Cross-Cultural Adaptation and Validation of The Geriatric Depression Scale Into Igbo Language: A Validation Study


 Background: Late-life experiences such as protracted and indisposing medical disorders can negatively impact older adults’ psychological and mental health, making them vulnerable to depression. Majority of the assessment tools for depression were developed for use in western countries. There is therefore need for availability of culture- and environment- specific tools for assessment of depression in low and middle income countries. This study was designed to cross-culturally adapt and validate the Geriatric Depression Scale-15 (GDS-15) into Igbo Language and culture. Methods: The English version of the GDS-15 was translated into Igbo language; synthesized, back-translated, and underwent expert panel review, pretesting and cognitive debriefing interview, according to Beaton’s guideline. The Igbo version of the GDS-15 was tested for concurrent and structural validities and internal consistency among consecutively recruited 140 consenting older adults (62.9% females) in Enugu North Senatorial District, at 0.05 level of significance.Results: The English version of the GDS-15 was successfully cross-culturally adapted to Igbo with all the 15 items still retained on the Igbo version of the GDS-15. The Igbo version of the GDS-15 exhibited the same structure as the English version. The Igbo version of the GDS-15 has Cronbach’s alpha value of 0.53. The correlation between the participants’ total scores on the Igbo and the English versions of the GDS-15 (rho = 0.86) was excellent. There was no significant difference between corresponding scores in the English and Igbo versions of the GDS-15 (p = 0.89).Conclusions: The Igbo version of the GDS-15 is a valid and culturally specific instrument, and can be used for assessing depression among Igbo older adults in Nigeria.


Introduction
Globally, there is a rapid increase in the number of people aged 60 years and above [1,2]. In most of the developed and developing countries, the age of 60-65 is considered equivalent to retirement age and it is said to be the beginning of old age [3]. About 3.4% or 5.9 million of the total Nigerian population are older adults (aged 65 years and above) [4]. Though ageing population is a sign of victory of human development [5], it comes with challenges like depression in the society [1].
Depression is a common psychological illness characterized by persistent sadness and a loss of enthusiasm in activities that individuals usually enjoy, accompanied by an inability to accomplish daily activities [6]. It might be in uenced by incidental and unplanned negative life events, such as death of loved ones, elderly abuse, physical and cognitive decline, retirement, restricted mobility, limited nancial resources, seclusion, and an inability to participate in once-cherished activities, and so on [7][8][9][10]. Common indications of depression among older adults include feelings of hopelessness, helplessness, weight changes, persistent sadness, and unnecessary stresses over funds, and so on. It often spurs thoughts of suicide and suicidal mortality in this older population [11]. Individuals who are single, unmarried, separated or widowed, and female are at greater risk of having depressive symptoms [8,12].
However, late-life depression, in spite of its prevalence and effect, may be di cult to diagnose and insu ciently treated [23,24]. Active screening, early detection and treatment may be effective in reducing healthcare cost and improving case nding for depression among older adults [25,26]. Therefore, there is need for availability of validated instruments for assessing depression [17,26].
Many of the measurement tools used to study depression among older adults are adaptations of instruments developed in other cultural settings [27]. This includes the EURO-D, the Beck Depression Inventory, the Geriatric Depression Scale (GDS), the Center for Epidemiological Studies Depression Scale, and the Zung Self-Rating Depression Scale [26]. The most commonly used depression screening tools for older adults are the GDS and the Center for Epidemiological Studies Depression Scale [17]. The original GDS is a psychometrically-sound 30-item simple-to-administer questionnaire that has been utilized generally among community and institutionalized older adults [28,29]. Sheikh and Yesavage [30] developed and validated a 15-item GDS Short Form so as to improve its acceptability for the older adult population and make this measure an easier screening tool for depression.
Diagnostic scales developed in the western world may be inappropriate for use in developing countries.
Furthermore, the possibility of cultural effects confounding the measurement of depression may in uence the individual responses to questionnaires; prompt misdiagnosis or improper assessment of depression [26, 31,32]. It is therefore advisable and pertinent to cross-culturally adapt an instrument before use in a different culture and setting [33]. Cross-cultural adaptation of an already existing instrument instead of developing a new one has often been the choice when addressing a population whose language is different from the source language, as it is more economical and ensures ease of comparisons among populations [33][34][35][36]. Cross-cultural adaptation comprises the translation, adaptation and validation phases.
According to the National Bureau of Statistics [37], the Nigerian adult literacy in English Language is 42.1%. With this literacy level, it is imperative to make provision of a validated instrument for the measurement of depression within Igbo population especially for older adults who are non-literate and can neither communicate nor understand English language. Furthermore, mere translating the tool to this group of participants by different assessors will equally introduce some level of biases as the translations are not validated and may differ signi cantly with individuals. The GDS-15 has been utilized in various languages and cultural settings to assess depression, and has proved to be a reliable and valid instrument [29,38,39]. The GDS-15 has been translated into various languages: Iranian [40], Japanese [41], Malay [42], Chinese [43], Spanish [44], Nepalian [38]. However, there is no Nigerian adapted and validated version of the GDS-15. This study was therefore designed to translate, culturally adapt and validate the English version of the short form of GDS-15 into Igbo language. Igbo language is one of the three major Nigerian native languages (making up about 18% of the whole Nigerian population) and a minor language in Equatorial Guinea, with over 24 million speakers [45][46][47].

Design
This was a validation study that utilized the recommendations for cross-cultural adaptation established by Beaton et al. [33] for the American Academy of Orthopaedic Surgeons. The study protocol was approved by the Ethical Review Committee of the University of Nigeria Teaching Hospital, Ituku-Ozalla, Enugu (NHREC/05/01/2008B-FWA00002458-1RB00002323). Permission to translate and validate the original English version of the GDS-15 was gotten from the developers. Informed consent was obtained from all the consecutively recruited older adult participants (65 years and above) in Enugu North Senatorial District, Nigeria. The purpose and procedures of the study were thoroughly explained to the respondents. Participants were assured of con dentiality and the freedom to withdraw from the study when they were no longer interested. The inclusion criteria for this study include older adults (≥65 years) who could understand both English and Igbo languages; were well-oriented in time, place and person; and were not on any antidepressants. These eligibility criteria were for both the adaptation and validation phases of the questionnaire. Information on participants' socio-demographic: age, gender, occupation, level of education, and marital status was obtained through oral interview.

Instruments For Data Collection
The GDS-15 The Geriatric Depression Scale is used in the assessment of depression among older adults. The GDS, rst created by Sheikh and Yesavage [30] has been tested and used extensively with the older population.
The GDS is a time-saving instrument with a response format of (Yes/No), which brought about its use in different healthcare setting [48]. The GDS Long Form is a brief, 30-item questionnaire in which participants are asked to respond by answering yes or no in reference to how they felt over the past week.
Questions from the Long Form GDS which had the highest correlation with depressive symptoms in validation studies were selected for the short version. A scoring grid accompanies the GDS.

Cross-cultural Adaptation And Validation Of The Gds-15
This study followed the American Academy of Orthopaedic Surgeons' guidelines for cross-cultural adaptation developed by Beaton et al. [33]. The procedure for the study was performed in sequential order: Phase 1-translation phase, Phase 2-adaptation phase and Phase 3-validation phase.

Phase 1 -Translation Phase
This stage involves forward translation, synthesis of the translation and back translation [33,34]. The forward translation involved the translation of the English version of the GDS-15 into Igbo language. Two bilingual translators, a linguist and a physiotherapist, whose mother tongue was Igbo language, independently translated the questionnaire to produce two different Igbo translations (FT1 and FT2). The physiotherapist was aware of the concepts of the questionnaire being translated. The purpose of this translation was to offer equivalence from a more clinical perspective and also to produce a translation providing a more reliable equivalence from a measurement perspective. The linguist has no medical or clinical background, and was neither aware nor informed about the concepts being quanti ed. This second translator was supposed to provide a language translation as used by the population, in order to accentuate obscure meanings in the original questionnaire [46].
The two translations were then harmonized and synthesized by the two translators to produce a single quality-assured version (ST-12). The ST-12 was translated back to English language (as a quality control step) by two different bilingual translators who were oblivious of the concept under study, producing two back-translated English versions BT1 and BT2. These translators were physiotherapy lecturers who were uent in both English and Igbo languages, and also knowledgeable in cross-cultural adaptation procedure. The essence of this procedure was to ascertain the validity of the questionnaire in order to ensure that the translated version reproduced similar item content as the original version [34].

Phase 2 --Adaptation Phase
The original English GDS-15 version (E-GDS), the two forward translations, the synthesized version, and the back translations underwent an expert panel review in order to produce a version expertly crossculturally adapted to Igbo language, culture and environment. The expert panel comprised the four translators, four physiotherapy researchers, and a lay person. The expert committee were very familiar with the Igbo culture and environment, and identi ed the comprehensiveness, relevance and comprehensibility of the instructions, items and response options. All the translations were carefully deliberated on while ensuring experiential, operational, measurement, conceptual, idiomatic, and semantic equivalences. All disagreements in the translations were settled by consensus, thus producing a pre-nal Igbo version of the questionnaire which was subjected to eld testing by administering it to 30 older adults (62.9% female) who were consecutively recruited from conveniently selected communities in Enugu North Senatorial District (who met the inclusion criteria) after informed consent had been sought and obtained from the participants. These participants were taken through cognitive debrie ng interview in order to ascertain what each participant thought was meant by each item on the questionnaire and the responses; ease of understanding each item; if there was ambiguity in each item; if the response items for each item were di cult to understand; and if the activity described in each item was being experienced in Igbo culture. Each participant was expected to answer 'YES' or 'NO' for each question.
The answers on the cognitive debrie ng checklist were then presented to the expert panel on a second meeting. Items or response option with less than 80% YES were supposed to be amenable to changes. Item 4 "Ọ bụ na ịdịghị enwekarị mmụọ mkpali?" scored 60%. This item was misunderstood/misconstrued by the participants. It was therefore amended to "ị naghị enwe uchu ịme ịhe ọ bụla?". However, the remaining items had at least 80% YES. No further modi cation was made on the questionnaire by the panel of experts. The nal Igbo version of the GDS-15 (I-GDS-15) was consequently produced. PCA is a dimensionality-reduction method that is often used to transform the variables into simples ones, increasing interpretability, while still retaining the original information. The standard error of mean (SEM) and the minimal detectable difference (MDD) of the total and item scores on the I-GDS were determined. The MDD was estimated using the following formula: MDD = 1.96 x SEM x √2 [50].
Before performing the PCA, the Barlett's test of sphericity and the Kaiser-Meyer-Olkin (KMO) were used to check the appropriateness of the data for factorial analysis. To ensure the factorability of the data, Barlett's test of sphericity value must be signi cant [51], and the KMO value must exceed the recommended value of 0.6 [52]. All the coe cients of correlation of each of the items on the I-GDS-15 with one another must all exceed 0.3; and the communalities must all be above 0.3 [53,54]. Communality values of less than 0.3 might suggest that the item did not t well with the other items loading on the same component. During PCA, any factor with its eigen values exceeding one are often retained. The number of retainable factors could also be revealed through the scree plot by counting off the number of points before a clear point of re ection. The scree plot was used to examine further the number of retained factors [55]. Every component with initial eigen value lower than the random eigen value are usually rejected. Level of signi cance was set at p <0.05.

Cross-cultural adaptation of the GDS-15 into Igbo
The pre-nal version of the I-GDS-15 was pretested on 30 older adults who were taken through cognitive debrie ng interview. Some modi cations were made on four of the items during the expert panel meeting. The word "basically" in item 1 was deliberated on for the best Igbo synonym that captured the context. "Ozuzu oke" was contemplated on but was dropped by the panel as it was thought not to convey the indepth meaning of "basically". After deliberations on several likely alternative/synonym for it, the panel opted for "o pekata mpe". In item 3, the word "empty" was taken to mean "void". The initial Igbo translation "abaghịzi uru" was argued to mean "useless". Studying one of the forward translations, the panel agreed with the translation "chakoo" to mean "emptiness". "Good spirits" in item 5 was initially translated to be anụrị /obiọma. Being in good spirits and being happy were argued to mean the same thing. While some argued that good spirits could mean cheerful and so "obi ọma/ anụrị", others disagreed. After many considerations, it was literally translated to "ezi mmụọ". However, it was argued that it may seem offensive to some individuals on cultural/belief basis. The panel therefore deliberated further for a more polite appropriate and acceptable word. "Obi sara sara" was nally agreed to be a better option that conveys the meaning. The word "helpless" in item 8 was also literally translated to "enweghị enyemaka". This was argued that it does not represent the contextual meaning. The panel agreed that it does not mean being without assistance/aid. The word "helpless" was highlighted as being a dead-end situation that only a miracle can solve (that can be salvaged only by a miracle). After several disagreements, the panel opted for "nkoropu obi".
However, during the pretesting of this pre-nal I-GDS-15, all the participants indicated clarity of language and ease of understanding of all the items during cognitive debrie ng interview except for two items, and so led to some modi cations at the second expert meeting. The term "mmụọ mkpali" in item 4 was misunderstood to be "insult". It was modi ed to "Ị naghị enwe uchu ịme ihe ọ bụla?". In item 13, "ume" and "ike" were thought to mean similar things. The expert review opted for "ike". No further modi cations were made, and thus the nal I-GDS-15 was produced. Table 1 summarises the modi cations ( nal I-GDS-15).

Validation Of The Igbo Version Of The Gds-15
Socio-demographic pro le of the participants One hundred and forty apparently healthy older adults (37.1% males, 62.9% females) who were residents in Enugu North Senatorial District were involved in the psychometric testing of the I-GDS-15. The description of the participants is displayed on Table 1.

Concurrent Validity Of The Igbo Version Of The Gds-15
Spearman correlation coe cients of the items and total scores ranged from (rho = 0.271 -1.000; p<0.001) indicating evidence of good concurrent validity in the scores on the Igbo version, thus supporting the use of the components as separate scales ( Table 2). The correlation coe cients of many of the items on the I-GDS with one another did not exceed the recommended value of 0.3. This explains the not-so-high internal consistency coe cient (0.53) reported in this study. Total score on the E-GDS-15 and the I-GDS-15 as pictorially shown on the Bland-Altman plot revealed evidence of homoscedasticity of the two scores (Fig. 1). The Scatter diagram demonstrates the correlation between the total scores on the E-GDS-15 and the I-GDS-15 (Fig. 2). The Mann-Whitney U test of difference between the total score on the two versions of the questionnaire showed no signi cant difference (U= 9710.500; p = 0.89), thus revealing that the two instruments produced equivalent conceptual and linguistic meanings (Table 3).

Reliability Analysis
Cronbach's alpha for the internal consistency coe cient of all the items on the I-GDS-15 was 0.53. This value suggests poor/adequate internal consistency of the items on the I-GDS-15. The standard error of mean and minimum detectable difference values for the total and item scores on the I-GDS ranged from 0.01 (item 11) to 0.16 (total score) and 0.03 to 0.44 respectively (Table 4). This revealed evidence of the responsiveness of the item and total scores on the scale.

Structural Validity Of The Igbo Version Of The Gds-15
Page 9/23 The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy was 0.67 exceeding the recommended value of 0.6 [52], while the Barlett's test of sphericity was statistically signi cant (X 2 = 306.352; p < 0.001); thus suggesting that factor analysis was appropriate and feasible.  (Table 5). Scree plot also depicted the presence of these ve factors (Fig. 3). The communality values of all the items were all above 0.3 indicating that all the items t well with other items (Table 6).

Discussion
This study was performed to cross-culturally adapt and validate the Igbo version of the GDS-15 among Igbo older adults in Enugu, Southeast Nigeria following an established and accepted protocol [33] in order to minimize bias and eliminate erroneous results. The GDS-15 has been translated into various languages: Iranian [40], Japanese [41], Malay [42], Chinese [43], Spanish [44], The Cronbach's alpha of the I-GDS-15 was 0.53, thus suggesting a poor internal consistency. This value is lower than the 0.82 Cronbach's alpha of the original version [30]; and those reported in the Iranian: 0.92 [40] We make recommendations for further studies to address test-retest reliability. This could be helpful in the context of the low alpha reliability (0.53) reported in this present study.

Conclusions
The I-GDS-15 is a valid and an acceptable instrument for use in measuring depression among Igbo- All methods were carried out in accordance with relevant guidelines and regulations.

Consent for publication
Not applicable Availability of data and materials The dataset used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
The authors did not receive any support from any organization for the submitted work. Scree plot of the components on the Igbo version of the Geriatric Depression Scale-15