Elsevier

General Hospital Psychiatry

Volume 28, Issue 1, January–February 2006, Pages 55-58
General Hospital Psychiatry

Psychiatric–Medical Comorbidity
Validity and reliability of the Hospital Anxiety and Depression Scale in a hypertrophic cardiomyopathy clinic: the HADS in a cardiomyopathy population

https://doi.org/10.1016/j.genhosppsych.2005.08.004Get rights and content

Abstract

Objective

The purpose of this study is the validation of the Hospital Anxiety and Depression Scale (HADS) in patients suffering from hypertrophic cardiomyopathy in an inner city teaching hospital. The secondary objective was to establish whether the use of the total HADS score to detect “caseness” is justified.

Methods

One hundred fifteen patients in a cardiac outpatient clinic completed the HADS, which was compared against the gold standard Structured Clinical Interview for DSM-III-R nonpatient version (SCID-np). Receiver operating characteristic (ROC) curves were created for the anxiety and depression subscales, as well as the total score, then sensitivity, specificity, likelihood ratios and positive and negative predictive values were calculated. Retest correlation was assessed at 2 weeks in 31 patients.

Results

The depression subscale was highly sensitive (100%) when the cutoff score 8 was used; however, the corresponding specificity was poor (79%). Raising the cutoff to 10 improved specificity but compromised sensitivity. The anxiety subscale was performed similarly though with less accuracy. The total HADS score produced a poor ROC curve and performed best when the cutoff was 14.

Conclusions

The HADS performs well as a screening instrument for anxiety and depression in this population at the designed cutoff score 8. However, its use as a research instrument and the practice of using the total score to detect caseness are not supported by this study.

Introduction

Anxiety disorder and depression frequently occur in the physically ill. Some studies postulate that depression has a prevalence of between 20% and 50% [1], [2] in the general hospital setting. Alternatively, this may be an artifact of the screening process. Early studies investigating the prevalence of psychiatric disorder in medical patients used scales that include somatic symptoms. These scales, therefore, lose their ability to distinguish depression from malaise in the physically ill.

In 1983, Zigmond and Snaith [3] developed the Hospital Anxiety and Depression Scale (HADS) for use as a screening questionnaire in medical outpatient clinics. The HADS is a 14-item self-report questionnaire incorporating anxiety and depression subscales. Each item is scored 0 to 3, and a score of 8 or greater on one or both of the subscales indicates the presence of a depressive or anxiety disorder. Completion of the questionnaire takes only a few minutes. The developers aimed to discard all ambiguous somatic symptoms such as dizziness and lethargy and instead comprised the depression subscale around the psychopathology of anhedonia and the anxiety subscale based upon the cognitive symptoms of anxiety. The HADS has been found to perform reliably in psychiatric, nonpsychiatric and well populations [4] as a screening tool; however, its use as a diagnostic instrument for research may be inappropriate [5]. A low cutoff point on the depression subscale provides adequate sensitivity, but the corresponding positive predictive value (PPV), the probability that a subject who tests positive is a true positive, is disappointingly low. A satisfactory balance cannot be achieved by modifying the cutoff point on this subscale [6], and so, it has been recommended that “caseness,” the presence of depression, anxiety or both, is better defined by the total HADS score [7].

The tripartite model of depression and anxiety, which divides symptoms into three groups [8], poses a further problem for the HADS. According to this model, low positive affectivity or anhedonia characterizes depression, anxiety states are characterized by physiological hyperarousal (PH) and both are nonspecifically associated with general negative affectivity. Hence, the somatic symptoms of PH that the HADS was designed to exclude are those that differentiate anxiety from depression. Nevertheless, the HADS is a simple, self-rating scale that patients find acceptable [9], its use has been validated across a number of different languages [10], [11] and settings [12], [13], and perhaps, more importantly, its widespread use has made it familiar and acceptable to clinicians and researchers alike.

Patients suffering from hypertrophic cardiomyopathy experience raised levels of anxiety, depression and psychosocial dysfunction compared with population norms. These consequences are particularly common in those with chest pain and dyspnea, but are not associated with a family history of hypertrophic obstructive cardiomyopathy (HOCM) or premature sudden death [14]. This study used the HADS as a diagnostic instrument, though it has not yet been validated in this population. The aim of the current study is to assess the performance of the HADS in a clinical population with hypertrophic cardiomyopathy.

Section snippets

Methods

One hundred fifteen consecutive English-speaking patients from a sample of 134 (86% response rate) at a specialist Hypertrophic Cardiomyopathy clinic at St. George's Hospital and Medical School in London agreed to complete the study. The participants first completed the HADS questionnaire, and then one trained researcher (JM) who was blind to the subjects' HADS score performed the Structured Clinical Interview for DSM-III-R nonpatient version [15] (SCID-np), which was considered to be the gold

Results

The sample was comprised of 68 males and 47 females with a median age of 43 years ranging from 23 to 63 years. The prevalence of depression as defined by the SCID-np in this population was 21%, major depression accounting for 75% of this, and anxiety disorders occurred in 39%. Both conditions coexisted in 12% of the population. The mean score on the depression subscale was 6.2 with an S.D. of 3.15. The mean score on the anxiety scale was 7.89 with an S.D. of 2.84, whereas the respective values

Discussion

This is the first study to validate the use of the HADS questionnaire in an HOCM population, and the results affirm those from other studies across differing populations. As can be seen from the AUC, both subscales, but particularly the depression scale, perform well. It can therefore be concluded that the HADS performs well as a screening instrument. The ROC curves confirm the original design as they demonstrate that a cutoff point of 8 on both subscales offers the best compromise between

References (18)

There are more references available in the full text version of this article.

Cited by (58)

  • Emotional and behavioral problems in children with autism spectrum disorder and psychological distress in their parents in Saudi Arabia: The moderating effect of marital relationship satisfaction

    2022, Journal of Pediatric Nursing
    Citation Excerpt :

    About 62.4% (58) were unemployed. For the HADS results, a cutoff score of 8 on the respective depression and anxiety subscales was used to determine whether a person was suffering from depression or anxiety (Poole et al., 2006). Around 56% of the parents had anxiety symptoms (53) and around 61% (57) had depressive symptoms; the mean score for anxiety was 8.72 (SD = 3.46, range = 0–21), and was 8.41 (SD = 3.20, range = 0–21) for depression.

  • Anxiety and depression following cumulative low-level exposure to organophosphate pesticides

    2016, Environmental Research
    Citation Excerpt :

    Why this might be the case is unclear, but it is possible that self-report measures are not valid measures of psychiatric disorder for OP-exposed cohorts, possibly because many of the symptoms included in these questionnaires are non-specific and occur in a wide range of medical conditions (including several cognitive and somatic symptoms often reported following OP exposure). This notion is supported by a recent study which has found that the HADS and Beck Inventories may not be valid measures of psychiatric illness in OP-exposed participants (Harrison and Mackenzie Ross, In preparation); nor in many other patient groups with concurrent ill-health symptoms (e.g. Golden et al., 2007; Mitchell et al., 2010; Poole et al., 2006; Watson et al., 2014). Thus it is unclear how meaningful the information derived from self-report measures is in clinical, or real-world terms.

View all citing articles on Scopus
View full text