This methodological study was approved by Institutional Ethics Committee of Ahmedabad Institute of Medical Sciences with ethics approval number AIMS/2016/75. The study was carried out for the period of seven months, from November 2016 to May 2017 at Ahmedabad Institute of Medical Sciences, Gujarat, India and various old age centres at Ahmedabad.

The method of translation and cultural adaptation is described in [Table/Fig-1]. The developers and authors of ACE III were approached with the proposal to translate and culturally adapt ACE III to Gujarati. The developers provided the Australian version of ACE III with its scoring guidelines [15]. The ACE III and its components were studied in detailed. The ACE III was then translated to Gujarati by two independent translators who were native speakers of Gujarati language and fluent with English language. As per the guidelines by Rose M et al., the translators were chosen with different background [16]. Out of the two translators, one translator was familiar with the concepts being translated whereas the other was not. The first translator possessed the background of physiotherapy and healthcare, and the other translator had a background of literature and education. The two translations were then reconciled to one as first translation copy. The first translation copy was then studied by the first and the third author and report was prepared by the first author. The report included requirement of changes of certain terminologies and changes in different domains.

[Table/Fig-1]:

Method of translation of ACE III to Gujarati.

For Fluency domain, “The Letter Task” instructions were modified into relevant Gujarati instructions instead of opting for a literal translation. Words starting from letter “C” like cat, cry, and clock are included as an example in the original English instruction for the letter task. Instead of translating the words of letter “C”, different set of Gujarati words starting with same alphabet were used to give example.

Repeating and recalling name and address for memory domain was changed to a Gujarati name and a local address. General knowledge 12questions for memory domain were changed to relevant questions for Gujarati population.

Language domain of ACE III includes six tasks. Out of which “pencil and paper task” and “repeat the words” were retained as it is with Gujarati translation. In “repeat the proverbs”; the proverbs were changed as per the relevance in Gujarati language. “Name the pictures”; has 12 pictures to be named and 4 questions to be answered based on the pictures. Out of 12, the picture of “kangaroo was changed to cow”, “penguin to peacock”, “anchor to hammer”, “harp to tabla”, “barrel to bucket”, “crown to bicycle”, and “accordion to flute”. Rest of the pictures were retained as they were. The questions based on the pictures were changed accordingly. “Read the words” were changed to Gujarati words having same level of difficulty but different meaning.

Visuospatial Ability domain was retained without any changes except “identifying the letters”. The letters were changed to Gujarati alphabets having same pronunciation as English letters.

The first copy thus formed, was then translated back to English by an independent non-professional translator who was a bilingual physiotherapist but who was not involved with geriatric, neurology or neuropsychiatric field. As the translator was not aware or informed of the concepts being considered in the translation, it not only prevented information bias but also better prompted, the unexpected meanings of the items in the translated tool. The differences between the translated copy and original English version were documented. The first copy was further modified as per the requirements.

The process of forward-backward-forward translation method continued till the acceptable second copy of Gujarati translation was obtained [17].

The procedure for the validity of the ACE III is described in [Table/Fig-2]. For the face validity, the second copy of the instrument was given to seven professionals working in neurological medicine and neuro-rehabilitation field with the mean experience of 12 years [18]. Each item was analysed as accepted, accepted with modification or rejected. The report of the items requiring modification was prepared and modified till all the items reached common consensus. None of the items were listed as rejected. The third copy of Gujarati translation was prepared and tested for content validity.

[Table/Fig-2]:

ACE III validity.

For the content validity, Rose M et al., and Haccoun RR, technique of single group was used. It requires bilingual participants [16,17]. The sample of 10 adults, both males and females in the age group of 60 to 75 years, were selected. They were native speaker of Gujarati and were fluent in English language as well. They were tested and interviewed with both English and Gujarati versions of ACE III in random orders. The subjects were asked if they understood the components in both the scales equally well or if they felt that there was a need to replace any word or instructions for better understanding. Accordingly, a report was written and minor corrections were made as suggested.

After all modifications, the fourth version of the Gujarati translation was prepared. It was thoroughly proof read for any errors. The final version was then submitted to the developers.

The data was analysed using Microsoft Excel 2010 and SPSS version 16.0. The data was analysed for the test of normality by using normal distribution curve and where it was found to be normally distributed, the parametric test was applied and where it was not found to be normally distributed, the non-parametric test was applied.

For face validity, the expert panel consisted of seven professionals working in Neurological medicine and Neuro-rehabilitation Department of Civil Hospital, SBB College of Physiotherapy and Ahmedabad Institute of Medical Sciences, Ahmedabad, Gujarat, India. Their age and experience in years are listed in [Table/Fig-3].

The mean±SD age of the older adults who participated in the validation was 67±4.71 years. There were six males and four females. The [Table/Fig-4] shows Pearson’s correlation of English and Gujarati version of ACE III. The [Table/Fig-5] shows domain wise results of Gujarati and English ACE III.

The older participants’ mean±SD age at the time of completion of their formal education was 20±3.45 years. The [Table/Fig-6] shows Spearman’s correlation of level of education with English ACE III and Gujarati ACE III.

The purpose of the study was to translate, culturally adapt and validate the Gujarati version of ACE III. Mathurnath PS et al., performed a study to translate ACE to Malayalam and culturally adapt and validate ACE to Malayali speaking population in southern India [14]. Ganguli M et al., performed a study to translate Mini Mental Status Examination (MMSE) to Hindi speaking population in North India [19]. There is a vast cultural difference between different regions of India. So, it requires changes in an English tool to culturally adapt it for Indian population. Also, different language and region prompts for changes unique for that particular culture.

The result indicates that there is a strong positive correlation between the original English version and translated Gujarati version of ACE III. Rose M et al., in their study described that strong correlation between the original tool and the translated version of the tool not only indicates that the quality of the translation is good but also indicates that both the measures provide same information in both the languages [16]. dos Santos Kawata et al., performed a study to validate Japanese Addenbrook’s Cognitive examination - Revised (ACE-R) but, they correlated ACE-R with clinical dementia rating having moderate correlation between the said two outcomes [20].

While analysing domain wise results, attention, fluency and visuospatial abilities shows no difference between Gujarati and English versions. But, in memory, subjects have scored more in Gujarati version compared to English version. For Language as well, the subjects scored more in Gujarati version compared to English version. In a previous study of translation of ACE to Malayalam language, the researchers assessed the cognitive level of older individuals residing in Kerala with ACE Malayalam, and older individuals residing in United Kingdom with ACE UK version. In their study, the scores of both the groups were comparable for all the domains [14]. This is contrary to the present study, as in the previous study [14]; the individuals were tested in their place of residence in their spoken language. In the present study, the English version required the subjects living in India, to remember and recall the address which has English name and the areas of Australia, for which most subjects failed. The name and the address were Harry Barnes, 73, Market Street, Rockhampton, Queensland. Compared to this, Gujarati name and address were more relevant and easier to remember and recall.

The language has a component of naming several pictures. The subjects also have to answer the questions based on those pictures. This particular task has maximum score of 12. Many pictures in the English version were not relevant to Gujarati population and that explains why subjects in Gujarat scored less in Language domain in English version.

The correlation of education with English ACE III is 0.75 with p-value 0.05, and correlation with Gujarati ACE III is also 0.75 with p-value 0.05. This indicates strong positive correlation between Gujarati ACE III and level of education. Matias-Guiu JA et al., while translating the ACE III to Spanish had similar findings and also indicated that education exerts significant effect on test scores [21].

The strength of this study is that the translation procedure followed a strict translation and validation protocol described by WHO [22]. The comparative analysis between both the versions was done extensively.

The limitation of the study is that the average time to execute the Gujarati ACE III was not established as the participants underwent an interview while going through Gujarati ACE III test.

This study produced equivalent Gujarati version of ACE III which is comparable and significantly correlating to Original English version of ACE III. Early detection and early intervention can be of great importance for the individuals with cognitive decline. Gujarati ACE III becomes one of those few tools available for free to detect cognitive impairments. Screenings and differential diagnosis of conditions like MCI, Dementia and Alzheimer’s disease can be done easily for Gujarati population with the help of Gujarati ACE III. In future, a study can be carried out to determine an average time to execute the Gujarati ACE III. In addition to this, a study to establish optimal cut off score for Gujarati population and a study to establish intra rater and inter rater reliability can also be carried out. Also, studies to find prevalence of MCI and Dementia in Gujarati population can be conducted.