Results
K-prototypes cluster analysis of integrated circadian, physiological, and psychosocial data identified four distinct subgroups among 191 inpatients with cancer pain.
- Explanatory sequential mixed-methods design with a quantitative phase followed by semi-structured interviews analyzed thematically.
- Participants were 191 adult inpatients from a tertiary care hospital.
- Measures included 24-h blood pressure, salivary melatonin, pain beliefs, coping styles, social support, and clinical variables.
- Cluster sizes were 44 (Cluster 0), 54 (Cluster 1), 74 (Cluster 2), and 19 (Cluster 3).
Results
Cluster 0 (elderly reverse-dipper, n = 44) had moderate pain, poor sleep quality, emotional lability, and social withdrawal, and showed no significant analgesic effect on either outcome.
- Mean pain intensity: coefficient = 0.063, 95% CI: -0.585 to 0.711, p = 0.856.
- Pain variability: coefficient = -1.466, 95% CI: -4.178 to 1.246, p = 0.335.
Results
Cluster 1 (non-dipper severe pain, n = 54) had disrupted circadian rhythm, negative pain cognition, and avoidant coping; analgesics reduced pain variability but not mean pain intensity.
- Pain variability: coefficient = -2.333, 95% CI: -4.001 to -0.665, p = 0.008.
- Mean pain intensity: coefficient = 0.072, 95% CI: -0.300 to 0.444, p = 0.712.
- Analgesic efficacy was evaluated using multiple linear regression.
Results
Cluster 2 (dipper mild pain, n = 74) maintained normal circadian rhythm, positive mental status, and active social interaction; analgesics reduced pain variability but not mean pain intensity.
- Pain variability: coefficient = -3.427, 95% CI: -5.391 to -1.463, p = 0.002.
- Mean pain intensity: coefficient = -0.248, 95% CI: -0.682 to 0.186, p = 0.274.
Results
Cluster 3 (high-melatonin, n = 19) was the only subgroup in which analgesics significantly reduced both mean pain intensity and pain variability, and it was characterized by passive coping and emotional suppression.
- Mean pain intensity: coefficient = -1.259, 95% CI: -2.031 to -0.487, p = 0.001.
- Pain variability: coefficient = -5.602, 95% CI: -8.664 to -2.540, p = 0.001.
- This was the smallest cluster (n = 19), so estimates are based on a small sample.
Results
Across subgroups, analgesics more consistently affected pain variability than mean pain intensity.
- Pain variability significantly decreased in Clusters 1, 2, and 3 but not Cluster 0.
- Mean pain intensity decreased significantly only in Cluster 3.
Conclusions
The authors conclude that cancer pain is heterogeneous in circadian, endocrine, and psychosocial profiles, and that interactions among these factors may contribute to distinct pain mechanisms and variable analgesic responses.
- Comprehensive multidimensional assessment may facilitate more accurate patient subtyping and support individualized pain management.
- The study was registered with ChiCTR (ChiCTR2500096806).