Normative values for the Hospital Anxiety and Depression Scale (HADS) in the general German population

Abstract

Objective

The aim of this study was to present normative values for the Hospital Anxiety and Depression Scale (HADS).

Methods

A representative sample of the German general population (N=4410) was tested with the HADS.

Results

Females are more anxious than males, and older subjects are more depressed than younger subjects. The mean scores for anxiety / depression are 4.4 / 4.8 (males) and 5.0 / 4.7 (females). Using the cut-off 8+, the percentages of elevated anxiety and depression in the total sample are 21 % and 23 %, respectively. Regression analyses proved a linear but not a curvilinear age trend of anxiety and depression. Percentile rank norms are given for anxiety, depression, and the HADS total score.

Conclusion

The regression coefficients allow the calculation of expected mean scores for each age and gender distribution of any sample of patients. HADS mean scores are better suited to describe the degree of anxiety and depression in patient samples compared to percentages of subjects with elevated values.

Introduction

Anxiety and depression are frequently observed symptoms in cases of severe illness, especially in the fields of cardiovascular diseases [1], [2] and oncology [3], [4], [5]. However, even pathologically elevated degrees of anxiety and depression often remain undetected in clinical practice. Screening instruments have been developed that can be used to detect patients with high levels of psychological distress.

One such screening instrument is the Hospital Anxiety and Depression Scale (HADS). It was developed by Zigmond and Snaith [6] to identify cases of anxiety and depression disorders among patients in nonpsychiatric hospital clinics. The questionnaire consists of 14 items, seven anxiety and seven depression items. Symptoms relating also to physical disorder, such as dizziness, headache or insomnia were not included. The answer format of the items has four degrees scored with values 0 to 3. Anxiety and depression scores are obtained by simply summing up the scores of the seven items, yielding values between 0 and 21. The HADS authors define three ranges for each subscale: 0-7 (non-cases), 8-10 (doubtful cases) and 11-21 (cases). These cut-offs (8+ and 11+) were defined on the basis of psychiatric ratings of anxiety and depression disorders.

The HADS has been extensively used in clinical practice [7], and there are many papers testing the psychometric quality of the HADS. Bjelland and colleagues [8] give a review of 747 papers using the HADS, testing the factorial structure, the internal consistency, sensitivity, specificity (together with the calculation of optimal cut-off scores) and correlations to other questionnaires and interview-based measures. Though in some studies the original cut-off scores (8+ and 11+) failed to be optimal, until now there is no convention concerning the use of other cut-offs. Most of the factor analyses confirmed the two-factorial structure, but other investigations (especially those using structural equation modelling) detected three or even four factors. Rasch analyses have also been used to test the psychometric quality of the HADS scales [9], [10]. The internal consistency was appropriate in most cases with Cronbach alpha values above 0.80 [8]. A recent study comparing various patient-reported outcome measures of anxiety, depression and general distress used in RCTs of psychosocial interventions for people with cancer [11] found out that the HADS scored highest overall.

In order to assess whether a sample of patients is especially affected by psychological distress, comparisons with the general population are necessary. Even if several groups of patients are to be compared, one is often faced with the problem that gender and age distributions of the samples are not identical. Here it is necessary to know to which degree gender and age differences affect the results. In most HADS studies, anxiety and depression are considered separately. The original test description does not recommend the calculation of a combined total score. In clinical practice, however, clinicians are often interested in a summarising assessment of the patient concerning his or her general psychological distress. One possibility is to decide whether at least one of the scales exceeds a cut-off score. Since there are two cut-off scores (8+ and 11+) for each scale, and since there are several possibilities to combine the scales (heightened score in at least one scale or in both scales), the procedure is ambiguous. Several researchers therefore decide to calculate a total score, simply summing up the anxiety and the depression score. Several attempts have been undertaken to define an optimal cut-off score for this total score, based on sensitivity and specificity calculations. In the review of Bjelland et al. [8], 16 papers are listed with 11 different cut-off scores for the HADS total score, ranging from 8+ to 21+. A recent paper of Singer et al. [12] recommends the use of the cut-off score of 13+ (“medium model”), which is nearly in the middle of the scores presented in the literature.

Though there are very many studies using the HADS in clinical samples, our knowledge about normative data of the general population (including age and sex differences) is limited. The most comprehensive HADS depression study was performed in Norway with more than 60.000 subjects of the general population [13]. This study proved a clear and nearly linear age trend of the HADS depression scale. A Swedish [14], a Dutch [15] and a Korean [16] study also detected increasing depression scores with age. However, two British studies [17], [18] (which partly used the same data set) failed to detect such an age trend. Most investigations found only small gender differences in depression. Five studies [13], [17], [18], [19], [20] obtained gender differences less than 0.3. Two other investigations [14] detected higher depression mean scores in males compared to females (mean score difference about 0.5), while the Korean study [16] found an opposite trend. HADS anxiety shows other age und gender differences than depression. While the age effect is negligible or unclear [14], [15], [18], most studies found a gender effect with higher anxiety mean scores for females. The corresponding differences between females and males were 0.5 [14], [21], 0.7 [19], [16], 1.3 [18], and 2.1 [20].

Beyond age and gender differences, it is important to compare the raw mean scores obtained in these studies. Depression mean scores are relatively homogeneous, at least in Europe. Most studies [13], [15], [18], [19], [21], [22] presented mean scores between 3.3 and 3.7, the Swedish study [14] obtained 4.0. Anxiety mean scores are more heterogeneous. The following mean scores are reported: 3.9 [15], 4.1 [19], [22], 4.5 [14], 4.7 [21], 5.3 [20] (young persons only), 6.0 [23] (women only) and 6.4 [18]. Percentile norm values are only given in the British studies [17], [18] with the very high anxiety mean scores.

The differences among the studies require further investigations of normative values based on large and representative samples of the general population. Since age influences depression, and gender influences anxiety, these normative values should be given for age groups and both genders separately. In the large Norwegian study [13], an interesting approach is presented. A linear regression analysis was performed, and the coefficients of this regression equation can be used to calculate expected mean scores for each age and for both genders. This approach was also used in the British study [18], including a squared and a cubic component, which, however, contributed only marginally to the explained variance.

The objectives of this study are

• to describe age and gender differences for HADS anxiety and depression in terms of mean scores and percentages above the cut-offs,

• to test linear and nonlinear age effects,

• to calculate regression equations and to present normative values (percentile ranks) for anxiety, depression, and the HADS total score.