Latent class analysis differentiation of adjustment disorder and demoralization, more severe depressive and anxiety disorders, and somatic symptoms in patients with cancer
Published online on June 05, 2018
Abstract
---
- |2+
Abstract
Objectives
Demoralization as a form of existential distress involves poor coping, low morale, hopelessness, helplessness, and meaninglessness. In a secondary analysis of a cohort of German cancer patients, we aimed to explore latent class structure to assess the contribution that symptoms of demoralization make to anhedonic depression, anxiety, adjustment, and somatic disorders.
Methods
Measures of demoralization, depression, anxiety, physical symptoms, and functional impairment had been completed cross‐sectionally by 1527 patients with early or advanced cancer. Latent class analysis used maximum likelihood techniques to define the unobserved latent constructs that can be predicted as symptom clusters. Individual patients were assigned to the most probable class. Classes were compared on demographics, and logistic regression assessed the odds of individual items predicting each class.
Results
A 4‐class model provided the best fit. Class 1 (n = 829, 54.3%) was defined by the absence of distress; Classes 2 to 4 all carried functional impairment. Class 2 (n = 333, 21.8%) was differentiated by somatic symptoms (sleep, tiredness, and appetite); Class 3 (n = 163, 10.7%) by anhedonia, anxiety, and severe demoralization; and Class 4 (n = 202, 13.2%) by adjustment and moderate demoralization. Members of Class 3 were more likely to be younger, female, anhedonic, depressed, and anxious. In both Classes 3 and 4, functional impairment, physical symptom burden, and suicidal ideation were present.
Conclusions
In contrast with the severe symptom cluster carrying anhedonia, anxiety, and demoralization, the moderate symptom cluster was formed by patients with demoralization and impaired functioning, a clinical picture consistent with a unidimensional model of adjustment disorder.
- Psycho-Oncology, EarlyView.