Symptom Clustering in Advanced Cancer
Walsh, Declan, Cleveland Clinic Foundation, USA, Rybicki, Lisa, Cleveland Clinic Foundation, USA

Multiple symptoms are common and often severe in patients with advanced cancer. Symptom profile appears to be influenced by demographic characteristics and may be helpful in predicting prognosis. We conducted a cluster analysis of symptoms collected prospectively amongst 1000 patients with advanced cancer referred to a Palliative Medicine Program. Eight symptom clusters were identified as showing high absolute correlation. Gastrointestinal symptoms were important in the clusters. Depression did not cluster with fatigue, pain with depression, nor early satiety with nausea. Recognition of symptom clustering may be important in understanding the natural history of cancer, the pathophysiology of advanced cancer, and improving symptom control.