Identification of symptom clusters by occurrence and severity in lung cancer patients receiving chemotherapy.

e21657 Background: Lung cancer patients experience a high prevalence and severity of symptoms, which often co-occur to form symptom clusters. Symptom clusters can be identified using different measures such as occurrence or severity but the optimal method is not known. We aimed to compare symptom clusters identified using occurrence versus severity among lung cancer patients receiving chemotherapy. Methods: As part of a larger cohort study of cancer patients receiving chemotherapy, lung cancer patients self-reported symptom occurrence and severity 1 week after receipt of chemotherapy using the Memorial Symptom Assessment Scale. We characterized the occurrence and severity of symptoms and performed separate exploratory factor analyses (EFA, unweighted least squares extraction method, geomin oblique rotation) to identify symptom clusters using occurrence versus severity ratings (0 = not present to 4 = very severe). Results: 145 lung cancer patients completed the symptom assessment (57% female; mean age 64 years; 88% non-small cell lung cancer, 12% small cell). Patients experienced a mean of 14 symptoms (SD 7.1). The most prevalent symptoms were lack of energy (79%), feeling drowsy (66%), difficulty sleeping (66%), and pain (61%). Among only patients with each symptom, the most severe symptoms were problems with sexual interest or activity (mean severity rating 2.43), lack of energy (mean 2.38), constipation (mean 2.27), and hair loss (mean 2.21). EFA using symptom occurrence identified 5 symptom clusters: sickness behavior (which included physical symptoms), lung cancer, psychological, nutrition, and epithelial symptoms. EFA using symptom severity identified 5 similar clusters but the sickness behavior cluster included both physical and psychological symptoms. Conclusions: EFA using symptom occurrence versus severity identified the same number of symptom clusters with similar themes. However, the symptoms within each cluster were more comprehensive when symptom severity was used. The presence of physical and psychological symptoms in the sickness behavior cluster identified using symptom severity highlights the importance of addressing these symptoms together.