Purpose: The primary objective was to explore how patients' worst pain clustered together with functional interference items. Secondary objectives were to determine whether symptom clusters change with palliative radiotherapy (RT) and to compare the difference between responders and nonresponders to radiation.
Materials/methods: Worst pain at the site of treatment and functional interference scores were assessed using the Brief Pain Inventory (BPI). Patients provided their scores at baseline, 4, 8, and 12 weeks post-RT. A principal component analysis was performed on the 8 items (worst pain and 7 functional interference items) at all time points to determine interrelationships between symptoms. Principal components with an eigenvalue higher than 0.90 and explaining more than 10% of the variance were selected. The Cronbach alpha statistic was used to estimate the internal consistency and reliability of the derived clusters at baseline and at subsequent follow-ups. Robust relationship and correlation among symptoms were displayed with a biplot graphic.
Results: From May 2003 to January 2007, 348 patients with bone metastases that were referred for palliative RT were accrued into the study. There were 206 males (59%) and 142 females (41%), with a median age of 68 years (range, 30-91). Lung (26%), breast (25%) and prostate (24%) were the most common primary cancer sites. Treatment ranged from single to multiple fractions, with the majority of patients receiving a single 8 Gy (58%) and 20 Gy/5 (35%). The most prevalent sites of RT were spine (31%), pelvis (16%), and hips (15%). Two symptom clusters were identified. Cluster 1 included walking ability, general activity, normal work, enjoyment of life and worst pain. Cluster 2 included relations with others, mood and sleep. The two clusters at baseline accounted for 67% of the total variance with a Cronbach alpha of 0.87 and 0.70, respectively. In responders to radiation treatment, the two symptom clusters disintegrated at 4, 8, and 12 weeks post-RT. All symptom severity items improved over time (p < 0.0001). In nonresponders, two clusters had disappeared at week 4, reemerged at week 8, and disintegrated at week 12.
Conclusion: Symptom clustering has proved to be therapeutically important because treatment of one symptom may affect others within the same cluster. The significant correlations between worst pain and the functional interference items reaffirm the importance of pain reduction as a treatment goal for palliative radiotherapy. By treating a patient's symptom of worst pain, it would subsequently ease their response burden on their daily functional activities by decreasing symptom severity, increasing function, and improving overall quality of life.