Background: Psychological distress is frequently observed,1 however, it is underestimated in cancer patients. The aim of this study is to develop a simple battery for screening for psychological distress, adjustment disorder and major depressive disorder in Japanese cancer patients.
Methods: One hundred and twenty-eight cancer patients were interviewed by psychiatrists and tested using the Hospital Anxiety and Depression Scale (HADS), a 14-item self-assessment questionnaire. Psychiatric diagnoses were performed according to the Diagnostic and Statistical Manual of Mental Disorders, third edition-revised.
Results: Cronbach α values of the Japanese version of the scale were 0.77 for the subscale for anxiety and 0.79 for depression. By a receiver operating characteristic analysis, we determined that an optimal cut-off point for screening for adjustment disorder and major depressive disorder was 10/11, which gave high enough sensitivity and specificity (91.5 and 65.4%, respectively). To screen for major depressive disorder alone, 19/20 was an optimal cut-off point with 82.4% sensitivity and 96.3% specificity. The subscales of HADS (anxiety and depression) also had high screening performance.
Conclusions: The Japanese version of HADS is a sensitive and specific tool for screening for psychological distress in Japanese cancer patients. This scale can be used for an early detection of patients' psychological distress which may be followed by psychiatric interventions.
Cancer patients must face several stresses, and factors which cause their psychological distress, especially anxiety and depression. Previous reports indicated that the most frequent psychiatric diagnoses of cancer patients were adjustment disorder with anxiety and/or depression and they also suffered from major depressive disorder (1).
Although patients with these psychiatric diagnoses need to take specific psychiatric therapies, including medication, they tend to be over- or underdiagnosed by medical staff (2,3). One reason for this tendency is that medical staff, even the family, take it for granted that cancer patients have some psychological distress so that it is not worthwhile taking psychiatric interventions (4), and the other is a difficulty in assessing their distress because of the physical symptoms (5,6). However, this psychological distress can disturb the patient's quality of life and affect their choice of treatments for cancer.
Some screening batteries have been developed for the early detection of patients' psychological distress (7,8). The Hospital Anxiety and Depression Scale (HADS; refer to the Appendix), a self-report questionnaire, was developed by Zigmond and Snaith (9) and is well accepted to screen for psychiatric problems in medically ill patients because it has only 14 items and does not include questions about physical symptoms (10,11).
In the present study, we investigated the reliability and the validity of the Japanese version of HADS, and its usefulness to screen for psychological distress in Japanese cancer patients.
Materials and Methods
Between August 1996 and April 1997, 128 cancer patients were interviewed by three trained psychiatrists (AK, TA, YU) according to the Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders, third edition-revised (SCID) and simultaneously filled out HADS after the interview at the National Cancer Center Hospital East. These patients had been selected in two ways. There were 31 who had been referred by oncologists to the psychiatrists and diagnosed as adjustment disorder or major depressive disorder (including eight ambulatory patients). The diagnosis of adjustment disorder is defined as emotional or behavioral disturbances in response to an identifiable stressor (e.g. diagnosis of cancer) which cause a significant impairment in social or occupational functioning. On the other hand, major depressive disorder is thought to be a more severe disorder and is defined as at least two weeks of depressed mood or loss of interest accompanied by at least four of the following symptoms: 1. decrease in appetite, or weight loss; 2. sleep disturbance; 3. psychomotor retardation or agitation; 4. fatigue or loss of energy; 5. feeling of worthlessness; 6. diminished ability to think or concentrate; 7. suicidal ideation. The other subjects were 55 advanced lung cancer patients and 42 head and neck cancer patients who were newly admitted to the hospital and waiting for treatment (mainly surgery or chemotherapy). They were sampled consecutively, and the patients with cognitive impairment or severe physical dysfunction were excluded. Those patients gave written informed consent.
Ninety-seven of the newly admitted patients also filled out the Profile of Mood States (POMS) as an external standard. Test-retest reliability of HADS was examined in 109 ambulatory breast cancer patients who also gave written informed consent.
This study was approved by the Institutional Review Board of the National Cancer Center.
The HADS is a four-point, 14-item self-assessment scale to measure psychological distress, and has two factors: anxiety and depression. The Japanese version of HADS was backtranslated by Kitamura (12).
POMS is known as a useful measurement of mood in cancer patients and consists of six factors: Tension-Anxiety, Depression-Dejection, Anger-Hostility, Vigor, Fatigue, and Confusion (13). The validation study of the Japanese version of POMS was reported by Yokoyama et al. (14).
In order to test the specificity and sensitivity of HADS, receiver operating characteristic (ROC) analysis was performed twice: once for assessing its validity as a method of screening for adjustment disorder and major depressive disorder taken together, and once for screening for major depressive disorder alone, excluding adjustment disorder.
Representing ROC analysis on a curve is a way of expressing the relationship between the true positive rate (sensitivity) and the false-positive rate (1 - specificity). The curve is a representation of the ability of the screening instrument to discriminate between ‘cases’ and ‘non-cases’. The desired cut-off point is generally chosen in order to minimize the sum of false-positive and false-negative test results.
The other analyses described in the article were conducted using SAS statistical software.
Table 1 shows the patients' sociodemographic and medical data. The mean age of the subjects was 61.1 ± 10.8 (SD) years (52–91 years, median 61), and 62.5% of the subjects were male. The cancer sites were mainly lung and head and neck. Most of the subjects were married and 41.4% had a full-time job. As to phase of illness in hospitalized patients (n = 120), 85.0% were waiting for the initial treatment, and the others were post operation (10.0%) or during chemotherapy or radiation (3.3%). Two were in terminal phase. The median performance status (the Eastern Cooperative Oncology Group's Scale) was one. The referred patients were significantly younger (mean age 54.7 vs 63., t = 3.98; d.f. = 126; p < 0.001), consisted of higher percentages of females (61.3% vs 29.9%, χ2 = 10.4; d.f. = 1; p < 0.01), and had poorer performance status (mean 1.32 vs 0.58, z = 2.72; p < 0.01).
Reliability and Validity of the Japanese Version of HADS
Cronbach's α of the Japanese version of HADS was 0.77 for the anxiety subscale and 0.79 for the depression subscale. The correlations of the scores between test and retest were 0.82 for the total score of HADS, 0.73 for the anxiety subscale, and 0.82 for the depression subscale (p = 0.0001; Pearson correlation test). The mean interval between test and retest was 7.9 days (4–23 days). The correlation between the total scores of HADS and that of POMS, between the anxiety scores of HADS and the Tension-Anxiety scores of POMS, and between the depression scores of HADS and the Depression-Dejection scores of POMS were 0.60, 0.60 and 0.49, respectively (p = 0.0001; Pearson correlation test). The results of confirmatory factor analyses were as follows: Goodness of Fit Index (GFI), 0.90; GFI adjusted for degrees of freedom, 0.86; root mean square residual, 0.057: χ2 = 97.1; and probability based on a χ2 > 0.05.
Screening for Psychological Distress by HADS
Thirty (23.4%) of 128 patients were diagnosed as adjustment disorder and 17 (13.3%) were major depressive disorder. The mean HADS score for the patients without adjustment disorder or major depressive disorder was 8.7, for the patients with adjustment disorder 18.0, and for the patients with major depressive disorder 23.5.
Figure 1 shows the ROC curve for screening for adjustment disorder and major depressive disorder expressing the relationship between the true positive rate (sensitivity) and the false-positive rate (1 - specificity) for each of the total scores of HADS. The ROC curve for screening for major depressive disorder alone is shown in Fig. 2. The optimal cut-off point for the screening of adjustment disorder and major depressive disorder seems to be 10/11. This cut-off point is associated with 91.5% sensitivity and 65.4% specificity. The optimal cut-off point for the screening of major depressive disorder seems to be 19/20, which gave 82.4% sensitivity and 96.3% specificity (Table 2).
For the screening of adjustment disorder and major depressive disorder, the probability of a high HADS score of 15 being associated with a case [the positive predictive value (PPV)] is 72.3%, and the PPV of a low score of four is 41.6%. For the screening of major depressive disorder taken alone, the PPV of a high score of 26 is 92.3% and that of a low score of 12 is 61.2%.
The HADS anxiety and depression subscales were also tested and optimal cut-off points determined for screening for adjustment disorder and major depressive disorder or major depressive disorder alone (Table 2). These subscales, except the depression subscale for screening for adjustment disorder and major depressive disorder (specificity 58.0%), provided high sensitivity and specificity.
The present results indicated that the Japanese version of HADS had well-accepted reliability and validity, and the scale was a useful screening battery for screening for psychological distress in cancer patients. Sensitivity and specificity for the optimal cut-off points were high enough compared with the previous reports (7,8,10,15,16–18).
HADS was developed by Zigmond and Snaith in 1983 (9) and widely used as an instrument for detecting states of anxiety and depression in a healthy population (19), psychiatric patients (20), or medically ill patients (21,22) such as those with stroke (23) and rheumatoid arthritis (24). Most studies indicated its validity and usefulness as a screening battery, and our results confirmed that the Japanese version of the scale is also well accepted.
Cancer patients whose psychological distress is mainly characterized by anxiety and depression can benefit from using HADS to detect this distress. Carrol et al. (17) tested 809 cancer patients using HADS and reported that 23.1 % of them had scores of 11 or greater for the anxiety or depression scale, and appeared to have anxiety or depressive disorders. The distress of Japanese cancer patients measured by HADS as well as POMS was obviously higher in the subjects who were diagnosed as adjustment disorder or major depressive disorder; however, we could not say that the prevalence of distress or scores of these scales were high or not by this study. In Carrol's report, clinical psychiatric interviews were not performed. Ibbotson et al. (7) used HADS as a screening battery and interviewed simultaneously. They determined an optimal cut-off point of HADS as 14/15 to screen for psychiatric problems in 284 cancer patients. Razavi et al. (10) tested HADS in 210 cancer in-patients with variable sites and performed psychiatric diagnosis simultaneously. The design and sampling of the study were similar to those of our study. Their result indicated that optimal cut-off points were 12/13 or 18/19 for screening for adjustment disorder and major depression or major depression alone, respectively. The cut-off point which we suggested (10/11) was lower than in these reports. The optimal cut-off point gave high sensitivity (91.5%), and the false-positive rate was 34.6%. Razavi et al. (10) indicated the possibility of a false-positive rate caused by the frequency of acute stress reactions, which may give rise to a high level of psychological distress and thus high scores on HADS. The false-positive rate of our study may be associated with the patient's subclinical high distress in Japanese cancer patients.
We can choose 14/15 as a cut-off point if we need higher specificity, resulting in 76.5% (76.6% sensitivity) in our samples, which is superior to that of Razavi's report. Which cut-off points we should select depends on the purpose of using HADS. The anxiety subscale of HADS also gave high sensitivity and specificity at the cut-off point 7/8. This result supports the previous report indicating the usefulness of the separate subscales of HADS as a screening battery (11). However, the anxiety subscale gave the same cut-off point for major depressive disorder as for adjustment disorder and major depressive disorder, indicating the difficulties in distinguishing this type of psychological distress by the subscale.
The limitation of our study is that the subjects consisted of two different groups. Ibbotson et al. (7) suggested that HADS had a good screening performance, especially for those free of disease and when the disease was judged to be more stable than in those with progressive disease, and concluded that we should take disease and treatment status into account. Furthermore, the previous studies suggested that we need to investigate the screening performance of the scale in the setting of site-specific malignancies (8,15). Future research in a specific setting will provide us with more specific cut-off points.
In conclusion, this is the first report that proved the reliability and validity of the Japanese version of HADS, and the scale appeared to be a simple, sensitive and specific battery for screening for psychological distress in Japanese cancer patients. For the early detection in general cancer patients of adjustment disorder and major depressive disorder for which we should consider psychiatric interventions or careful observations of the patient, an optimal cut-off point of HADS of 10/11 is recommended. Further research that we investigates the screening property of the scale in a more specific setting will be required.
Hospital Anxiety and Depression Scale
the Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders, third edition-revised
the Profile of Mood States
receiver operating characteristic
Goodness of Fit Index
positive predictive value
This work was supported in part by a Grant-in-Aid from the Ministry of Health (9–31) and Welfare for the Second term Comprehensive 10-Year Strategy for Cancer Control, from the Ministry of Health and Welfare, and by a grant from the Pfizer Health Research Foundations, Tokyo, Japan.
The authors thank Ms. Keiko Takeda, Yoko Kunimoto, Naomi Suimon, Yuko Kojima, Kumiko Harada, Yukie Sugihara and Ms. Miho Sakai, at the Psycho-Oncology Division, National Cancer Center Research Institute East, Japan, for their research assistance, and all staff at the Head and Neck Ward, National Cancer Center Hospital East, Japan.
A. The Hospital Anxiety and Depression Scale (HADS)
Read each item and please circle the answer which comes closest to how you have been feeling, on the average, in the past week. Don't take too long over your answers; your immediate reaction to each item will probably be more accurate than a long thought out response.
I feel tense or ‘wound up’:
Most of the time
A lot of the time
From time to time, occasionally
Not at all
I still enjoy the things I used to enjoy:
Definitely as much
Not quite so much
Only a little
Hardly at all
I get a sort of frightened feeling as if something awful is about to happen:
Very definitely and quite badly
Yes, but not too badly
A little, but it doesn't worry me
Not at all
I can laugh and see the funny side of things:
As much as I always could
Not quite so much now
Definitely not so much now
Not at all
Worrying thoughts go through my mind:
A great deal of the time
A lot of the time
From time to time but not too often
I feel cheerful:
Not at all
Most of the time
I can sit at ease and feel relaxed:
Not at all
I feel as if I am slowed down:
Nearly all the time
Not at all
I get a sort of frightened feeling like ‘butterflies’ in the stomach:
Not at all
I have lost interest in my appearance:
I don't take as much care as I should
I may not take quite as much care
I take just as much care as ever
I feel restless as if I have to be on the move:
Very much indeed
Quite a lot
Not very much
Not at all
I look forward with enjoyment to things:
As much as ever I did
Rather less than I used to
Definitely less than I used to
Hardly at all
I get sudden feelings of panic:
Very often indeed
Not very often
Not at all
I can enjoy a good book or radio or TV program:
B. Calculation of HADS
Each item is scored 0–3 (items 1,3,5,6,8,10,11,13 are reversely scored). The maximum score of this scale is 42. The odd-numbered items indicate anxiety subscales, and the even-numbered items depression subscales.
C. The Japanese Version of HADS
The copyright of the Japanese version of HADS belongs to Dr Kitamura, and you can refer to the full scale of this version in reference 12 (the original reference is 9).