Elsevier

General Hospital Psychiatry

Volume 36, Issue 1, January–February 2014, Pages 74-80
General Hospital Psychiatry

Psychiatric–Medical Comorbidity
Comparison of the Hospital Anxiety and Depression Scale and the Center for Epidemiological Studies Depression Scale for detecting depression in women with breast or gynecologic cancer

https://doi.org/10.1016/j.genhosppsych.2013.08.010Get rights and content

Abstract

Objective

Depression is common in cancer patients but frequently undetected. Consensus regarding validity and optimal thresholds of screening measures is lacking. We investigated the validity of the Hospital Anxiety and Depression Scale (HADS-D) and Center for Epidemiological Studies Depression Scale (CES-D) relative to a referent diagnostic standard in women with breast or gynecologic cancer.

Method

Participants were 100 patients who completed the CES-D and HADS-D within a larger study. The Mini International Neuropsychiatric Interview was the criterion standard. Sensitivity, specificity, predictive values and likelihood ratios for various thresholds were calculated using receiver operating characteristics. Participants were assigned to two diagnostic groups: ‘major depressive disorder’ or ‘any depressive disorder’.

Results

Separate analyses were conducted whereby participants found to be receiving depression/anxiety treatment at the time of validation (n= 28) were excluded. Both measures had good internal consistency and criterion validity. There were no statistical differences in global accuracy between the measures for detecting either group. For optimal sensitivity and specificity in both groups, generally recommended thresholds were lowered for the HADS-D. For the CES-D, the threshold was lowered for ‘any depressive disorder’ and raised for ‘major depressive disorder’. Negative predictive values associated with our recommended cutoffs were excellent, but positive predictive values were poor.

Conclusions

The HADS-D and CES-D have acceptable properties and are equivalent for detecting depression in this population. Depending on the purpose of screening, the CES-D may be more suitable for identifying major depression. Threshold choice may have serious implications for screening program effectiveness, and the use of generally recommended thresholds should be cautious.

Introduction

Depression is common in cancer patients. Reported prevalence rates vary, but a frequently cited estimate is that 20%–30% of cancer patients experience clinically significant depressive symptoms [1]. Depression is associated with deleterious effects across a range of outcomes including quality of life [2], [3], [4], health service utilization and health care costs [4], [5], and even survival [6], an association potentially mediated by poorer adherence to anticancer treatments [7]. Indeed, results of a large meta-analysis [8] showed a 39% higher mortality rate for cancer patients with minor or major depression.

Despite its importance, the ability of oncology professionals to reliably detect depression is undisputedly poor [9], [10], [11]. Obstacles to appropriate recognition of depression in cancer settings include lack of clinician time, poor communication skills, nihilism about depression and its treatment, and unfamiliarity with assessment methods [12]. Recorded discussions between patients and oncologists showed that discussions of emotional health featured in less than one third of consultations [13]. Low concordance between patients' self-report and oncologists' clinical impression was clearly demonstrated in a study of 143 oncologists and 2297 patients across 34 cancer centers in the United Kingdom in which, compared to results of a validated screening inventory, clinician sensitivity and specificity were 29% and 85%, respectively [9].

The need for easy, accurate methods of identifying depressed individuals who may benefit from referral to specialist psychological services is obvious. Consequently, a large number of screening tools have been developed and used in cancer populations. Cutoff scores designating clinically significant depressive symptoms are typically recommended for each instrument, yet there is lack of consensus regarding the validity, optimal thresholds and relative merits of these measures [14], [15], [16]. The validity of a screening instrument is the extent to which it actually measures the specified construct (i.e., depression). When screening for depression in cancer patients, an appropriate balance must be struck between detection sensitivity and specificity, or the screening program may be rendered ineffective. An insufficiently sensitive instrument may fail to identify a substantial number of depressed patients in need of psychological intervention, resulting in aversive outcomes as outlined above, but when instrument specificity is low, resources may be unnecessarily expended on patients who have mistakenly been referred for specialist psychological services. If patients are offered services they do not need, uptake of these referrals may be lower than expected and clinical resources wasted, potentially resulting in funding cuts to psycho-oncology services [12].

Two measures of particular interest for detecting depression are the Hospital Anxiety and Depression Scale (HADS) [17] and the Center for Epidemiological Studies Depression Scale (CES-D) [18]. The HADS, which comprises two 7-item subscales for measuring anxiety and depression (HADS-D), respectively, is the most well-known and widely used screening measure, particularly in Europe. Developed for use in medically ill populations, the HADS-D purposively does not contain any somatic items to avoid confounding psychological symptoms with those of disease or treatment. The emphasis of the content is on anhedonia. While there is a wealth of evidence attesting to its psychometric properties [19], there is controversy regarding the HADS-D's discriminant validity relative to structured clinical assessment [20]. There is also lack of consensus as to the optimal cutoff score to detect major or minor depression among cancer patients [16], [21]. Reviews of the literature show that, in 2009, the most commonly used HADS-D threshold for determining caseness in the empirical literature was ≥ 8 [21]. This finding is poorly supported by validation studies that recommend cutoff scores ranging from ≥ 5 to ≥ 11 [20], [21].

By comparison, the 20-item CES-D, developed for use in nonpsychiatric community samples and more popular in North America, includes somatic items, and its content is biased towards experience of dysphoria. The CES-D has adequate psychometric properties [16] and has performed consistently well in depression screening studies and as an outcome in randomized trials of cancer patients [14], but its criterion validity in oncology settings has previously been evaluated in only two small studies of patients with mixed cancer diagnoses (n= 33) [22] and head and neck cancer (n= 60) [23].

In this paper, we describe a head-to-head comparison of the HADS-D and CES-D in detecting depression in patients with breast or gynecologic cancer using the Mini International Neuropsychiatric Interview Version 5 (M.I.N.I.) [24], [25] as the referent diagnostic tool. The M.I.N.I. is a validated instrument used to diagnose minor and major depression according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) [26] and is similar to the Structured Clinical Interview for DSM-IV [27] in operation and principle. The HADS, CES-D and M.I.N.I are described in greater detail below (see Methods). To our knowledge, only one previous study [23] has evaluated the relative operating characteristics of the HADS-D and CES-D in cancer patients. Given the differences between these instruments as outlined above and the not insignificant consideration that the CES-D is freely available while use of the HADS attracts substantial cost, this comparison is thought to be useful.

The aims of this study were to (a) investigate internal consistency and intercorrelations of the HADS-D and CES-D, (b) analyze the operating characteristics of the HADS and CES-D according to an independent criterion standard for depressive disorders, (c) determine if either screening instrument is superior for detecting DSM-IV depressive disorders and (d) determine optimum cutoff scores for discriminating between patients with and without depressive disorders.

Section snippets

Participants

The sample comprised women with newly diagnosed breast or gynecologic cancer attending two colocated major teaching hospitals in metropolitan Australia. Women younger than 18 years of age, with ductal carcinoma in situ, with previously diagnosed invasive carcinoma, or with cognitive or physical impairment affecting ability to consent or participate and those with poor English language comprehension were excluded from the study. Participants were recruited in person during outpatient visits

Study participation

As noted above, the 100 participants interviewed in this study were drawn from a larger study involving repeated administrations of the CES-D and HADS-D at 8-week intervals. Consequently, participants in the current series ranged from those at the early stages of the larger study (i.e., with multiple, future administrations still scheduled) to those whose telephonic interview corresponded with the completion of their project participation (i.e., with no future self-report administrations

Discussion

The main aim of this study was to determine the validity of two self-report depression screening instruments relative to a referent diagnostic standard in women with breast and gynecologic cancer. The substantial intercorrelations between the CES-D and HADS-D (r= 0.723) showed the extent to which the scales measure the same construct despite their differing emphases on dysphoria and anhedonia, respectively. Excellent internal consistencies for both instruments were demonstrated.

In its purest

Conclusions

Our results showed that the HADS-D and CES-D were statistically equivalent for detecting major and subthreshold depression in women with breast and gynecologic cancer. Despite comparable global accuracy though and depending on the purpose of screening, the CES-D may be more suitable for identification of major depression. To our knowledge, this study is one of only three CES-D validation studies in oncology patients and one of only two to compare the HADS-D and CES-D with a diagnostic referent

Acknowledgments

This study was supported in part by seed funding from the Western and Central Melbourne Integrated Cancer Service. The Centre for Women's Mental Health is supported by the Pratt Foundation. The authors wish to thank the Breast Care Nurses Robyn Cordner, Monique Baldacchino, Kerry Shanahan, Bronwyn Flanagan and Claire Noonan for their assistance in recruiting patients to the study. We also thank the participating patients for their support of this research.

References (39)

  • B. Lowe et al.

    Comparative validity of three screening questionnaires for DSM-IV depressive disorders and physicians' diagnoses

    J Affect Disord

    (2004)
  • C. Herrmann

    International experiences with the Hospital Anxiety and Depression Scale: a review of validation data and clinical results

    J Psychosom Res

    (1997)
  • E. Frick et al.

    Anxiety, depression and quality of life of cancer patients undergoing radiation therapy: a cross-sectional study in a community hospital outpatient centre

    Eur J Cancer Care (Engl)

    (2007)
  • A. Alacacioglu et al.

    Quality of life, anxiety, and depression in Turkish colorectal cancer patients

    Support Care Cancer

    (2010)
  • L.E. Carlson et al.

    Benefits of psychosocial oncology care: improved quality of life and medical cost offset

    Health Qual Life Outcomes

    (2003)
  • J.S.A. Simpson et al.

    Impact of a group psychosocial intervention on health care utilization by breast cancer patients

    Cancer Pract

    (2001)
  • D. Kissane

    Beyond the psychotherapy and survival debate: the challenge of social disparity, depression and treatment adherence in psychosocial cancer care

    Psychooncology

    (2009)
  • J.R. Satin et al.

    Depression as a predictor of disease progression and mortality in cancer patients: a meta-analysis

    Cancer

    (2009)
  • L. Fallowfield et al.

    Psychiatric morbidity and its recognition by doctors in patients with cancer

    Br J Cancer

    (2001)
  • Cited by (27)

    • Performance of four diagnostic approaches to depression in adults with cancer

      2018, General Hospital Psychiatry
      Citation Excerpt :

      While individual settings may take for granted that one approach has been determined to be optimal, the literature is mixed at best and clear guidelines on which approach to use do not exist. Many cancer settings and psycho-oncology researchers, for example, continue to opt for screening measures that exclude somatic symptoms despite equivocal findings (i.e., the Hospital Anxiety and Depression Scale [5]; [11–13]), and the potential to “miss” patients with clinically significant depressive symptoms. The current study compares three of the most widely recognized approaches to depression assessment among medically ill patients, inclusive, exclusive, and substitutive, in a large sample of ambulatory cancer patients.

    • Predictors of depression and anxiety symptom trajectories in the 24 months following diagnosis of breast or gynaecologic cancer

      2016, Breast
      Citation Excerpt :

      Recruitment occurred over 24 months beginning in August 2008. Recruitment and data collection procedures have been previously described [27–29]. In brief, women completed a self-report questionnaire comprising clinical and demographic data, measures of anxiety and depression symptoms, and a measure of neuroticism at recruitment.

    • Transitions in functional status of community dwelling older adults: Impact of physical performance, depression and cognition

      2016, European Geriatric Medicine
      Citation Excerpt :

      Additionally, although mortality was considered as a decline in FS, SPPB, GS or worsening depression, limiting the analysis only to those survivors in the second wave (n = 485) yielded similar results to those obtained in the overall sample of 607 (see previous sensitivity analysis). Another limitation could be the fact that depression was measured with two different scales in 2008 and 2011, although the HADS-D and CES-D have been shown to have acceptable properties and are equivalent for detecting depression [31,32]. The ability of older persons to be functionally independent in the community is a public health issue, therefore knowledge and understanding of the factors responsible for functional limitations is important so that strategies to prevent or delay these limitations can be developed.

    View all citing articles on Scopus

    Conflict of interest: The authors have no conflict of interest to declare.

    View full text