The diagnosis and treatment of depression are complicated by the presence of a serious medical illness, such as cancer. The role of inflammation in the pathophysiology of depression remains unknown; however, the symptom cluster of sickness behavior is more clearly related to inflammation. Symptom clusters may provide the specificity needed to improve treatment outcomes. The purpose of this study is to use confirmatory factor analysis to examine the construct of sickness behavior in patients with advanced cancer using the Beck Depression Inventory–II (BDI–II).

Patients (N = 167) were recruited from chemotherapy clinics and the Department of Psychiatry at Memorial Sloan Kettering Cancer Center. Symptoms were a priori delineated using the factor labels sickness behavior and negative affectivity (two-factor model), and affective, cognitive, and sickness behavior (three-factor model). These data were also fit for a more traditional model using affective, cognitive, and somatic factors.

The mean total BDI–II score was 14.74 (SD = 8.52; range = 1–46). Fit statistics for all models were good, but the novel three-factor model with sickness behavior provided the best fit: χ2(186) = 273.624; p < 0.001; root-mean-square error of approximation = 0.053; comparative fit index = 0.949.

Both the two- and three-factor models provide support for the unique construct of sickness behavior in patients with advanced cancer. High factor correlations with the BDI–II and other measures of distress were observed, which raises questions about theoretically distinct, but related, constructs.