Comparing depressive symptom representation between individuals with cancer, non-cancer controls and healthy...

Individuals affected by cancer are vulnerable to depression, but prevalence rates vary greatly between studies. The preponderance of somatic rather than psychological symptoms, overlapping with cancer-related symptoms, has spurred controversy on depressive symptom assessment in cancer. This work moves beyond sum scores of symptoms and explores symptom interrelations in individuals affected by cancer (compared to non-cancer and healthy controls) to deepen the understanding of depressive symptom expression in the context of cancer. It also compares the symptom network of individuals with cancer to a) non-cancer and b) healthy controls.

Network analyses and comparisons were conducted on 3512 participants from the Gutenberg-Health-Study (1230 with cancer; 1230 non-cancer controls; 1230 healthy individuals). Gaussian graphical models were estimated for each group, and network structures were compared regarding symptom predictability and centrality of PHQ-9 items.

Depressive symptoms were more frequent in individuals affected by cancer than in non-cancer controls (d = 0.131) and healthy individuals (d = 0.172), but network structures did not differ. Mean predictability (amount of symptom variance explained through all other symptoms) ranged between 25 % and 29 %; with depressed mood highest in individuals with cancer and energy loss highest in non-cancer and healthy controls. Both symptoms were the most central nodes in all groups.

Symptom structure did not differ between individuals affected by cancer, non-cancer controls and healthy individuals. While somatic symptoms were more frequent in individuals with cancer, they were as closely associated with other symptoms as in the control populations. Along with psychological symptoms, somatic symptoms should be considered in screening as well as targeted by interventions in those with cancer.