Comparing Self-Reported Sleep Quality and Wearable-Derived Sleep Metrics in Middle-Aged and Older Adults with Chronic Pain: A Psychometric Study

Soamy Montesino-Goicolea¹Pedro A. Valdes-Hernandez²Olga Nin³Cameron Smith³Eric C Porges⁴Yenisel Cruz-Almeida^1*

• ¹Department of Community Dentistry and Behavioral Science, University of Florida, Gainesville, United States
• ²Department of Health Outcomes & Biomedical Informatics, University of Florida, Gainesville, United States
• ³Department of Anesthesiology, University of Florida, Gainesville, United States
• ⁴Department of Clinical and Health Psychology, University of Florida, Gainesville, United States

Objectives Our primary aim was to evaluate the agreement between subjective and objective methods of measuring sleep quality in people with musculoskeletal pain and poor sleep. Secondly, we aimed to explore the relationship between subjective and objective sleep quality with clinical and experimental pain as well as its impact on function. Methods Participants with musculoskeletal pain (intensity>5/10 most days in past 3 months) and poor sleep (PSQI total>5) (n=33), completed the Pittsburgh Sleep Quality Index (PSQI) and wore a wearable ring that characterizes sleep stages (i.e., Oura© ring). The equivalent of PSQI's answers over the last month were compared to the average appropriate Oura measure using zero-order correlations (primary aim). Partial Pearson correlations were used to assess sleep-pain relationships (second aim)—controlling for age and sex. Statistical significance was set at α < 0.05 with Bonferroni correction. Results PSQI responses for total bedtime (p < 0.0005), sleep duration (p < 0.0005), and the PSQI duration component (p < 0.003) significantly correlated with their Oura-derived equivalents, whereas Oura measures of sleep latency, efficiency, and disturbances showed no alignment with PSQI metrics. PSQI total score and its sleep latency component were significantly associated with pain measures, including WOMAC pain (p = 0.022; latency p = 0.009), MPQ (total p = 0.026; latency p = 0.008; neuropathic p = 0.026, latency p = 0.011; continuous p = 0.026; intermittent p = 0.026, latency p = 0.008; affective—latency p = 0.008), and GCPS pain intensity (p = 0.026; latency p = 0.012) as well as interference (latency p = 0.008). By contrast, Oura's sleep measures showed no significant associations with pain, except for sleep latency, which correlated with conditioned pain modulation (p = 0.049). All p values are Bonferroni-corrected. Conclusions This preliminary study provides valuable insights into the complementary roles of subjective and objective sleep assessments in older adults with chronic pain. The findings underscore the importance of integrating both approaches to refine sleep evaluation in individuals with musculoskeletal pain. Future research should further examine the feasibility and clinical utility of combining subjective and objective assessments to enhance understanding of sleep-and pain-related health outcomes.