(2007) 19, 321 325 & 2007 Nature Publishing Group All rights reserved 0955-9930/07 $30.00 www.nature.com/ijir ORIGINAL ARTICLE Validation of the Hospital Anxiety and Depression Scale and the psychological disorder among premature ejaculation subjects QK Fatt 1, AS Atiya 2, NGC Heng 3 and CC Beng 4 1 School of Medicine & Health Sciences, Monash University Malaysia, Bandar Sunway, Petaling Jaya, Selangor Darul Ehsan, Malaysia; 2 Faculty of Medicine, Department of Social & Preventive Medicine, University Malaya, Kuala Lumpur, Malaysia; 3 Faculty of Medicine, Department of Medicine, University Malaya, Kuala Lumpur, Malaysia and 4 Sunway Medical Center, Bandar Sunway, Petaling Jaya, Selangor Darul Ehsan, Malaysia Premature ejaculation (PE) is a common sexual dysfunction among the general population. PE has often been associated with a psychological state of mind. Hospital Anxiety and Depression Scale (HADS) can be used as an instrument to assess the emotional and psychological state. The present study was designed to assess the reliability and validity of the HADS in a Malaysian population. The validity and reliability were studied in subjects with and without PE. Test retest methodology was used to assess the reliability whereas Cronbach s alpha was used to assess the internal consistency. In the control and the PE groups, the internal consistency was good and a high degree of internal consistency was observed for all 14 items. In the control group, the Cronbach s alpha values at baseline were from 0.811 to 0.834, whereas for retest, the Cronbach s alpha values were from 0.821 0.838 items. Intraclass correlation coefficient (ICC) was high for the control (0.797 0.868: baseline and 0.805 0.872: retest) and PE group (0.822 0.906: baseline and 0.785 0.887: retest). The high value of ICC and the internal consistency was due to high reliability and consistency of the items at 2-week interval. A degree of significance between the baseline and week-2 scores was observed across all items in the PE group but not in the control group. The HADS is a suitable, reliable, valid and sensitive instrument to measure the clinical change for anxiety and depression in the Malaysian population. (2007) 19, 321 325. doi:10.1038/sj.ijir.3901528; published online 30 November 2006 Keywords: premature ejaculation; Hospital Anxiety and Depression Scale; internal consistency; intraclass correlation coefficient; test retest reliability and validity Introduction Premature ejaculation (PE) is a common sexual dysfunction that may adversely affect men, their partners and their sexual and nonsexual relationships. 1,2 PE is often associated with a psychological state of the subjects such as stress, anxiety, pessimism and low self-esteem. 3 5 Self-assessment scales have shown to be useful instruments and an aid to practice. Numerous questionnaires are available to assess male vulnerability to PE in relation to their Correspondence: Dr QK Fatt, School of Medicine and Health Sciences, Monash University Malaysia, 20 and 22, Jalan PJS 11/5, Bandar Sunway, Petaling Jaya, Selangor Darul Ehsan 46150, Malaysia. E-mail: Quek.Kia.Fatt@med.monash.edu.my Received 5 August 2006; revised 21 September 2006; accepted 28 September 2006; published online 30 November 2006 emotional and psychological states and if the questionnaires are routinely used to screen individual, who are at risk through the recommendation of a clinician and psychologist, it can improve the patient s sexual life, self-esteem and marital relationships. The most widely used questionnaire is the Hospital Anxiety and Depression Scale (HADS). 6,7 It consists of two subclasses, one measuring anxiety whereas the other measuring depression, and both are scored separately. The seven-item questions in each subclass was developed to assess the severity of anxiety and depression, which are categorized into normal, mild, moderate and severe in their respective subclasses. 8,9 The anxiety score addressed the subject s anxious mood, restlessness and anxious thoughts, whereas the depression score addressed the state of loss of interest and diminished pleasure response. 10,11 The HAD is the present-state instrument and may be re-administered at weekly intervals to assess the subject s
322 progression. 12 Each question is scored on a numeric scale (maximum score of 3), with higher scores indicating a worsened function. The questionnaire is simple, short and requires only 5 10 min for completion and is easy to administer and is suitable for routine clinical practices. This is particularly important in PE subjects, as subjects suffering from such disorders may have an emotional state that varies from one day to another. The retest allows serial follow-up and will provide a useful record of progress. The HADS is validated and has the ability of both diagnosis and grading psychological states in many population, 13 16 but has not been used among the Malaysian population. The present study was conducted at the University Malaya Medical Center, Kuala Lumpur, and was designed to assess the reliability and validity of the HADS in a Malaysian population. Materials and methods The 14-item HADS questionnaire consisted of two sections each with seven items for the Anxiety scale and seven items for the Depression scale. Each of these items is rated from 0 (not at all) to 3 (almost always). The total score is the sum of items 1 7 (range: 0 21) according to the severity of the symptoms. The subjects were selected based on the inclusion and exclusion criteria set in the study and their willingness to participate in the study. The samples were recruited as they presented in the Urological and primary care clinic for other medical complaints. During the 2-week interval, there was no treatment for PE involved. Therefore, there should be no much changes in terms of symptomatology. These patients also do not have any major disease that can affect their ejaculation. For subjects with PE, the inclusion criteria were intravaginal ejaculation latency time (IVELT) at less than 2 min; 17 and their literacy and ability to provide an informed consent. For the control group, the inclusion criteria were IVELT of more than 2 min; and their literacy and ability to provide an informed consent. Subjects who consented to take part in the study were given and asked to complete the HADS questionnaire (self-administered). The HADS questionnaire was administered to the PE group (n ¼ 82; mean age: 44.4710.2 years) and to the control group (n ¼ 185; mean age: 44.179.8 years) and the validity and reliability of the questionnaire were studied. All PE and control subjects who were included in the reliability study were retested approximately 2 weeks after the initial administration of the HADS questionnaire. The English version of HADS was utilized in this study and since the cultural background of the population in this country differs from other countries; therefore, there is a need to validate the HADS before actually using it in future studies in Malaysia. Data analysis Validity and reliability were determined based on a standard protocol. Cronbach s alpha (CA) coefficient was used to assess the internal consistency of the HADS questionnaire. Test retest reliability was assessed using the intraclass correlation coefficient (ICC) with two-way random effects model. The values of ICC vary from 1 (perfectly reliable) to 0 (totally unreliable). Sensitivity to change (responsiveness) was assessed in the patients with PE and this was analyzed by assessing the difference between mean HADS before (baseline) and after and dividing it by the standard deviation baseline. Guyatt statistics is assessed by calculating the mean difference of HADS score between baseline and week 2 and dividing it by the standard deviation of the stable group. Here, the stable group is also considered as a control group. Results A total of 267 subjects had participated in the validity and relativity study. The socio-demographic characteristics of the subjects are shown in Table 1. For the control and PE groups, the internal consistency was good and a high degree of internal consistency was observed for all 14 items. The high value of CA in the control and PE group (CA40.800) at both baseline and 2-week interval indicated a high level of homogeneity among items (Table 2). ICC was high for the control group (ICC40.800) and Table 1 Socio-demographic characteristics of the subjects Control (n ¼ 185) (%) PE (n ¼ 82) (%) Age group 20 29 years 16 (8.649) 14 (17.073) 30 39 years 68 (36.757) 15 (18.293) 40 49 years 41 (22.162) 25 (30.488) 50 59 years 39 (21.081) 17 (20.732) 460 years 21 (11.351) 11 (13.415) Ethnicity Malay 78 (42.162) 37 (45.122) Chinese 84 (45.405) 18 (21.951) Indian 21 (11.351) 27 (32.927) Others 2 (1.081) 0 Ejaculation time (min) 16.038 2.049 9.791 0.607 Abbreviations: PE, premature ejaculation;, standard deviation.
the PE group (ICC40.800) (Table 3). The high value of ICC and internal consistency was owing to the high reliability and consistency of the items. A degree of significance between baseline and week-2 scores was observed across all items in the PE group but not in the control group. In the PE group, a significant difference was observed at test retest between baseline and after 2-week interval for mean HADS anxiety score (7.585 vs 6.841, Po0.001) and HADS depression score (7.476 vs 6.671, Po0.001). Several items also noted a significant difference after 2-week interval and it included items 3 7, 9 12 and 14 (Table 3). On the other hand, none of the anxiety and depression items in the control group were statistically significant. Discriminant validity or ability of the HADS to discriminate reliably between the PE group and control group was assessed. As shown in Table 4, a high significant difference was observed between the PE group and the control group for all the anxiety and depression items scores of the HADS. The mean difference between item scores for these groups was greatest in the total scores. 323 Table 2 Reliability for control group (n ¼ 185) Item test score (week 0) retest score difference Lower Upper 1 0.827 0.746 0.437 0.836 0.741 0.439 0.005 0.195 0.023 0.034 2 0.825 0.432 0.568 0.829 0.454 0.616 0.022 0.328 0.069 0.026 3 0.823 0.622 0.632 0.826 0.595 0.654 0.027 0.284 0.014 0.068 4 0.826 0.335 0.595 0.832 0.319 0.542 0.016 0.194 0.012 0.044 5 0.824 0.432 0.596 0.833 0.427 0.586 0.005 0.244 0.030 0.041 6 0.824 0.454 0.570 0.831 0.443 0.549 0.011 0.255 0.026 0.048 7 0.826 0.665 0.648 0.832 0.654 0.625 0.011 0.312 0.034 0.056 8 0.833 0.746 0.763 0.834 0.735 0.759 0.010 0.104 0.001 0.030 9 0.834 0.470 0.608 0.838 0.449 0.588 0.022 0.254 0.015 0.058 10 0.824 0.341 0.588 0.827 0.335 0.558 0.005 0.195 0.023 0.034 11 0.824 0.514 0.562 0.834 0.503 0.553 0.010 0.295 0.030 0.050 12 0.819 0.384 0.650 0.830 0.351 0.599 0.032 0.293 0.010 0.075 13 0.815 0.249 0.514 0.822 0.276 0.516 0.030 0.243 0.060 0.001 14 0.811 0.308 0.578 0.821 0.324 0.564 0.020 0.304 0.060 0.030 Abbreviations: CI, confidence interval; HADS, Hospital Anxiety Depression Scale; ICC, intraclass correlation coefficient; ¼ standard deviation. ICC ¼ 0.835 (: 0.797 0.868) (before retest). ICC ¼ 0.841 (: 0.805 0.872) (after retest). þ Po0.001 for all ICC and internal consistency; t-tests for paired comparison are not significant. Table 3 Reliability for PE group (n ¼ 82) Item test score retest score difference Lower Upper 1 0.854 1.200 0.456 0.823 1.158 0.508 0.037 0.292 0.028 0.101 2 0.854 1.220 0.522 0.832 1.171 0.517 0.049 0.268 0.010 0.108 3 0.859 1.110 0.544 0.824 0.976 0.608 0.134* 0.409 0.044 0.224 4 0.859 1.160 0.638 0.825 0.976 0.471 0.183 z 0.389 0.098 0.268 5 0.851 0.900 0.659 0.824 0.756 0.534 0.146* 0.475 0.042 0.251 6 0.854 1.040 0.532 0.839 0.927 0.438 0.110* 0.315 0.041 0.179 7 0.855 1.150 0.611 0.833 0.988 0.458 0.159* 0.429 0.064 0.253 8 0.858 1.120 0.596 0.825 1.050 0.586 0.007 0.409 0.020 0.160 9 0.861 1.170 0.663 0.830 1.050 0.607 0.122* 0.329 0.049 0.194 10 0.864 1.050 0.646 0.827 0.915 0.632 0.134* 0.327 0.051 0.217 11 0.858 1.020 0.684 0.832 0.910 0.613 0.110* 0.352 0.030 0.190 12 0.856 0.880 0.692 0.832 0.719 0.614 0.159* 0.484 0.052 0.265 13 0.876 1.040 0.656 0.841 1.000 0.648 0.040 0.367 0.040 0.120 14 0.866 1.010 0.657 0.841 0.910 0.670 0.100 # 0.372 0.020 0.180 Abbreviations: CI, confidence interval; HADS, Hospital Anxiety Depression Scale; ICC, intraclass correlation coefficient; PE, premature ejaculation;, standard deviation. ICC ¼ 0.868 (: 0.822 0.906) (before retest). ICC ¼ 0.841 (: 0.785 0.887) (after retest). þ Po0.001 for all ICC and internal consistency; t-tests for paired comparison: # Po0.05, *Po0.01, z Po0.001.
324 Table 4 HADS item characteristics: discriminant validity at baseline HADS items Control mean s.e.m. PE mean s.e.m. difference s.e.m. ESI Lower Upper P 1 0.746 0.032 1.195 0.050 0.449 0.060 1.014 0.567 0.331 o0.001 2 0.432 0.042 1.220 0.058 0.788 0.071 1.421 0.928 0.647 o0.001 3 0.622 0.046 1.110 0.060 0.488 0.076 0.805 0.638 0.338 o0.001 4 0.335 0.044 1.159 0.070 0.824 0.083 1.354 0.987 0.660 o0.001 5 0.432 0.044 0.902 0.073 0.470 0.085 0.763 0.638 0.302 o0.001 6 0.454 0.042 1.037 0.059 0.583 0.072 1.044 0.725 0.440 o0.001 7 0.665 0.048 1.146 0.067 0.481 0.083 0.755 0.645 0.318 o0.001 8 0.746 0.056 1.122 0.066 0.376 0.086 0.525 0.547 0.205 o0.001 9 0.470 0.045 1.171 0.073 0.701 0.086 1.121 0.870 0.531 o0.001 10 0.341 0.043 1.049 0.071 0.708 0.083 1.168 0.873 0.543 o0.001 11 0.514 0.041 1.024 0.076 0.510 0.086 0.847 0.681 0.340 o0.001 12 0.384 0.048 0.878 0.076 0.494 0.090 0.745 0.672 0.316 o0.001 13 0.249 0.038 1.037 0.072 0.788 0.082 1.404 0.095 0.641 o0.001 14 0.308 0.042 1.012 0.073 0.704 0.084 1.167 0.087 0.538 o0.001 Anxiety score 3.697 0.179 7.585 0.311 3.889 0.359 1.517 4.598 3.178 o0.001 Depression score 3.000 0.206 7.476 0.313 4.476 0.375 1.593 5.216 3.735 o0.001 Abbreviations: CI, confidence interval; ESI, effect size index; HADS, Hospital Anxiety Depression Scale; ICC, intraclass correlation coefficient; PE, premature ejaculation;, standard deviation; s.e.m., standard error of mean. Discussion The ability of the HADS to discriminate between subjects with and without PE indicates that it could be a valid, reliable and sensitive tool in assessing anxiety and depression owing to PE in a Malaysian population. There are many more instruments that can be used like PHQ and QIDS-SR but we decided to use HADS because it is quite widely used as it has long been used as a screening instrument in research as well as in clinical setting. For the test retest reliability, high ICC and internal consistency were noted for both the control and PE groups. This is not surprising as the test retest was carried out at 2-week interval, where there were not much changes that can be observed in the normal control subjects. However, a significant change was noted in subjects with PE after the 2-week interval, suggesting that emotional and psychological symptoms may relate directly to the occurrence of PE. This PE group was not given any treatment; therefore, the changes could be due to natural psychological disturbance owing to external factors such as social and economic factors. Apart from these, this PE group has a lifelong PE and there is no confounding diseases affecting their ejaculation. Similarly, all the items in the HADS had a high discriminant validity between the PE group and the control group (all items Po0.001). The positive value of effect size index (ESI) obtained when the questionnaire was administrated at the time of study (baseline) and after week 2 indicates a relatively high degree of sensitivity to change (responsiveness) in both anxiety and depression items. In the present study, we assessed the correlation between changes in the HADS scores for both anxiety and depression. The substantial improvement of PE subjects after a 2-week interval is indicated by a significant lower score of HADS in week 2 compared with baseline. The subjects with PE show a relatively unstable emotion and influenced by psychological factors such as tension, anxiety, life stressors associated with work and family and problems in their interpersonal relationships and communication. This can be seen by their responsiveness towards a series of items present in the HADS, which was statistically significant compared to that after the 2-week interval. The score after 2-week interval was significantly lower than that at week 0, which may be perceived as an improvement over their emotional and psychological state. These results suggest that the change in HADS scores may reflect meaningful clinical changes where the psychological factors have shown a close association with the incidence of PE. These findings provide a substantial assurance that scores obtained by using the HADS are relatively sensitive and reliable and can be used as a diagnostic tool by clinicians or psychologists to identify subjects who are prone to PE and to assess their anxiety and depression states and also for the follow-up improvements of PE subjects after psychological treatment. Some of the PE group subjects may have experienced the PE for quite some time and got used to it. Therefore, the natural psychological disturbance was not so great. The lack of intervention to measure the true responsiveness to change was found to be a limitation in this study.
In conclusion, the reliability and validity of the HADS for anxiety and depression were tested. The high ICC of HADS for both the control and the subjects with PE indicates excellent intraclass reliability whereas the high CA indicates that the HADS has a good internal consistency. These findings suggest that the HADS is a useful tool for assessing the occurrence of psychological problem in relation to disease. References 1 Lewin J, King M. Sexual medicine. Towards an integrated discipline. BMJ 1997; 314: 4132. 2 Rust J, Golombok S, Collier J. Marital problems and sexual dysfunction: how are they related? Br J Psychiatry 1988; 152: 629 631. 3 Jack TM, Walker VA, Morley SJ, Hanks GW, Finlay-Mills BM. Depression, anxiety and chronic pain. Anesthesia 1987; 42: 1235 1236. 4 Klein DF. Endogenomorphic depression. Arch Gen Psychiatry 1974; 31: 447 454. 5 Snaith RP. The concepts of mild depression. Br J Psychiatry 1987; 150: 387 393. 6 Norton GR, Jehu D. The role of anxiety in sexual dysfunctions: a review. Arch Sex Behav 1984; 13: 165 183. 7 Zigmond A, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983; 67: 361 370. 8 Typer S. Psychiatric assessment of chronic pain. Br J Psychiatry 1992; 160: 733 741. 9 Snaith RP, Taylor CM. Rating scales for depression and anxiety: a current perspective. Br J Clin Pharmacol 1985; 19: 175 205. 10 Carroll BT, Kathol RG, Noyes R, Wald TG, Clamon GH. Screening for depression and anxiety in cancer patients using the Hospital Anxiety and Depression Scale. Gen Hosp Psychiatry 1993; 15: 69 74. 11 Snaith RP, Taylor CM. Rating scales for depression and anxiety: a current perspective. Br J Clin Pharmacol 1985; 19: 175 205. 12 Bramley PN, Easton AME, Morley S, Snaith RP. The differentiation of anxiety and depression by rating scales. Acta Psychiatr Scand 1988; 77: 136 139. 13 El-Rufaie OEFA, Absood G. Validity study of the HADS among a group of Saudi patients. Br J Psychiatry 1987; 151: 687 688. 14 Wattis JP, Davies Kn, Burn WK, McKenzie FR, Brothwell JA. Correlation between hospital anxiety depressions (HAD) scale and other measures of anxiety and depression in geriatric inpatients. Int J Geriatr Psychiatry 1994; 9: 61 63. 15 Aylard PR, Gooding JH, McKenna PS, Snaith RP. A validation study of three anxiety and depression self assessment scales. J Psychosom Res 1987; 31: 261 268. 16 Barczak P, Kane N, Andrews S, Congdon AM, Clay JC, Betts T. Patterns of psychiatric morbidity in a genitor-urinary clinic: a validation of the hospital anxiety and depression scale. Br J Pyschiatry 1988; 152: 698 700. 17 Rowland DL, Cooper SE, Schneider M. Defining premature ejaculation for experimental and clinical investigations. Arch Sex Behav 2001; 30(3): 235 253. 325