Introduction

Dementia is a progressive disease of the brain that impairs several neurocognitive functions including memory, orientation, and reasoning. Individuals with dementia often experience deficits in learning ability, emotional regulation, and social interaction1. In 2019, the estimated number of dementia cases worldwide was 57.4 million, and this number is expected to nearly triple by 20502.

In the elderly, dementia is a chief contributor to disability and dependency, resulting in a substantial economic burden on the government and the community. This can devastate health and social amenities due to an estimated increase in the cost of dementia care to two trillion US dollars by 2030. As a response, the World Health Organization (WHO) launched a Global Action Plan aimed at addressing the impact of dementia3,4,5.

Obesity, diabetes, and hypertension are among the 12 modifiable risk factors associated with the disease6. In fact, the prevalence of these risk factors is notably high in the Middle East and North Africa (MENA) region7,8,9. By modifying the risk factors, 40% of dementia cases might be delayed or prevented, placing immense importance on raising awareness in the region6.

Globally, studies have shown that the level of knowledge about dementia is fair to moderate, with a common misconception that dementia is a natural part of aging10. In the MENA region, there is a limited number of studies on the public’s knowledge of dementia. Based on the existing literature in the region, there is an inadequate level of awareness regarding the disease11,12,13.

In the UAE, the number of dementia cases is expected to see an exponential increase of 1795% by 20502. This can be partially attributed to the anticipated growth in the percentage of UAE residents above the age of 65 from 1.6% in 2020 to a projected 18.5% in 205014.

To reduce the burden of dementia and facilitate early prevention and management, a reliable tool is essential for measuring dementia knowledge. The Dementia Knowledge Assessment Scale (DKAS), developed by Annear et al. (2015), fulfills this requirement effectively15.

The DKAS has been previously translated and validated in many languages including Spanish, Chinese, Japanese, and Turkish16,17,18,19. Despite the large number of Arabic speakers worldwide, the DKAS has yet to be psychometrically evaluated in an Arabic-speaking population. Thus, an intervention is necessary to narrow the knowledge gap in the Arabic population in the MENA region20.

In the current literature, there are no studies assessing the level of dementia knowledge in the UAE. Hence, this study aims to evaluate the knowledge of dementia among the UAE population and to develop and assess the validity and reliability of the Arabic version of the DKAS (DKAS-A).

Methodology

Study design and population

In this study, we utilized a cross-sectional design to evaluate the knowledge of dementia among the general population of the UAE. We chose the convenience sampling method to recruit participants by distributing an online questionnaire on several social media platforms (WhatsApp, Instagram, Facebook, Reddit, and X) between December 2023 and January 2024. Concurrently, we collected data from UAE residents in public places (malls, parks, and bus stations) to increase the diversity of our sample. Eligibility criteria included being a resident of the UAE, aged 18 years or older, and understanding Arabic or English.

Data collection tool

The questionnaire contained a sociodemographics section to obtain the participants’ age, sex, education level, and employment status. This was followed by five questions assessing participants’ personal and caregiving experiences with people with dementia (PwD), their perceived knowledge, and sources of information on the disease. The final section comprised the 25 items of the DKAS with the survey taking approximately 10 min to complete. Participants were given the option to complete the questionnaire in either English (the original DKAS) or Arabic (the translated DKAS-A), based on their language preference.

The DKAS consists of 25 statements regarding dementia that require participants to identify whether each statement is “false”, “probably false”, “probably true”, or “true”, with an extra “I don’t know” option. Those statements belong to one of four subscales (causes and characteristics, communication and behavior, care considerations, and risk and health promotion) that add up to a final total score between 0 and 5021.

Each item was scored based on both accuracy and confidence. For true statements, participants received 2 points for a confident correct answer (“true”) and 1 point for a less confident but correct answer (“probably true”). Incorrect responses, including “false,” “probably false,” and “I don’t know,” were given 0 points. Items 1, 3, 4, 5, 6, 7, 9, 12, 13, 14, 15, 16, 19, and 20 were reverse coded, as they were false statements.

Sample size

We calculated the minimum required sample size to be 375 participants with a 5% margin of error, a 95% confidence level, and an estimated knowledge of 57.83% as reported by a similar study conducted in Saudi Arabia22. The sample size was calculated using OpenEpi online calculator. Prior to the initiation of data collection, we conducted a pilot study to collect feedback for the finalization of the questionnaire.

A total of 912 participants completed the questionnaire and were included in the analysis, of which 507 completed it in English (original DKAS) and 405 in Arabic (DKAS-A).

Translation process

The translation process involved the use of the Translation, Review, Adjudication, Pretest, and Documentation (TRAPD) translation method as described by Dorer (2020)23. This approach ensures a culturally appropriate high-quality translation due to its multiple quality checks. The questionnaire was independently translated by two expert bilingual translators of whom both were native Arabic and fluent English speakers with experience in the medical field. They followed the parallel translation technique to produce two separate Arabic versions. The two translators then met with a third individual, an independent reviewer and adjudicator, to finalize a single version which accurately reflects the original meaning.

Pretesting was done through individual debriefing assessments with 30 participants24. Each participant was asked to provide feedback on their interpretation of the meaning of each statement focusing on identifying any elements that may be misleading or confusing. Those 30 responses were not included in any further analysis. Adjustments were made by the reviewer based on this feedback to improve clarity and convey the intended meaning. Additionally, a third independent translator performed a back translation into English as an added quality check.

Reliability

To assess the instrument’s reliability, we evaluated its internal consistency by calculating Cronbach’s alpha and split-half reliability coefficients, as well as conducting test-retest analysis. We assessed split-half reliability by dividing the scale into 2 groups of even and odd-numbered items. As recommended by Kennedy (2022) and Dutil (2017), we assessed test-retest reliability on a sample of 100 individuals, who completed the questionnaire twice, with a two-week gap between administrations)25,26. We then calculated the intraclass correlation coefficient (ICC) using the two-way mixed effects model with absolute agreement to estimate the retest reliability27,28. Furthermore, we examined equivalent forms reliability by administering the instrument in Arabic and English to two separate groups, each consisting of 50 participants. Group A filled the English version followed by the Arabic version, while Group B filled them in the opposite order. We computed the Pearson product-moment correlation coefficients between scores to estimate equivalent forms reliability. We finally analyzed the data to check for the presence of floor and ceiling effects. It is worth noting that the responses included in the test-retest and equivalent forms reliability analyses were only used in these tests and not in any other sections of the study.

Validity

A bilingual consultant neurologist assessed the face validity of the final translation to ensure its linguistic accuracy and cultural appropriateness for the target population.

We assessed construct validity by using the known-groups method in which the total knowledge scores were compared between caregivers of PwD with non-caregivers and medical professionals with caregivers29. Our hypothesis was that people with a caregiving background would have a higher knowledge score than non-caregivers. Additionally, we expected similar knowledge levels between medical professionals and caregivers.

We further verified the tool’s construct validity by conducting a Confirmatory Factor Analysis (CFA) using the lavaan package in R (version 4.5.1). The CFA was based on the original four-factor model of the DKAS. Several indices were used to evaluate the model, and we reported the results according to the recommendations of Jackson et al. (2009) and Schreiber et al. (2006)30,31.

Recommendations for CFA suggest 10 participants for each item to establish reliable estimates32. The DKAS is a 25-item questionnaire; thus, a minimum of 250 participants is needed. We performed all validity analyses, including CFA, on the 405 respondents who filled the DKAS-A, exceeding the recommended number and ensuring the reliability of our findings.

Ethical considerations

Before beginning the questionnaire, all participants were presented with a participant information sheet detailing the purpose of the study and the voluntary nature of their participation. Informed consent was obtained from all participants prior to their involvement in the study. This study was approved by the Ethics Committee at the University of Sharjah prior to the survey’s distribution (REC-23-15-01-SE). All methods have been performed in accordance with the relevant guidelines and regulations.

Statistical analysis

We analyzed the characteristics of the participants, including their sociodemographic variables, frequency of interaction with PwD, and sources of dementia knowledge, using a descriptive analysis model. The results were reported in percentages and frequencies. The Kolmogorov–Smirnov test was statistically significant, indicating deviation from normality for the DKAS total and subscale scores (p < 0.05). However, due to the test’s oversensitivity in large samples, visual inspection of histograms and Q–Q plots, as well as skewness (–0.172) and kurtosis (0.425) values within the acceptable range, suggest that the data were approximately normally distributed. Therefore, we calculated these scores and reported their means and standard deviations33. For bivariate analysis, DKAS total and subscale scores were correlated with sociodemographic variables using Independent T-test and One-Way ANOVA. We conducted all analyses on SPSS Statistics for Windows Version 28.0 and R (version 4.5.1), considering p-values < 0.05 to be statistically significant.

Results

Sociodemographic characteristics of the sample

A total of 912 participants were included in the study, of which 507 answered in English and 405 answered in Arabic. Among them, 52.1% were female, and 28.6% were between 21 and 30 years of age. Non-Emiratis constituted 86.1% of the sample, and about one-third of the total population resided in Abu Dhabi. Approximately 44.5% had a high school degree or less, and about half of the sample was employed. The majority (90.2%) did not have any medical or healthcare background.

Only 35.1% encountered PwD, out of which 36.3% reported rarely interacting with these individuals. Most participants (82.1%) perceived their dementia knowledge to be between low and moderate levels (See Table 1).

Table 1 Sociodemographic characteristics of the sample.
Full size table

Reliability analyses

Using a sample of 100 individuals who completed the questionnaire twice, with a two-week gap between administrations, the intraclass correlation coefficient (ICC) was calculated to assess the stability of the results across the test and retest administrations. The DKAS-A achieved an ICC of 0.670, 95% CI [0.546, 0.765], indicating moderate reliability based on Koo & Li (2016)28.

When assessing equivalent forms reliability between two groups of 50 participants each, the DKAS-A showed very strong equivalence to the original English tool (r >0.8) as reported by Akoglu (2018)34. Pearson’s correlation coefficient between total scores was 0.875 for Group A (English followed by Arabic), with results ranging from 0.638 (communications and behavior) to 0.822 (causes and characteristics) across the subscales. For Group B (Arabic followed by English), results ranged from 0.660 (⁠risks and health promotion) to 0.800 (causes and characteristics), with a 0.827 correlation between total scores. All correlations were statistically significant (p < 0.001).

Further reliability testing was done on the 405 Arabic responses within our sample. The DKAS-A demonstrated acceptable internal consistency, with a Cronbach’s alpha of 0.70 and a McDonald’s omega of 0.77. Upon split-half reliability assessment, the Spearman-Brown coefficient was 0.808, exceeding the 0.8 threshold to indicate a very strong correlation based on Akoglu (2018)34. Furthermore, the Guttman Split-Half coefficient was 0.807.

Assessment of floor and ceiling effects showed minimal limitations. Only 3 respondents got the minimum score, and no participant achieved the maximum score of 50. This absence of score aggregation at either end of the range suggests good discrimination power across the entire scale.

Validity analyses

Construct validity

In the first comparison for the known-groups method, the T-Test was able to show a statistically significant difference in scores (t = −2.978, p = 0.003) between caregivers of PwD (mean = 21.73, SD = 5.79) and non-caregivers (mean = 19.01, SD = 6.44). In the second comparison, there was no significant difference in scores (t = −0.170, p = 0.865) between medical professionals (mean = 21.31, SD = 6.85) and caregivers of PwD (mean = 21.55, SD = 5.66). All 405 Arabic responses were used for validity analyses.

We initially conducted the CFA on all 25 items of the DKAS-A which showed insufficient model fit. The model had a significant chi-square value, χ² (269, N = 405) = 783.40, p < 0.001. This significance is expected considering the large sample size and hence should not be considered for model fitting35,36. While the root mean square error of approximation (RMSEA) (0.069; 90% CI [0.063, 0.074]) and standard root mean square residual (SRMR) (0.063) were within acceptable thresholds (< 0.08), both the comparative fit index (CFI) (0.855) and the Tucker-Lewis Index (TLI) (0.838) indicated limited support for the model’s structure31.

Examination of standardized factor loadings and squared multiple correlations revealed that items 1 and 20 had very low factor loadings (< 0.4) and squared multiple correlations (< 0.2), whereas items 5 and 14 showed borderline factor loadings (0.415 and 0.403, respectively) and very low squared multiple correlations (0.172, and 0.162, respectively)37. Further assessment showed that items 5 and 20 had high standardized residual covariances (>2) with multiple items, and items 1, 5, 14, and 20 demonstrated the lowest mean inter-item correlations (all ≤ 0.258). Therefore, these four items were excluded from the model.

We then conducted a revised CFA on the remaining 21 items. Model fit improved across all indices: χ² (183, N = 405) = 452.546, p < 0.001, RMSEA = 0.060 (90% CI [0.053, 0.067]), SRMR = 0.049, CFI = 0.913, and TLI = 0.900. While the increase of CFI and TLI exceeded the leniently accepted threshold of 0.9, they did not reach the strict threshold of 0.9531,38. Generally, the overall fit was improved, supporting the construct validity of the 21-item version (DKAS21-A). The accepted CFA model is presented in Fig. 1.

Fig. 1
figure 1

Confirmatory factor analysis of the DKAS21-A (without items 1, 5, 14, 20).

Full size image

Subsequently, we tested the psychometric properties of the DKAS21-A to compare it to the 25-item version. The full results are provided in Table 2. The overall quality of the questionnaire did not significantly change. Test-retest, equivalent forms, and split-half reliability showed comparable values between the two versions. A slight decrease was seen in Cronbach’s alpha that can be explained by the reduction in the number of items rather than an alteration of the tool’s reliability39. This is further supported by the persistence of McDonald’s omega coefficient at the acceptable value of 0.77. For construct validity, the T-Test comparing caregivers of dementia patients and non-caregivers remained significant, while the comparison between medical professionals and caregivers was non-significant.

Table 2 Psychometric properties of the DKAS-A compared to DKAS21-A.
Full size table

Despite the improved construct validity of the DKAS21-A in comparison to the DKAS-A, we opted to maintain the results of the 25-item version due to the lack of significant changes to the psychometric properties after the reduction of items and to support merging the Arabic and English responses, allowing us to provide a more comprehensive understanding of the UAE’s population as a whole. See supplementary files for the full version of both DKAS-A and DKAS21-A.

Resources of dementia knowledge

Media (TV, radio, Internet) was the most commonly reported source of knowledge, with nearly 64.0% of participants using it. Approximately 29.4% and 15.7% reported experience and education, as their primary knowledge sources on dementia, respectively. Around one-tenth of the participants reported obtaining information on the disease from their physicians, while support groups and other sources of knowledge were minimally utilized, with each being used by 3.6% of participants (See Fig. 2).

Fig. 2
figure 2

Sources of knowledge about dementia.

Full size image

Knowledge of dementia

The DKAS sample scores ranged between 0 (minimum) and 43 (maximum) (mean = 19.19; SD = 7.13), corresponding to 38.38% correct responses. There is no set cutoff score to represent a certain level of dementia knowledge; however, scores of more than 25 (out of 50) have been previously considered in the literature as indicative of moderate knowledge40. Using that cutoff, only 22.8% (n = 208) of the sample have adequate knowledge of dementia.

At the subscales level, the mean scores in “Care consideration” (mean = 5.74, SD = 3.21) and “Causes and characteristics” subscales (mean = 5.5, SD = 2.7) were higher than “Risks and health promotions” (mean = 4.23, SD = 2.23) and “Communication and behaviors” (mean = 3.71, SD = 2.11) (See Fig. 3).

Fig. 3
figure 3

Dementia knowledge assessment scale (DKAS) subscales scores.

Full size image

At the individual statements level, correct and partially correct responses were merged and presented in Fig. 4. A few statements had a relatively lower rate of correct answers including those related to the belief that the sudden onset of cognitive problems is a characteristic feature of dementia (24.3% answered correctly), the role of medications in treatment (34.2%), the prevalence of vascular dementia (35.1%), the effect of dementia on lifespan (37.4%), and the importance of correcting confused patients (39.6%). On the other hand, the statements with the highest scores included the ones about the prevalence of Alzheimer’s disease (78.8%), the benefit of exercise (73.4%), the effectiveness of daily care (72.8%), and the difficulty of learning new skills (72.0%). The percentage of participants who answered each DKAS item correctly or partially correctly is shown in Fig. 3.

Fig. 4
figure 4

Percentage of the participants’ correct answers in DKAS.

Full size image

Sociodemographic factors and DKAS

Upon association of demographic characteristics & DKAS scores using the Independent T-test and One-Way ANOVA test, working or studying in the healthcare sector demonstrated a significant correlation with both total and subscale DKAS scores (p < 0.05). Female sex was significantly associated with higher DKAS total score and all of its domains except for risk and health promotion subscale. The “Causes and characteristics” subscale score demonstrated a significant association with age (p = 0.045). Participants with higher education (e.g. Master’s or Doctoral degrees) demonstrated better awareness on the “risk and health promotion” subscale and had higher overall DKAS knowledge scores than those with a high school diploma or less.

Participants’ subjective knowledge of dementia showed a significant relationship across the total score and all DKAS subscales, except for the “Communication and behavior” and “Risks and health promotion” subscales. In these two subscales, participants with high or moderate subjective knowledge showed better awareness than those with low subjective knowledge. However, no significant difference was noted between the high and moderate groups.

Regarding sources of knowledge, a significant correlation was found between total knowledge scores with direct experience with PwD (p < 0.001) and having taken educational courses (p < 0.001). However, no significant correlation was found between total knowledge scores and exposure to media, support groups, information from physicians, or other sources (See Table 3).

Table 3 Demographic characteristics & DKAS.
Full size table

Discussion

In this study, the DKAS was successfully translated from English to Arabic. The Arabic Dementia Knowledge Assessment Scale (DKAS-A) and its shortened 21-item version (DKAS21-A) showed satisfactory validity and reliability. To our knowledge, the DKAS-A is the first measure of dementia knowledge that has been psychometrically validated in the Arabic language, and the only measure that was tested on a diverse cohort from the general adult population of the UAE. The Arabic version of Alzheimer’s Disease Knowledge Scale (ADKS-A) has been developed and psychometrically tested in Egypt; however, when comparing the ADKS and DKAS, the latter has proven to be a reliable, valid measure of dementia knowledge that performs better than the ADKS when administered to a large and diverse international cohort41,42.

Melchior, F. & Teichmann, B. (2023) previously suggested that the DKAS has excessive items that are often deleted in translation papers. This hypothesis is further supported by our findings where the CFA revealed that exclusion of items 1, 5, 14, and 20 improves the Arabic version’s construct validity without significantly altering its psychometric properties. This is consistent with the results from the German translation where deletion of 5 items allowed the CFI and TLI indices to cross the threshold of 0.9 while maintaining the tool’s reliability43. While the Spanish translation retained the original 25-item, four-factor model of the tool, the Turkish and Japanese versions removed eight and nine items, respectively, to improve psychometric performance. These translations also opted for a unidimensional structure instead of the four-factor model16,17,18.

Notably, the items deleted were not consistent across translation studies, suggesting that the reasons for exclusion likely stem from an absence of culturally or linguistically equivalent wording or concepts in the translated tools’ languages44. As highlighted by Sung et al. (2021) and Melchior, F. & Teichmann, B. (2023), another factor that contributes to the differences in model fitting between DKAS translations is the sample used in each paper. To illustrate, the Turkish and traditional Chinese translations were validated in samples that included a large number of nursing professionals and home care workers, respectively. These versions demonstrated better fitting compared to validations conducted in general population samples, including our study and the German validation16,19,43.

Given that the DKAS21-A demonstrated comparable psychometric properties to the DKAS-A but showed better construct validity based on CFA results, we recommend that future studies assessing dementia knowledge in Arabic-speaking populations use the DKAS21-A when collecting data exclusively in Arabic. However, the absence of significant changes to the tool’s reliability may justify using the 25-item version for the purpose of merging responses if data is also collected in English. Thus, our results support the use of the DKAS-A and DKAS21-A as a generalized measure of dementia knowledge, applicable to diverse populations. This adaptation provides a crucial tool for enhancing the understanding of dementia within Arabic-speaking communities.

Upon investigating dementia knowledge levels, we found that only 22.8% of our sample had moderate knowledge (mean = 19.19, SD = 7.13), corresponding to 38.38% correct responses. Comparatively, our population’s scores were consistent with a study conducted among Italian and Swiss populations (mean = 22.4; SD = 8.88), yielding 44.86% correct responses45. In contrast, our findings were lower than those reported in Germany (mean = 28.80; SD = 9.46) and Greece (mean = 25.87; SD = 8.6). This discrepancy may be attributed to the fact that 75% of the German sample and 82.5% of the Greek sample know a PwD, in comparison to our sample where only 35.1% know a PwD43,46.

Regarding key demographic variables, females demonstrated better knowledge than males, aligning with previous studies45,47. Several factors may contribute to this observation. Primarily, in both familial and professional settings, women are more likely to assume caregiving responsibilities than men48,49. Moreover, because women are at a higher risk of developing dementia, they tend to be more proactive in seeking health information50. In terms of age, there were no clear trends found with total knowledge score in our study. Other studies have also reported a similar absence of a relationship between age and knowledge scores, while some have found a positive correlation between these variables40,51,52,53.

Among the four DKAS subscales, high scores were reported in the “Care Considerations” subscale. This is consistent with other study populations in the MENA region that share similar cultural and religious values to those of the UAE54,55. Traditionally, family-centered care for the elderly is prioritized in the region as opposed to nursing home care. This may contribute to more supportive attitudes towards caring for PwD56,57.

In contrast, respondents scored lower on the Communication and Behavior subscale, with nearly half stating that it is impossible to communicate with someone who has advanced dementia (DKAS 14). Given the importance of communication in improving the quality of life and managing behavioral problems in this patient population, this discovery is concerning58. Limiting social isolation in patients and enhancing community engagement is imperative to reducing the disease burden. As a result, greater priority should be placed on imparting effective communication strategies, such as Montessori-based programs, that are designed to address the increasing linguistic and cognitive deterioration associated with dementia59,60.

Our study revealed two key misconceptions about dementia. The first involves the commonly held belief that dementia is a natural part of aging61. Notably, nearly half (48%) of our participants incorrectly agreed with this statement. This finding is concerning especially when compared to another study where almost one-third (36.7%) of their sample had the same misbelief45. This misconception can lead to serious consequences in which people may ignore early warning signs and delay seeking diagnosis and treatment.

The second misconception is evident in that 56.6% of respondents were not able to recognize high blood pressure as a risk factor for developing dementia. This finding highlights a significant knowledge gap, particularly in the UAE where approximately one-third of the population is hypertensive and does not adequately manage this modifiable dementia risk factor62. Thus, by targeting individuals at risk of hypertension and educating those already diagnosed, we can improve health outcomes and reduce the incidence of dementia in the nation.

Participants’ knowledge of dementia varied based on their sources of information. A notable positive correlation was observed among individuals who reported direct exposure to PwD or participation in educational courses. This underscores the effectiveness of firsthand exposure and structured educational interventions in promoting a deeper understanding of dementia17,63. Hence, implementing social recreational programs where PwD are paired with volunteers may prove to be beneficial. This approach can increase dementia knowledge in volunteers across all age groups while also reducing social isolation in patients64.

On the other hand, no significant correlation was found between total knowledge scores and sources such as media, physicians, or support groups. Almost two-thirds (64%) of participants reported using media as their primary source of knowledge while only 10.4% acquired their information from physicians. The latter low percentage may explain the lack of correlation between this knowledge source and total scores. Furthermore, the variability in depth and quality of information provided by physicians may contribute to this finding. Nonetheless, existing literature suggests that competent physicians can improve patient’s understanding of their conditions and adherence to treatment65. Similarly, due to its widespread use, the appropriate utilization of media holds great potential for enhancing public awareness. One dual approach involves the development of online health communities that individuals can integrate into their healthcare routines to easily obtain reliable information from medical professionals66.

The low knowledge scores presented in this study may be attributed to the absence of a national dementia plan in the UAE. In comparison, the UAE population has demonstrated good knowledge levels on conditions that have been extensively targeted by awareness campaigns in the country, such as breast67 cancer. From this evident success, similar efforts should be done to help improve knowledge on dementia. A crucial step in this direction is through the implementation of a stand-alone dementia care strategy. In the meantime, dementia may be included as a priority in the country’s existing National Policy for Mental Health Promotion68. This approach has been implemented in countries like Australia and Switzerland, allowing for a prompt and multi-sectoral response to address dementia needs and raise public awareness69,70,71.

In our study, physicians were not a significant source of knowledge for dementia. Therefore, it is essential to guide them with the proper methods to effectively educate the population about the disease. This approach is seen within the Saudi Protocol for the Prevention of Cognitive Impairment that calls not only for education but also for the reduction of modifiable dementia risk factors tailored to the region72. These recommendations are highly relevant to the UAE population, especially due to the high prevalence of dementia risk factors such as obesity, diabetes, and hypertension62,73,74. Educating the public about these risk factors will encourage health prevention methods and delay cognitive decline.

Strengths and limitations

This study is not without its limitations. The cross-sectional design of the study highlights correlations between knowledge and sample demographics but does not establish causality. The use of convenience sampling restricted the sample to individuals within our reach, which may limit the generalizability of the findings. While the sample generally represented the UAE’s sociodemographic diversity, participants aged 18–20 were fairly overrepresented. This skew is likely due to the prevalent use of social media among this demographic. Additionally, the use of a self-reported questionnaire may lead to response biases.

The concurrent validity of the DKAS-A was not assessed in this study. Future research could address this by administering the DKAS-A alongside the ADKS-A to the same sample. Additionally, the statistical power of the known-groups analyses may be limited since these were conducted without prior determination of subgroup sample sizes, resulting in a relatively small sample size in the caregivers and medical professionals subgroup. However, the absence of sample size justification is common in similar validation studies, as noted in previous literature75.

The utilization of the DKAS over the ADKS in this study was due to the DKAS having a more-favorable response profile, better internal consistency, better construct validity, and overall performed better than ADKS42. Additionally, the translation of DKAS facilitated broader participation from Arabic-speaking individuals in the UA E. This resulted in a large sample of 912 participants which allowed for a high power of statistical analysis. The design of the online questionnaire minimized missing data, as all questions required responses. This study marks the first attempt in the UAE to measure dementia knowledge levels within the population.

Conclusion

This study successfully translated and validated a reliable Arabic version of the DKAS and evaluated the level of dementia knowledge among the UAE public. Our results confirm that knowledge on dementia is insufficient among the country’s adult population. This knowledge gap poses a significant public health challenge, particularly as the country’s population ages and the incidence of dementia is expected to rise exponentially. Our findings highlight the necessity for comprehensive dementia education programs and the development and implementation of a National Dementia Plan, in line with WHO recommendations5,76. Future research should focus on assessing the effectiveness of educational interventions and investigating the barriers to dementia knowledge among various demographic groups in the UAE, to help tailor programs to specific needs.