Is asking patients in palliative care, “Are you depressed?” appropriate? Prospective study

Introduction

Depression is a serious problem that affects about a quarter of patients in palliative care.1 Diagnosing depression is difficult because most patients are understandably sad. Lack of recognition of depression can lead to further morbidity and a detrimental effect on the quality of life.2

Screening for depression attracts wide interest, but as patients are frail and unwell, a screening tool needs to be brief. Asking North American patients in palliative care the single question, “Are you depressed?” had perfect sensitivity and specificity and 100% positive predictive value compared with the schedule for affective disorders.3 Other studies of patients with stroke and elderly patients found that a single statement with a “yes” or “no” response correctly classified more than 80% of people as depressed or not depressed.4 5

Participants, methods, and results

Patients were eligible to participate if they received only palliative and supportive day care; 106 were eligible to participate in the study during six months. A total of 74 patients consented to participate; we got baseline demographic information on age, ethnicity, diagnosis, past history of depression, and performance status (all patients had an Eastern Cooperative Oncology Group (ECOG) performance of 2 or 3—that is, physical function was limited). Age, sex, disease state, and performance status did not differ between consenters and non-consenters. Patients were aged 28-89 (mean 68) years.

We compared the result of the single question, “Are you depressed?” with a semistructured clinical psychiatric interview based on the criteria of the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders. We calculated the sensitivity, specificity, and the positive and negative predictive values.

We found that 27% of patients had depression (95% confidence interval 17% to 37%). To the single question, “Are you depressed?” 25 patients responded “yes”; in the clinical interview we found that 11 were depressed. Nine patients who replied “no” to the question were considered depressed by clinical interview (table). A yes answer had a sensitivity of 55% (34% to 77%) and specificity of 74% (61% to 84%) and positive and negative values of 44% and 82% respectively.

Comment

The single question, “Are you depressed?” is widely believed to successfully screen for depression in palliative care—many clinicians use it routinely—but it does not have the perfect sensitivity and specificity in the UK population that it achieved in North America.3

Although this is a small study, it is larger than many other studies concerning patients in palliative care. A quarter of patients had depression. The single question in this study correctly identified just over half of the patients diagnosed as depressed and approximately three quarters of the patients diagnosed as non-depressed.

In the North American study, the researchers included the single question, “Are you depressed?” in the schedule of the interview, which may have biased the results. Our results are also less favourable than those found in a study of stroke patients and elderly patients which compared the discriminating power of the question with an existing scale and not a semistructured interview.4 5 Such a subjective question requires further rigorous evaluation before we advocate its widespread application to screen for depression in patients with advanced cancer.