Validation of symptom clusters in patients with metastatic bone pain

PURPOSE Symptom clusters (scs) are a dynamic construct. They consist of at least 2 or 3 interrelated symptoms that may be a significant predictor of patient morbidity. In a previous study, we identified 2 scs in patients with bone metastases: An activity-related interference cluster, psychology-related interference cluster. These scs may be clinically important in the pain and symptom management of patients with metastatic bone pain. It is therefore important to validate the reported scs to determine if they hold true across similar patient populations. PATIENTS AND METHODS From February to September 2007, our study accrued 52 patients with bone metastases [29 men (56%), 23 women (44%); median age: 68.5 years (range: 39-87 years)] who were referred for palliative radiotherapy (rt). Prostate (31%), breast (29%), and lung (19%) were the most common primary cancer sites. Treatment arms ranged from single to multiple fractions, with most patients receiving a single 8-Gy fraction (77%) or 20 Gy in 5 fractions (21%). The most prevalent sites for rt were spine (42%), hips (17%), and pelvis (14%). Worst pain at the site of rt and functional interference scores were assessed using the Brief Pain Inventory (BPI), a multidimensional pain instrument that uses 11-point numeric rating scales. Patients provided their symptom severity scores on the BPI at baseline and at 4, 8, and 12 weeks post rt. At all time points, a principal component analysis with varimax rotation was performed on 8 items (worst pain and 7 functional interference items) to determine relationships between symptoms before and after rt for bone pain. RESULTS Two scs were identified. Cluster 1 included worst pain and interference with general activity, normal work, and walking ability; cluster 2 consisted of interference with mood, sleep, enjoyment of life, and relations with others. Our statistical analysis produced varied results for the 2 clusters found in our previous investigation. These differences may be an indicator for the instability of scs or may be a result of the fewer number of patients accrued in the present validation study. CONCLUSIONS The scs in our two studies were not identical for patients receiving palliative rt for symptomatic bone metastases. Another sc validation study should be conducted with a larger sample before a conclusion is drawn about the existence of an unstable phenomenon in sc research.