Vietnamese Version of the Geriatric Depression Scale (30 Items): Translation, Cross-Cultural Adaptation, and Validation

The proportion of geriatric depression recorded in Vietnam was 66.9%. Depression in older people is a risk factor for problems related to dementia, poor quality of life, and suicide. To have a good Vietnamese questionnaire for assessing geriatric depression, we conducted the study to translate and cross-culturally adapt the Geriatric Depression Scale—long-form with 30 items (GDS-30). The study has two steps. Step 1 is a translation of the GDS-30 scale. We followed the guideline by Beaton et al., (2000 & 2007). Firstly, two translators (informed and uninformed) translated the questionnaires. Secondly, the translations were synthesized. Thirdly, back translation was performed by two translators fluent in both Vietnamese and English but completely unknown of the original version of the scale and did not have medical expertise. Finally, seven experts reached a consensus on the pre-final Vietnamese version (GDS-30). Step 2 is a field test of the questionnaires on people 60 years or older. Then, we determined the internal consistency and test-retest reliability of the questionnaire in 55 Vietnamese inpatients in a geriatric department. Construct validity was determined by examining the relationship between depressive scores and patient characteristics. The Vietnamese version of GDS-30 was built with the agreement of all experts on the semantic, idiomatic, experiential, and conceptual equivalences between the original and pre-final Vietnamese versions of the GDS-30. The Cronbach’s alpha coefficient value was 0.928, indicating the items’ adequate internal consistency. Spearman’s correlation coefficient value of total scores between the first and second interviews showed medium correlation (0.479, p < 0.001), and the stability is acceptable. The GDS-30 scale reached the construct validity because the proportion of geriatric depression according to GDS-30 was significantly different between characteristics groups, such as gender, employment, level of education, economic status, and sleep disturbance. The Vietnamese version of the GDS-30 scale had high consistency, satisfactory reliability, and understanding and can be used as a screening tool for depression in elderly patients in primary healthcare centers. This is the first depression rating scale for the elderly in Vietnam to be translated and validated. Non-psychiatric health professionals or patients can quickly self-assess and screen for the illness.


Introduction
In 2020, the proportion of people aged 65 and over accounted for 9.3% of the total population, equivalent to 727 million people in the world [1]. In Vietnam, the proportion of older people (aged 60 years and older) has increased quite rapidly since the beginning of the 20th century, rising from 8.1% (in 1999) to 8.6% (in 2009) and reaching 11.9% in 2019. This rate has been projected to increase to 28.3% by 2050 [2][3][4]. Aging is associated with a number of factors that result in the deterioration of physical and mental health.
Depression is one of the most common mental disorders in the elderly [5]. In 2018, the proportion of geriatric depression recorded in Vietnam was 66.9% [6]. Older people with depression are at risk for several additional problems, such as dementia, poor quality of life, and suicide [7][8][9]. However, a 2012-2013 survey of 33,653 physician-patient encounters found less than 5% of adults were screened for depression in primary care [10]. Depression is popular in older adults, especially those with multimorbidity, such as thyroid disease, diabetes, heart disease, and other chronic medical conditions [11][12][13]. Depression and hypertension have also been shown to have interaction effects in a physiological way. In addition, depression could seriously affect their attitude of medication adherence, thereby decreasing their blood pressure control and quality of life, further aggravating the situation and creating a pathological spiral [14].
In 1982, Yesavage et al. developed the self-rated Geriatric Depression Scale (GDS) to screen for depression in the elderly [15]. A team of clinicians and researchers involved in geriatric psychiatry selected 100 questions believed to have the potential for distinguishing elderly depressives from normal subjects with various elements addressing cognitive complaints, motivation, future/past orientation, personal mood, etc. 30 questions with the highest correlation with the total score were chosen to create the GDS-30 scale [15]. The original version of this scale was written in English, and it has subsequently been translated and widely used in many Asian countries, such as India [16], Korea [17], and the Philippines [18].
The India version of this scale was conducted in the rural community of Ballabgarh in northern India with 1554 samples, mostly illiterate Hindi-speaking residents of Ballabgarh aged 55+. Although the large sample was selected, it was just distributed in a rural area, so the difference of depression in non-illiterate and illiterate people was unreliable [16]. The Korean version provided valid and reliable case-finding tools for screening major depression among the elderly psychiatric patients in Korea; however, the sample size of this study was not big enough to represent the Korean population [17]. The Philippines version of the GDS was performed in 505 elderly respondents who gave informed consent to participate in the study. Participants were required to be age 60 or older to comprehend both written and verbal English and Filipino and to exhibit no evidence of cognitive impairment. Respondents were required to complete all items from the English and the Filipino versions of the GDS. However, the result of this study was the GDS-15 (the short form of the original version), so it cannot fully evaluate different aspects of depression in the elderly [18].
Many popular depression rating scales have been translated into Vietnamese, such as the Center for Epidemiological Studies-Depression Scale (CES-D) [19], Patient Health Questionnaire-9 (PHQ-9) [20], and Zung Self-Rating Anxiety Scale (Zung SAS) [21]. However, these scales were not designed specifically for the elderly, so this study aims to translate and validate a Vietnamese version of the GDS-30 scale to screen for depression in the elderly.

Translation of the GDS-30 Scale
The GDS-30 scale was translated from the original English version of Yesavage et al. (1982) into a Vietnamese version [15]. It had 30 questions, including content related to depression of the elderly; the interviewees answered with two options (yes or no); depending on each question, the answer "yes" or "no" was counted as 1 point. Geriatric depression was assessed based on cumulative scores in two ways: (1) depression rating scores were divided into 2 groups, identified a depression (≥10 points) and no depression (<10 points); (2) depression rating scores were divided into 3 groups, identified as no depression (0-9 points), mild depression (10-19 points), and severe depression (20-30 points).

Study Design
The Vietnamese version of the GDS-30 scale was validated among elderly hypertensive inpatients (aged 60 years and above) at the Department of Geriatrics, Can Tho Central General Hospital, from May 2020 until March 2021. Sampling criteria included patients diagnosed with hypertension with systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg [24] or were being treated for hypertension and agreed to participate in this study. Regarding the subject exclusion criteria in our study: any participant who has an acute illness, hearing impairment, language barrier, poor communication, a serious life problem within two weeks, and severe dementia were not included in our study.

Data Collection
Participants were invited to a face-to-face interview through prepared questionnaires. We also conducted a second face-to-face interview on previously interviewed patients to assess the test-retest reliability, excluding patients discharged from the hospital. The length between the two interviews ranged from 2 weeks to a month [25]. This study conducted second interviews after 7-14 days (2 weeks) because of the limited time.

Sample Size
The response rate of 1:10 (i.e., each question required 10 patients) was chosen to calculate the required sample size using established methods [26,27]. The GDS-30 scale had 30 questions, so the estimated sample size was 300 patients. An additional 10% were selected to accommodate lost participants during research time; the required sample size was 330 patients.

Validation Criteria Reliability
Internal consistency was estimated using Cronbach's alpha coefficient. The scale was considered adequate internal consistency when the Cronbach's alpha value was > 0.5 [28].
Test-retest was based on the repetition between results of the first and second interviews and evaluated using Spearman's correlation coefficient. The Spearman's coefficient value > 0.3 and p-value < 0.05 was considered good test-retest reliability [29,30].

Validity
Content validity was based on the expert committee's review and the equivalence score between the source and target version of the scale. Content validity was accepted when all questions received the consensus of more than half of the expert committee members.
Construct validity was evaluated by assessing the relationship between results of the first response and depression-related characteristics, such as age, gender, current occupation, education level, marital status, economic status, sleep disorder according to the Pittsburgh Sleep Quality Index (PSQI). The outcomes were considered statistically significant if the p-values were < 0.05.

Data Analysis
Data were entered using Epidata 3.0 and were processed using SPSS version 18.0 (IBM Corp., New York, NY, USA). Descriptive statistics describe the proportions, frequencies of categorical variables, and the mean, standard deviation (SD) of continuous variables. The scale's reliability was assessed by internal consistency (Cronbach's alpha coefficient) and test-pretest (Spearman's correlation coefficient). The chi-squared test with odds ratios (OR) and 95% confidence intervals (CIs) was applied to assess the relationship between depression and depression-related characteristics. The principal component analysis method (PCA) with varimax rotation was used to determine the factor structure of the GDS-30 scale.

Ethics Approval
The Institutional Review Board approved this study for Ethics in Biomedical Research-Hanoi Medical University on 10 April 2020 (approval number: 72/GCN-HÐÐÐNCYSH-ÐHYHN). The purpose and content of the study were explained clearly and specifically to the participants, who signed a consent form. Participants also had the right to refuse to participate in the study without affecting treatment.

Results
The table showed the geriatric depression scale in the original English version and Vietnamese version after translation according to the above process (Table 1).

Consistency
The Cronbach's alpha coefficient value was 0.928, indicating the items' adequate internal consistency (Table 3).

Content Validity
Between the source and target version of the GDS-30 scale, the expert committee assessed the average score of 0.84 points for experience equivalence. The other criteria (semantic, idiomatic, and conceptual equivalence) achieved an average score of 0.88 (Table 5).

Construct Validity
There was a significant difference in the rates of depression between gender, employment, education level, economic status, depression diagnosed according to ICD-10 and sleep disturbance according to PSQI (Table 6).  The factor loadings ranged from 0.40 to 0.8. The factorial analysis produced 5 factors for the GDS that represents 52.874% of the variance (Table 7). The first factor was composed of 7 items: 7,8,6,13,9,4,16. This contributed to 33.57% of the total variance. This factor was called "sad mood".
The second factor was composed of 4 items: 15, 5, 19, 21. This contributed to 5.73% of the total variance. This factor was called "positive mood".
The third factor was composed of 5 items: 23,24,18,25,17. This contributed to 4.62% of the total variance. This factor was called "agitation and pessimistic" The fourth factor was composed of 3 items: 12, 28, 2. This contributed to 4.62% of the total variance. This factor was called "social withdrawal".
The fifth factor was composed of 2 items: 14, 30. This contributed to 3.88% of the total variance. This factor was called "cognitive inefficiency".

Discussion
This study aimed to translate and validate the first Vietnamese version of the GDS-30 scale for people over 60 years of age. The results showed that this scale could be a reliable screening tool for geriatric depression in Vietnam. Our findings are consistent with previous studies that have demonstrated that this scale can be used effectively to screen for depression in the elderly in many countries worldwide with different versions of the number of items [31,32].
Early diagnosis of depression in the elderly is essential. In our study, the proportion of elderly participants identified with depression was 41.5% according to the GDS-30 scale (≥10 points), a finding that is consistent with rates of depression in the elderly in many studies worldwide, which vary from 20.7 to 53.8% [33][34][35][36] among different countries and using varying diagnostic approaches In our study, the Cronbach's alpha coefficient value was 0.928, which showed very good internal consistency of the items [28]. All items had a suitable composite coefficient of correlation (≥0.3). "Cronbach's alpha coefficient value if items were deleted" of all items was lower or equal to Cronbach's alpha coefficient value of this scale, so no item was excluded from the scale. Spearman's correlation coefficient between the two interviews showed a medium correlation (0.479 with p < 0.001) regarding scale stability. After the first and second interviews, there were 6 items with no stability, including 3 positive items (1,15,19) and 3 negative items (3,23,26). Therefore, the patient's emotions were not affected by their response. This study also determined the efficacy of the Vietnamese version of the GDS-30 scale compared with the ICD-10 (gold-standard) in diagnosing depression in older adults. The GDS-30 scale reached construct validity because the proportion of geriatric depression according to GDS-30 was significantly different between characteristics groups.
Previous validation studies also illustrated the high consistency of the GDS-30 scale. In the Netherlands version, the Cronbach's alpha coefficient value was 0.88 [37] and showed a good level of consistency [28]. For other language versions, Cronbach's alpha coefficient values ranged from 0.839 to 0.91 [38,39]. These data showed that the Vietnamese version of the GDS-30 scale was reliable and had adequate internal consistency.
In the Korean version, the test-retest reliability (Pearson correlation) was 0.91 (p < 0.01), indicating that the performance of the GDS-30 is highly stable over time [40]. In the Italian version, the reliability of scale after re-testing (ICC) was 0.91 [39]. Period conducted a re-test in the two studies mentioned above ranged from 1-7 days, while our study conducted a re-test in 7-14 days. The older patients could not remember exactly what happened in the past. Therefore, Spearman's correlation coefficient value in this study was average.
Our study proved that the Vietnamese version of the GDS-30 scale had high consistency, satisfactory reliability, understanding. It can be used as a screening tool for depression in elderly patients in primary healthcare centers. Non-psychiatric health professionals or patients can quickly self-assess and screen for depressive symptoms. However, because of time limitation, our study could not be conducted in the community. Assessing the sensitivity and specificity of the Vietnamese version of the GDS-30 also needs to be considered.

Conclusions
This is the first depression rating scale for the elderly in Vietnam to be translated and validated. The Vietnamese version of the GDS-30 scale had high consistency, satisfactory reliability, and clarity. This study was conducted in a hospital and given the complexities of the population in this setting; future studies conducted among outpatients would be useful in further clarifying the factor analysis aspects of the Vietnamese translated GDS-30 items. Future studies will also be needed to assess the sensitivity and specificity of the Vietnamese version of the GDS-30. Still, it appears that the Vietnamese version of the GDS-30 can be used as a screening tool for depression in elderly patients in primary healthcare centers by non-psychiatric health professionals or patients to quickly self-assess and screen for depressive illness.