Depression and survival in acute illnesses
Synopsis: Among 3,529 of the original 9,105 patients enrolled in SUPPORT (Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments) who completed a questionnaire seven to 11 days after admission, depressed mood was associated with reduced survival, greater severity of illness, and worse levels of physical functioning.
Source: Roach MJ. Arch Intern Med 1998; 158: 397-404.
The SUPPORT project has produced a number of reports relating mainly to outcomes prediction and patient preferences for life-sustaining treatments during hospitalization for severe acute illness. This article reports the results of an examination of data from the SUPPORT project relating to depressed mood and its potential relationships to survival and other outcomes.
Of the 9,105 patients enrolled in the SUPPORT project, 3,529 completed the mood assessment portion of a questionnaire seven to 11 days after their initial hospital admission. The 3,529 patients represented those who were still alive: 4,092 had died, were still intubated, or were otherwise unable to communicate, and permission to complete the questionnaire could not be obtained from 1,031. Data were incomplete for 181 patients, and 272 could not be located following hospital discharge.
For the patients reported in this paper, admission diagnoses were acute respiratory failure (29%), congestive heart failure (24%), chronic obstructive pulmonary disease (COPD) (14%), lung cancer (12%), and a variety of other diseases (21%). Forty-two percent of the patients died within 4.5 years of study entry. Compared to the other SUPPORT patients, those who completed the mood questionnaire were younger (61 vs. 64 years old), tended to be men (58% vs. 42%), had lower severity of illness as assessed by Acute Physiology Score at day three (31 vs. 41), and had more comorbid illnesses (2.0 vs. 1.8). However, they were more functional, as assessed by fewer dependencies in activities of daily living (ADL) (1.2 vs. 1.8). All of these differences were statistically significant.
Measured by the Profile of Mood States (POMS) depression scale, the patients had an average score of 0.37 on a scale of 0 to 4. Patients with greater severity of illness and more ADL dependencies had significantly higher depression scores, and a higher estimate of the probability of six-month survival was negatively associated with depressed mood.
Older persons were less depressed than younger ones. Using a stratified Cox proportional hazards model for predicting survival, Roach and colleagues determined that survival time could be predicted (all, P <0.01) by ADL score, the number of comorbid illnesses, the six-month estimate of survival from the SUPPORT model, age, and depressed mood.
Depressed mood had approximately the same predictive effect as ADL dependencies and comorbid illnesses in predicting shortened survival. After adjusting for patient demographics and health status, depressed mood was associated with reduced survival time with a hazards ratio of 1.134 (95% CI, 1.071-1.200; P <0.001).
Comment by David J. Pierson, MD, FACP, FCCP.
This study shows that depressed mood is associated with slightly worse outcomes among patients acutely hospitalized with serious illness. But is there a causal relationship between depression and mortality in such patients, or is depression simply a marker for the severity of illness? As Roach and associates acknowledge, this question is at the heart of the issue addressed in this study.
Other investigators have demonstrated an association between depression and mortality, but the SUPPORT project has the potential advantage of inclusion of large numbers of patients who are demographically and medically diverse. Thus, the authors' findings, while relatively small in magnitude of effect, have statistical strength and possibly greater applicability to the general population of seriously ill patients.
Keep in mind the study's limitations
The study has several limitations, as discussed by Roach. Of all the patients enrolled in the SUPPORT project, only those who were alive and physically capable of responding to the questionnaire after a week of hospitalization and agreed to complete it, were included.
Whether data from more severely ill individuals would have produced the same results is unknown. Roach and his colleagues did not perform a formal psychiatric evaluation on the patients, so their findings with respect to depressed mood may or may not correlate with the clinical diagnosis of depression. In addition, as in any study demonstrating a statistical association, the researchers are unable to state whether the development of depressed mood was causally related to increased mortality.
Whether therapeutic intervention aimed at improving seriously ill patients' moods would have an effect on mortality is unknown. Nonetheless, this study makes a better case than anything previously available for a relationship between depressed mood and poor outcomes in severe acute illness and suggests that clinicians need to be aware of their patients' moods even in the absence of overt depression. Further studies should address the potential effect of interventions to decrease depression on survival and other outcomes in critically ill patients.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content