C 2008 International Psychogeriatric Association International Psychogeriatrics (2008), 20:3, 571–581 doi:10.1017/S1041610208006741 Printed in the United Kingdom
Validation of the Arabic version of the short Geriatric Depression Scale (GDS-15) ..............................................................................................................................................................................................................................................................................
Monique Chaaya,1 Abla-Mehio Sibai,1 Zeina El Roueiheb,1 Hiam Chemaitelly,1 Lama M. Chahine,2 Hassen Al-Amin2 and Ziyad Mahfoud1 1 2
Faculty of Health Sciences, American University of Beirut, Lebanon Faculty of Medicine, American University of Beirut, Lebanon
ABSTRACT
Background: This study aimed to examine the validity and reliability of an Arabic version of the 15-item Geriatric Depression Scale (GDS-15). Methods: 121 community-dwelling older adults and primary care patients aged 60 and above participated in this study. Older adults with dementia, those with thyroid dysfunction, and hearing or speech impairments were excluded. Testretest reliability was examined by re-administering the translated GDS-15 to a subset of 38 participants at least seven days after the initial interview. Results: The Arabic GDS-15 had good psychometric properties, but the best properties were reported for the 7/8 cutoff. Cronbach’s α as a measure of internal consistency reliability was high (0.88) and κ ranged from 0.57 to 0.75. The performance of the GDS-15 was equally good for both community-dwelling older adults and those in primary care settings, and for both forms of the GDS (examiner administered vs. self-administered). Conclusion: The Arabic GDS is a useful measure to assess depression among community-dwelling older adults and primary care patients who do not have dementia. Because of the use of formal Arabic, GDS-15 can be widely used with all Arabic-speaking people. Key words: Arabic GDS-15, elderly depression, validation, geriatric scale
Introduction Worldwide, mental illness represents the second most reported disabling morbidity, and its burden is even greater among older adults. Depression is a particularly significant problem among older adult in terms of its prevalence Correspondence should be addressed to: Ziyad Mahfoud, Assistant Professor, Department of Epidemiology and Population Health, Faculty of Health Sciences, American University of Beirut, Lebanon. Phone: +961 1 374374; Fax: +961 1 744470. Email:
[email protected]. Received 9 Aug 2007; revision requested 31 Aug 2007; revised version received 8 Oct 2007; accepted 9 Oct 2007. First published online 21 February 2008.
571
572
M. Chaaya et al.
and serious consequences. In the U.S.A., up to 19% of individuals aged 65 years and above suffer from severe depressive symptoms (Federal Interagency Forum, 2000). Depression in this age group coexists with other medical illnesses and disabilities and can lead to suicide. It has been reported that 70% of those who commit suicide visited their primary care physician one month prior to committing suicide, which raises concerns about the under-diagnosis of depression (U.S. Department of Health and Human Services, 1999). The assessment of depression among older adults is often challenging, especially in non-clinical settings, because of the comorbidities that often occur in this age group, such as cognitive decline or dementia (Hocking et al., 1995). Therefore, the recognition of this problem is imperative for diagnosis and treatment to prevent disability and premature mortality. While several screening instruments have been used in population-based surveys to assess depression among older adults, the Geriatric Depression Scale (GDS) is the most commonly used. The GDS exists in many versions, including the 30-item and 15-item versions. The GDS has also been used in a few studies of older Arab samples. In two studies conducted in Saudi Arabia, the GDS 30-item and the GDS 15-item versions were used without being validated (Al-Shammari and Al-Subaie, 1999; Abolfotouh et al., 2001). In another study that investigated depression among an Arab-American sample living in Detroit, U.S.A., the GDS-30 item was used and high reliability was reported (Wrobel and Farrag, 2006). An Arabic translated, non validated GDS-15 was used in a recent study on 740 older persons aged 60 years and over, living in poor areas of Beirut, one of which is a Palestinian refugee camp (Chaaya et al., 2006). The prevalence of probable depression (a score of 5–10) and definite depression (a score of > 10) was 43.8% and 24% respectively. These estimates are higher than those reported in other Arab countries such as Saudi Arabia (17.5%; Abolfotouh et al., 2001) and similar to the Detroit study (29% severe depression; Wrobel and Farrag, 2006). The high prevalence of depression found in the Lebanese study may be a true estimate, but it might also have resulted from a measurement bias due to the use of a non-validated tool. The internal consistency reliability coefficient of the study was very low (0.3), which raises further questions about the validity of the GDS-15. Validation of the Arabic version of the GDS is therefore imperative. This study aimed to examine the psychometric properties of an Arabic version of the GDS-15 and specifically to assess its reliability and criterion validity for two different types of populations: community-dwelling older adults and primary care patients
Methods Participants Participants were both outpatients and community dwelling individuals aged 60 years and older. A total of 170 individuals were approached to participate in the study; 49 were excluded and the final sample consisted of 121 elderly persons of whom 69 (57%) were recruited from community settings and 52 (43%)
Validation of the Arabic version of GDS-15
from primary care centers. Of the 121 elderly included in the present study and who answered the GDS-15, 105 were evaluated immediately afterwards by a psychiatrist, and 39 completed a second GDS-15 one week after the initial one. Excluded individuals were those who suffered from physical or medical conditions that would influence their ability to answer the GDS items. The exclusion criteria were: (1) dementia as measured by the Mini-mental Status Examination (MMSE) – an MMSE score of less than 25 for literate people and less than 23 for illiterate people was indicative of dementia (illiterate people cannot complete two items on the MMSE: writing a sentence and following a read one-step command; Farrag et al., 1998); (2) hearing impairment; (3) inability to speak or unintelligible speech; (4) use of antidepressants/mood stabilizers during the four weeks preceding the interview; (5) acute medical condition requiring immediate care; (6) psychotic and manic symptoms; (7) any thyroid disease (as thyroid dysfunction may lead to positive answers on some GDS items such as “lack of energy” or “worse memory than others” in the absence of major depression); (8) propranolol intake (propranolol may lead to depression or to depressive symptoms such as decreased concentration, low mood, and decreased energy; Patten and Love, 1993); (9) alcohol dependence; and (10) recent stroke (in past four weeks). Primary care patients were identified at three separate primary care centers: the family medicine clinics at the American University of Beirut Medical Center (AUBMC), the outpatient department at AUBMC, and a clinic affiliated with a volunteer organization that provides medical care for elderly individuals. Community older resident participants were identified through governmental and non-governmental organizations providing different types of services to older adults. Instrument and translation procedure The GDS, originally developed as a 30-item questionnaire and validated by Yesavage et al. (1983) was intended to assess depression in elderly people without dementia. In 1986, Sheikh and Yesavage developed the GDS-15, a shorter version of the original GDS (Sheikh and Yesavage, 1986), which became one of the most commonly used tools in assessing elderly depression. It consists of 15 short questions (yes/no) inquiring about symptoms of depression where a score of less than five indicates the absence of depression, a score 5 to 10 indicates probable depression and a score of more than 10 indicates definite depression (Allen et al., 1994). The advantage that the GDS-15 has over the original GDS30 is that it is less time-consuming and more suitable for patients with dementia. The GDS-15 has been tested in multiple settings and countries. In the U.K., its internal consistency reliability was found to be high with a Cronbach’s α of 0.80 (D’ath et al., 1994). In Greece and Brazil, the GDS-15 respectively showed a sensitivity of 92.2% and 92.7% and a specificity of 95.24% and 65.2% (Fountoulakis et al., 1999; Almeida and Almeida, 1999). For the current study, the GDS-15 was translated from the original English version into Arabic by a professional translator and two bilingual psychiatrists. The versions of the Arabic GDS-15 thus obtained were compared and properly
573
574
M. Chaaya et al.
merged to obtain one Arabic version which was then back-translated into English and compared with the original in order to check for consistency. Discrepancies were corrected accordingly. The final Arabic version was then piloted within a small group of older adults (n = 10) to ensure that all the terms were understandable to them. Ambiguities were adjusted accordingly. Interviewing procedures Permission was obtained from all institutions and organizations where the study was conducted. The study protocol was also approved by the Institutional Review Board (IRB) at the American University of Beirut. Verbal but not written informed consent was obtained from all subjects as a relatively high level of illiteracy was expected. The consent form was read by a general practitioner and included detailed information about all phases of the study. The participants were fully aware that their participation would include an interview by a general practitioner as well as a consultation by a psychiatrist. A subset of the participants was also informed about the reliability study and the need for a second interview in a week’s time. For each subject, two interviews were conducted during the same session: one interview by a general practitioner and another by a psychiatrist. The general practitioner’s interview comprised a set of questions first about age, sex, martial status, level of education, nationality, employment, and medication history. The physician then administered an MMSE test and conducted a physical examination. Participants who were found to be unfit based on the physical examination or who scored high on the MMSE test were excluded from the study (see exclusion criteria). The GDS-15 was self-administered to participants who were literate and had adequate vision while the general practitioner read the questions to illiterate subjects and those who had difficulty in reading the text. Immediately afterwards, a psychiatrist who was blinded to the results of the GDS-15 examined the participant using the DSM-IV criteria for depression and other psychiatric conditions, and concluded at the end of the interview what diagnosis, if any, the subject should receive. The psychiatrist’s assessment was the reference standard against which the results of the GDS-15 were compared. One hundred and five elderly were interviewed in this manner. In order to assess the reliability of the GDS-15, the instrument was readministered to a subset of 38 participants at least seven days after the initial interview by the same general practitioner in the same manner as initially done. A history of negative life events since the initial interview was obtained at the time of the second interview so as to identify possible factors contributing to any change that might have been observed in the GDS score. Data analysis Prevalence of depression according to the psychiatrist’s diagnosis (based on DSM-IV criteria) and the GDS results using different cutoffs were calculated. To check for significant differences between the two groups of older adults (community dwellers vs. primary care patient) by gender, age and other variables, χ 2 tests were used.
Validation of the Arabic version of GDS-15
Cronbach’s α coefficient of the Arabic GDS-15 was calculated for the entire sample to check for reliability. This measure assesses the internal consistency reliability of the GDS scale. The test-retest reliability was assessed by calculating the κ statistic, which measures the degree of agreement beyond chance, Cronbach’s α and Spearman’s correlation. This was performed on 39 elderly participants. As for the validity analysis, the psychiatrist’s diagnosis was used as the golden standard criterion. Sensitivity, specificity, positive and negative predictive values (PPV and NPV) were computed for different cutoffs (4/5, 5/6, 6/7, 7/8, 8/9). The area under the receiver operating characteristics (ROC) curve was computed to determine the best cutoff for the GDS-15, one that maximizes specificity and sensitivity.
Results The average age of the 121 participants was 69.8 years with a standard deviation of 6.3 years. The vast majority were females (75%), while high proportions of the participants were illiterate (46%), widowed (47%) or married (43%). Very few were working (12 %). There were no differences in the demographic variables between those recruited from the community and those recruited from primary care centers (Table 1).
Table 1. Distribution of study respondents by demographic characteristics across selection setting T O TA L
(N =121)
PRIM ARY CARE SETTING
COMMUNITY SETTING
P
VA L U E
..............................................................................................................................................................................................................................................
Age: mean (SD) Sex Female Male Marital status Married Widowed Other Education Illiterate Read and write Elementary education Other Work status Currently working Ex-worker Never worked
% 69.8 (6.3)
n % 69.5 (6.4)
n % 70.0 (6.3)
75.2 24.8
37 15
71.2 28.8
54 15
78.3 21.7
0.370
44.2 45.0 10.8
24 22 6
46.2 42.3 11.5
29 32 7
42.6 47.1 10.3
0.873
42.1 20.7 17.4
19 10 10
36.5 19.2 19.2
32 15 11
46.4 21.7 15.9
0.535
19.8
13
25.0
11
15.9
12.7 51.7 35.6
7 22 23
13.5 42.3 44.2
8 39 19
12.1 59.1 28.8
0.654
0.167
575
576
M. Chaaya et al.
1.0
0.8
Sensitivity 0.6
0.4
0.2
0.0 0.0
0.2
0.4
0.6
0.8
1.0
1 - Specificity Figure 1. Area under the receiver operating characteristics (ROC) curve = 0.892.
Table 2 shows the psychiatrist’s clinical evaluation of the participants, the GDS results, sensitivity and specificity, PPV and NPV obtained with several choices of the cutoff values for the GDS-15 scores, and the κ statistics. The proportion of depressed patients, including both major depressive disorder (MDD) and dysthymia, based on the psychiatrist’s diagnosis was 44.8% with similar proportions for community-dwelling older adults and primary care patients. Moreover, none of the demographic variables showed any significant difference between those diagnosed with depression and those with no depression (data not shown). Using different cutoffs for GDS score, the proportion of depressed patients ranged from 35.2% with 8/9 cutoff to 59% for the 4/5 cutoff. The lowest cutoff gave a high sensitivity (0.92) and a low specificity (0.67) and the agreement between GDS score and the psychiatrists assessment based on DSM-IV criteria was 57%. Both the 6/7 and the 7/8 cutoff values gave a good balance between sensitivity (0.89 and 0.83 respectively) and specificity (0.83 and 0.91 respectively). However the 7/8 cutoff value produced the maximum agreement with the clinical diagnosis (κ = 0.75). The area under the ROC curve shown in Figure 1 is 0.89 (P value < 0.001; 95% CI 0.82, 0.96) indicating the ability of the Arabic GDS-15 to discriminate between cases and non-cases of depression in this population. Cronbach’s α coefficient for assessing the internal consistency reliability was found to be 0.83. The test-retest reliability for the Arabic version of the GDS-15
Table 2. Clinical diagnosis, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and kappa statistic(κ)
T O TA L
PRIM ARY CARE SETTING
COMMUNITY SETTING
P - VA L U E
SENSITIVITY
SPECIFICITY
PPV
NPV
κ
...............................................................................................................................................................................................................................................................................................................................................................................
(N =105) (N =36)
(N =69) 0.962
47 (44.8%) 58 (55.2%)
62 (59.0%) 43 (41.0%)
56 (53.3%) 49 (46.7%)
52 (49.5%) 53 (50.5%)
44 (41.9%) 61 (58.1%)
37 (35.2%) 68 (64.8%)
16 (44.4%) 20 (55.6%)
21 (58.3%) 15 (41.7%)
19 (52.8%) 17 (47.2%)
17 (47.2%) 19 (52.8%)
13 (36.1%) 23 (63.9%)
10 (27.8%) 26 (72.8%)
31 (44.9%) 38 (55.1%) 0.914
0.92
0.67
0.69
0.91
0.57
0.934
0.89
0.76
0.75
0.90
0.64
0.733
0.89
0.83
0.81
0.91
0.71
0.385
0.83
0.91
0.89
0.87
0.75
0.248
0.68
0.91
0.87
0.78
0.61
41 (59.4%) 28 (40.6%)
37 (53.6%) 32 (46.4%)
35 (50.7%) 34 (49.3%)
31 (44.9%) 38 (55.1%)
27 (39.1%) 42 (60.9%)
Validation of the Arabic version of GDS-15
Physician Evaluation Probable Cases Non-cases GDS 15 with 4/5 cutoff Probable cases Non-cases GDS 15 with 5/6 cutoff Probable cases Non-cases GDS 15 with 6/7 cutoff Probable cases Non-cases GDS 15 with 7/8 cutoff Probable cases Non-cases GDS 15 with 8/9 cutoff Probable cases Non-cases
577
578
M. Chaaya et al.
was high with Spearman’s correlation of 0.79. For the test-retest, κ statistics were also computed and all were significant (data not shown).
Discussion This study examined the psychometric properties of an Arabic version of the GDS-15 and the results suggest that this scale is a useful screening test to detect depression among Arabic speaking older adults. It demonstrated criterion validity, and at different cutoffs GDS was able to differentiate between depressed and non-depressed individuals. The cutoff that provided the best sensitivity, specificity, area under the ROC curve and κ statistics was 7/8. The secondbest cutoff was 6/7. These findings are in line with those reported from the validation of the Farsi (Iranian) version of the GDS-15 which yielded very similar results in criterion validity and in reporting the 7/8 as the optimal score (Malakouti et al., 2006). However, the PPV of the Iranian version was much lower than the Arabic one. This is expected with a higher prevalence of depression among the study population, which also reflects the high prevalence of mental disorders in Lebanon (Karam et al., 2006). The Iranian study was also different from the current one in that it used the Composite International Diagnostic Interview (CIDI) as the gold standard for the diagnosis of depression. The high cutoff of 7/8 has been reported in recent studies carried out among elderly inpatients in Malaysia (Teh and Hasanah, 2004) and Korean psychiatric patients (Bae and Cho, 2004). It is important to note that there is an established literature that shows that structured or semistructured diagnostic interviews such as the SCID (Structured Clinical Interview for DSM Disorders) are sometimes superior and more accurate than psychiatrists in identifying psychiatric illnesses such as depression (Basco et al., 2000). However, in a systematic review of the criterion validity of the Geriatric Depression Scale in 42 published articles, a third of the 16 criterion validity studies of the 15-item GDS conducted between 1989 and 2003 did not use a structured or semi-structured interview or instrument (Wancata et al., 2006). Our study results are within the range of the psychometrics of GDS reported in the Wancata et al. review (2006) and adhere both to their recommendations regarding particular methodological requirements for validity studies, and the Standards for Accurate Reporting of Diagnostic Accuracy (STARD) (Bossuyt et al., 2003). The findings of this study also suggest a moderate to high reliability for both the test-retest and the internal consistency reliability of the scale. In general, validation studies of the short version of the GDS-15 have all reported a high internal consistency reliability with Cronbach’s α ranging from 0.77 to as high as 0.94 (Fountoulakis et al., 1999; Lam et al., 2004; Friedman et al., 2005; Malakouti et al., 2006). The κ statistic in this study, especially for the 7/8 cutoff, was higher than any reported. Our study differed from others in that the timing of the second interview was approximately seven days whereas in other studies it ranged from 2–3 days to three weeks after the initial interview. The performance of the Arabic GDS-15 was equally good for both older adults in communities and those in primary care settings. Because dementia
Validation of the Arabic version of GDS-15
patients, patients with thyroid dysfunction, and so on, were excluded, the validity of the Arabic GDS for these patients cannot be inferred. In addition, while the validity of the Arabic GDS for primary care patients and community-dwelling elderly has been confirmed in this study, this version of the GDS may not be valid for in-patients or residents of nursing homes or other long-term care facilities. The GDS was readministered at least seven days after the initial interview. Prior to its readministration, a history of negative life events that had occurred since the initial interview was obtained. Because a second interview by a psychiatrist was not also conducted at this time, subjects who had experienced negative life events and whose GDS had worsened were not fully evaluated (i.e. even if the GDS was reliable in the sense that it worsened if the subject’s mood had worsened, this was not confirmed based on DSM-IV criteria). Whether the GDS was read by the interviewer or self-administered, validity and reliability did not change, especially at the 7/8 cutoff. In its original form, the GDS was validated for self-administration (Yesavage et al., 1983). Cannon et al. (2002) have shown that there is a significant correlation between the oral and written administrations of the English version of the GDS among higher cognitive functioning participants, which applies to the tested Arabic version. Because the level of illiteracy among the Arab elderly is relatively high, and older patients may lack the visual or concentration abilities to fill-out self-rated scales accurately, and it is not usual for questionnaires to be filled out as part of a health examination in many Arab countries, this Arabic GDS is expected to be orally administered more commonly in day-to-day practice. It is important to confirm in future studies conducted in other Arab countries whether the oral (observer-administered) version of the Arabic GDS is as valid and reliable as the self-administered one. Because the GDS was translated into formal Arabic, it is available to all Arabic-speaking populations in the region, including the Gulf and most North African countries. In addition, studies on the mental health of Arab elderly emigrants in all countries can make use of this screening instrument. However, there remain some culture-specific limitations where a few items on the GDS may not be suitable for all Arab populations, and local researchers therefore ought to look at individual GDS items. Most of the elderly recruited for this study were illiterate and of low or middle socioeconomic status (SES) and therefore our sample may not be representative of all community and primary care elderly patients. It was not possible to look into differences in psychometric properties according to literacy and SES levels, because of the sample size. Additional studies which select a larger number of older individuals with high levels of education and/or of high SES are needed. In conclusion, the GDS-15 was proven to be a good instrument for assessing depression in older adults, without dementia, in community settings and primary care patients. Because it is short and easy to administer, it is recommended that it be used as a routine screening test to identify depression among older adults in primary care settings, and to use it in large population health surveys.
579
580
M. Chaaya et al.
Conflict of interest None.
Description of authors’ roles M. Chaaya contributed to the conceptualization, implementation and writing of the paper. Z. Mahfoud performed the data analysis and helped write the paper. A Sibai and H. Al-Amin contributed to the conceptualization of the project. Z. El Roueiheb helped in data analysis and writing the paper. H. Chemaitelly contributed to the conceptualization and data collection, and L. Chahine contributed to data collection and helped write the paper.
Acknowledgments We thank Dr. Marwan Tabbarah and Dr. Osman Mohammad for their meticulous work in interviewing the participants; Dr. Zeina Chemali for her valuable input during conceptualization of the study; Dr. Mounir Khani for helping with the translation; and all the older adults who participated in the study. We also thank the Welcome Trust for funding the study.
References Abolfotouh, M. A., Daffallah, A. A., Khan, M. Y., Khattab, M. S. and Abdulmoneim, I. (2001). Psychosocial assessment of geriatric subjects in Abha City, Saudi Arabia. Eastern Mediterranean Health Journal, 7, 481–491. Allen, N., Ames, D., Ashby, D., Bennetts, K., Tuckwell, V. and West, C. (1994). A brief sensitive screening instrument for depression in late life. Age and Ageing, 23, 213–219. Almeida, O. P. and Almeida, S. A. (1999). Short versions of the Geriatric Depression Scale: a study of their validity for diagnosis of a major depressive episode according to ICD-10 and DSM-IV. International Journal of Geriatric Psychiatry, 14, 858–865. Al-Shammari, S. A. and Al-Subaie, A. (1999). Prevalence and correlates of depression among Saudi elderly. International Journal of Geriatric Psychiatry, 14, 739–747. Bae, J. N. and Cho, M. J. (2004). Development of the Korean version of the Geriatric Depression Scale and its short form among elderly psychiatric patients. Journal of Psychosomatic Research, 57, 297–305. Basco, M. et al. (2000). Methods to improve diagnostic accuracy in a community mental health setting. American Journal of Psychiatry; 157, 1599–1605. Bossuyt, P. M. et al. (2003). Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD Initiative. Clinical Chemistry, 49, 1–6. Cannon, B. J., Thaler, T. and Roos, S. (2002). Oral versus written administration of the Geriatric Depression Scale. Aging and Mental Health, 6, 418–422. Chaaya, M., Sibai, A. M., Fayad, R. and El-Roueiheb, Z. (2006). Religiosity and depression in older people: evidence from underprivileged refugee and non-refugee communities in Lebanon. Aging and Mental Health, 11, 1–8. D’Ath, P., Katona, P., Mullan, E., Evans, S. and Katona, C. (1994). Screening, detection and management of depression in elderly primary care attenders. I: The acceptability and
Validation of the Arabic version of GDS-15 performance of the 15 item Geriatric Depression Scale (GDS15) and the development of short versions. Journal of Family Practice, 11, 260–266. Farrag, A. K., Farwiz, H., Khedr, E., Mahfouz, R. M. and Omran, S. M. (1998). Prevalence of Alzheimer’s disease and other dementing disorders: Assiut-Upper Egypt study. Dementia and Geriatric Cognitive Disorders, 9, 323–328. Federal Interagency Forum, (2000). Older Americans 2000: Key Indicators of Well-being. Updated on 13 January 2005 last accessed on: 9 May 2005. Available from: http://www.agingstats.gov/tables%202001/tables-healthstatus.html#Indicator%2016. Fountoulakis, K. N. et al. (1999). The validation of the short form of the Geriatric Depression Scale (GDS) in Greece. Aging – Clinical and Experimental Research, 11, 367–372. Friedman, B., Heisel, M. J. and Delavan, R. L. (2005). Psychometric properties of the 15-item Geriatric Depression Scale in functionally impaired, cognitively intact, community-dwelling elderly primary care patients. Journal of the American Geriatrics Society, 53, 1570–1576. Hocking, L. B., Koenig, H. G. and Blazer, D. G. (1995). Epidemiology and geriatric psychiatry. In M. T. Tsuang, M. Tohen and G. E. P. Zahner (eds.), Textbook in Psychiatric Epidemiology (pp. 437–452). New York: John Wiley & Sons. Karam, E. et al. (2006). Prevalence and treatment of mental disorders in Lebanon: a national epidemiological survey. Lancet, 367, 1000–1006. Lam, C. K., Lim, P. P. J., Low, B. L., Ng, L. L., Chiam, P. C. and Sahadevan, S. (2004). Depression in dementia: a comparative and validation study of four brief scales in the elderly Chinese. International Journal of Geriatric Psychiatry, 19, 422–428. Malakouti, S. K., Fatollahi, P., Mirabzadeh, A., Salavati, M. and Zandi, T. (2006). Reliability, validity and factor structure of the GDS-15 in Iranian elderly. International Journal of Geriatric Psychiatry, 21, 588–593. Patten, S. B. and Love, E. J. (1993). Can drugs cause depression? A review of the evidence. Journal of Psychiatry and Neuroscience, 18, 92–102. Sheikh, J. I. and Yesavage, J. A. (1986). Geriatric Depression Scale: recent evidence and development of a shorter version. Clinical Gerontologist, 5, 165–172. Teh, E. E. and Hasanah, I. (2004). Validation of Malay version of the Geriatric Depression Scale among elderly inpatients. Available at http://www.priory.com/psych/MalayGDS.htm. Accessed 30 March 2007. U.S. Department of Health and Human Services, (1999). The Surgeon General’s Call to Action to Prevent Suicide, 1999. At a Glance: Suicide Among the Elderly. Updated 8 July 2005. Accessed on 4 October 2005. Available at http://www.surgeongeneral.gov/library/calltoaction/fact2.htm. Wancata, J., Alexandrowicz, R., Marquart, B., Weiss, M. and Friedrich, F. (2006). The criterion validity of the Geriatric Depression Scale: a systematic review. Acta Psychiatrica Scandinavica, 114, 398–410. Wrobel, N. H. and Farrag, M. H. (2006). A preliminary report on the validation of the Geriatric Depression Scale in Arabic. Clinical Gerontologist, 29(4), 33–46. Yesavage, J. A. et al. (1983). Development and validation of a geriatric depression rating scale: a preliminary report. Journal of Psychiatric Research, 17, 27.
581
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.