted Regression.
Inclusion of sleep profiles increased the explained variance in clinical pain severity from 32% to 35%.
Subjective experiences of post-sleep activity and vigilance were the strongest predictors of clinical pain severity.
Physiological sleep parameters measured by polysomnography were weaker predictors of clinical pain severity compared to subjective sleep quality measures.
K-means clustering analysis independently confirmed that subjective sleep quality was more strongly associated with clinical pain severity than physiological sleep measures.
The study was a secondary analysis of cross-sectional data from the Cleveland Family Study examining sleep-pain relationships in 556 adults.
This research suggests that how people subjectively feel about their sleep quality — particularly how rested, active, and alert they feel after sleeping — is a stronger predictor of clinical pain severity than what objective sleep lab measurements show. The study analyzed data from 556 adults who underwent overnight sleep studies (polysomnography) and also completed questionnaires about how their sleep affected their daily functioning. Using several machine learning approaches, the researchers found that factors like how well people can perform activities and stay alert after sleep predicted pain levels more strongly than physical measurements of sleep like sleep stages or breathing patterns. The study found that basic personal and health information (such as demographics and clinical characteristics) explained about 32% of the differences in pain severity between individuals, and adding sleep information bumped this to 35%. While this increase may seem modest, the key insight is that the subjective sleep quality measures drove this improvement more than the objective polysomnography measurements. This pattern was consistent across multiple different analytical methods, including a separate clustering analysis, lending confidence to the finding. This research suggests that there may be value in focusing on how people perceive and experience their sleep, rather than solely on measurable physiological sleep parameters, when studying or treating pain conditions. It raises the possibility that interventions aimed at improving people's subjective sense of sleep quality — how refreshed and functional they feel after sleeping — could potentially have a meaningful impact on their experience of pain, though the cross-sectional design of this study means cause and effect cannot be determined from these data alone.
Raghuraman N, Duan E, Massalee R, Wang Y. (2026). Subjective sleep quality predict clinical pain severity more strongly than polysomnographic parameters: Machine learning findings from a cross-sectional study.. Sleep medicine. https://doi.org/10.1016/j.sleep.2025.108753