Community-Based Dementia Screening Initiative in Brunei: Final Results

Abstract

The objective was to expand community-based dementia screening across diverse settings in Brunei and evaluate the prevalence of risk factors, cognitive symptoms and factors associated with suspected cognitive impairment using the Mini-COG. Cross-sectional community screening was conducted across community and healthcare settings among adults aged 60 years and older, or aged 50 years and older with dementia risk factors. Participants completed a structured questionnaire on demographics, dementia risk factors and cognitive symptoms, followed by the Mini-COG. Multivariable logistic regression was used to examine factors associated with suspected cognitive impairment. A total of 1358 participants were included, with a median age of 60 years, and 63.5% were female. Two-thirds had hypertension, over half had hypercholesterolaemia and over one-third had diabetes mellitus. The most commonly reported symptoms were misplacing things (42.6%), forgetfulness (32.5%) and visuospatial difficulties (24.2%). Among those with symptoms, 14.5% reported worsening symptoms and 12.0% had impaired activities of daily living. From the Mini-COG, 14.4% had suspected cognitive impairment. Multivariable analysis revealed that older age (adjusted OR 1.06 per year, p < 0.001) and education level (adjusted OR 2.25 for primary versus tertiary, p = 0.007) were independently associated with Mini-COG impairment, while vascular risk factors showed no significant associations. Community-based dementia screening using a brief questionnaire and the Mini-COG was feasible across multiple settings in Brunei. There was a high prevalence of vascular risk factors and cognitive symptoms. Age and educational attainment were stronger predictors of abnormal Mini-COG results compared to vascular risk factors in this study. These findings support opportunistic screening in healthcare and community settings, and an emphasis on public education on dementia symptoms and risk reduction.

1. Introduction

Dementia represents a growing public health challenge globally, with the number of people living with dementia projected to triple from 57 million in 2019 to 153 million by 2050 [1]. This rise is more marked in low- and middle-income countries, particularly in Asia and the Western Pacific [2]. Brunei Darussalam is a small nation in Southeast Asia with a population of approximately 442,000 [3]. Based on the Global Burden of Disease Study 2019, there were an estimated 1574 people living with dementia in Brunei in 2019, which is projected to increase nearly five-fold to 7317 by 2050 [4]. This reflects the rapidly ageing population and high prevalence of vascular risk factors including hypertension, diabetes mellitus and cardiovascular disease in the country [3].

Early identification of cognitive impairment is a priority of the World Health Organization (WHO) Global action plan on the public health response to dementia 2017–2025, which aims for at least 50% of people with dementia to be diagnosed by at least 50% of member states by 2025 [4]. However, dementia was significantly underdiagnosed in Brunei, with only 18.5% of estimated cases identified through national electronic health records in 2022 [5]. Furthermore, many individuals with dementia present late to clinical services, often with complications such as falls, immobility, or aspiration pneumonia, rather than with early cognitive symptoms [6].

This diagnostic gap reflects multiple factors, including limited public awareness of dementia symptoms, cultural attitudes towards cognitive decline as a normal part of ageing and a lack of routine cognitive screening in primary care and community settings [7]. During the COVID-19 pandemic, efforts were made to disseminate dementia information through social media platforms [8]. The post-pandemic period presents an opportunity to strengthen community engagement and early detection strategies, particularly given the advances in disease-modifying treatments for Alzheimer’s disease [9] and increasing evidence for dementia risk reduction through management of modifiable risk factors [10].

Brief cognitive screening tools, such as the Mini-COG, offer a practical approach to identifying possible cognitive impairment in community and clinical settings. The Mini-COG combines three-word recall with a clock-drawing test, which can be administered in approximately three minutes with minimal training [11]. While cognitive assessment tools have not been formally validated in Brunei [12], the simplicity and cross-cultural applicability of the Mini-COG make it suitable for opportunistic screening across diverse settings and populations.

While many dementia risk factors are well established in the literature, there is limited evidence describing their distribution in Southeast Asian populations or examining how community-based cognitive screening can be implemented in real-world settings. The pilot phase of community-based dementia screening conducted in senior citizen activity centres in Brunei was previously reported [13]. This pilot demonstrated the feasibility and acceptability of screening, and identified a high prevalence of dementia risk factors and cognitive symptoms among these socially active older adults. Brunei represents a distinct sociocultural and healthcare context, with strong primary care coverage, a high non-communicable disease burden and an evolving public awareness of dementia.

Following this pilot, the screening initiative was expanded to a broader range of community and healthcare settings to reach a larger group. This paper reports the final results from the expanded community dementia screening initiative. The aims were to describe the demographic characteristics and distribution of dementia risk factors and cognitive symptoms across the full cohort, examine associations between participant characteristics and suspected cognitive impairment using multivariable analysis and provide implementation-level evidence to inform community-based dementia detection strategies in Brunei.

2. Materials and Methods

The cross-sectional community-based screening initiative was conducted between June 2022 and May 2023. The study design and screening approach have been described in detail previously [13]. Participants were recruited through consecutive sampling at designated screening venues during organised sessions. They completed a structured questionnaire covering basic demographics, risk factors for dementia, ten warning signs or symptoms of dementia [14] and the Mini-COG as a brief cognitive screening tool. The questionnaire was developed by the research team. Content validity was ensured through expert review by clinicians involved in geriatric medicine and primary care. As the questionnaire was designed for pragmatic community screening, formal psychometric validation was not performed. Screening was conducted face-to-face by trained volunteers under the supervision of the research team. For participants who had difficulty reading or writing, all questions were asked verbally to ensure comprehension and accurate recording of responses.

The Mini-COG was administered according to standard published procedures [11]. Participants were asked to recall three unrelated words, draw a clock with numbers correctly placed and set the hands to a specified time (10 past 11). Scoring followed established criteria: participants recalling all three words were classified as unlikely to have dementia; those recalling none were classified as likely dementia; and those recalling one or two words were further evaluated based on their clock-drawing performance. All screeners received structured training prior to data collection to ensure consistency in administration and scoring. The research team was also available for clarification of uncertain scoring during screening sessions.

The inclusion criteria were older adults aged 60 years and above, or those aged at least 50 years old who reported one or more risk factors for dementia. The inclusion of individuals aged 50–59 years was intended to facilitate early identification of at-risk individuals rather than to estimate the population prevalence of late-onset dementia. The locations for screening included public spaces (churches, mosques, libraries, community centres and government ministries), places where older people frequent (Senior Citizen Activity Centres, pension collection locations and marketplaces) and places where people with risk factors attend for evaluation (primary care, dialysis centres, hospital outpatient clinics and hospital lobbies). After completing the questionnaire, participants were given brief tailored advice on dementia risk reduction and were encouraged to discuss with their doctor if they had poorly controlled risk factors, symptoms or cognitive concerns identified on the Mini-COG.

Data collected were entered into Microsoft Excel and analysed using RStudio version 4.5.2. Descriptive statistics were used to summarise participant characteristics, while multivariable logistic regression examined adjusted associations between demographics, education and multimorbidity with suspected cognitive impairment based on the Mini-COG. The significance threshold was p < 0.05.

3. Results

Screening was undertaken in a range of community and healthcare settings.

Table 1. Location of dementia screening categorised by participant groups: 50–59 years with dementia risk factors, 60 years and older or excluded from subsequent analysis.

Screening Locations	50–59 Years	≥60 Years	Excluded	Total
Senior Citizen Activity Centres	22	180	12	214
Community and retirement group talks	48	80	100	228
Public awareness talks	28	27	82	137
Public awareness booths	5	21	15	41
Marketplaces	32	78	50	160
Primary care health centres	28	14	76	118
Dialysis centres	76	130	105	311
Hospital outpatient clinics	53	99	14	166
Government ministries	337	1	210	548
Total	635	723	604	1962

shows the number of participants recruited from each category.

There were 1962 participants screened, out of which 1358 (69.2%) were included for analysis. The median age of included participants was 60 years (range 50 to 91 years). There were 490 (36.1%) males, 862 (63.5%) females and 6 (0.4%) participants who did not disclose gender. The majority, 988 (72.8%) participants had at least a primary school education, with 584 (43.0%) having secondary education and 269 (19.8%) having tertiary education. In terms of occupation, participants included retirees, government employees, private sector workers and homemakers, which reflected the diverse recruitment settings.

The prevalence of self-reported dementia risk factors was high across the cohort (Figure 1). Hypertension was the most common risk factor in 892 (65.7%), followed by high cholesterol in 723 (53.2%), diabetes mellitus in 484 (35.6%) and overweight in 386 (28.4%). The main differences between male and female participants (shown in

Table 2. Participant characteristics according to gender.

Demographics:
Variable	Male (N = 490)	Female (N = 862)	p-Value
Age, Median (range)	63 (50–89)	60 (50–83)	0.78
Education Level:			0.39
Primary	37 (7.5%)	68 (7.9%)
Secondary	202 (41.2%)	382 (44.3%)
Tertiary	106 (21.6%)	163 (18.9%)
Vascular Risk Factors:
Variable	Male	Female	p-Value
High blood pressure	318 (64.8%)	570 (66.1%)	0.65
High cholesterol	252 (51.4%)	470 (54.5%)	0.27
Diabetes mellitus	173 (35.3%)	309 (35.8%)	0.84
Heart disease	100 (20.4%)	93 (10.8%)	<0.001
Lung disease	39 (8.0%)	52 (6.0%)	0.17
Liver disease	18 (3.7%)	38 (4.4%)	0.51
Kidney disease	102 (20.8%)	133 (15.4%)	0.01
Previous head injuries	24 (4.9%)	32 (3.7%)	0.29
Previous stroke	19 (3.9%)	22 (2.6%)	0.17
Overweight	121 (24.7%)	266 (30.9%)	0.02
Smoking	64 (13.1%)	11 (1.3%)	<0.001
Cognitive Symptoms:
Variable	Male	Female	p-Value
Memory loss/forgetfulness	162 (33.1%)	279 (32.3%)	0.79
Difficulty performing familiar tasks	54 (11.0%)	94 (10.9%)	0.94
Language/word finding difficulty	62 (12.7%)	93 (10.8%)	0.30
Disoriented to time/place	74 (15.1%)	125 (14.5%)	0.76
Difficulty making judgements	62 (12.7%)	101 (11.7%)	0.30
Problem keeping track of things	51 (10.4%)	104 (12.1%)	0.36
Misplacing things	203 (41.4%)	374 (43.4%)	0.48
Changes in mood/behaviour	92 (18.8%)	190 (22.0%)	0.16
Visuospatial difficulties	121 (24.7%)	206 (23.9%)	0.74
Withdrawing from work or social activities	37 (7.6%)	65 (7.5%)	0.99
Progression, Functional Impact and Need for Evaluation:
Variable	Male	Female	p-Value
Are symptoms worsening?	40 (8.2%)	64 (7.4%)	0.62
Are they affecting ADLs?	35 (7.1%)	55 (6.4%)	0.59
Should your memory get checked?	83 (16.9%)	152 (17.6%)	0.75
Mini-COG (suspected cognitive impairment)	78 (15.9%)	118 (13.7%)	0.26

Only ‘Yes’ responses are presented for dichotomous variables. Other responses are excluded from this table but accounted for in analyses.

) were that males had significantly higher rates of heart disease (20.4% vs. 10.8%, p < 0.001), kidney disease (20.8% vs. 15.4%, p = 0.01) and smoking (13.1% vs. 1.3%, p < 0.001), while females had higher rates of being overweight (30.9% vs. 24.7%, p = 0.02).

The distribution of self-reported cognitive symptoms is shown in Figure 2. The most commonly reported symptom was misplacing things in 578 (42.6%) of participants, followed by forgetfulness in 441 (32.5%), visuospatial difficulties in 328 (24.2%) and mood or behaviour changes in 283 (20.8%). Among participants who reported cognitive symptoms, 104 (14.5%) stated that they were getting worse, while 90 (12.0%) admitted to symptoms affecting their activities of daily living (ADLs). There were 235 (17.3%) who thought they should get their memory checked. There were no significant differences between genders for cognitive symptoms (Table 2). While females showed a trend towards higher reporting of changes in mood or behaviour, this did not reach statistical significance. There were 197 (14.5%) with suspected cognitive impairment based on the Mini-COG.

Multivariable logistic regression analysis adjusting for age, gender, education level and vascular risk factors found that increasing age and lower education level were independently associated with suspected cognitive impairment on the Mini-COG. Increasing age was associated with a higher likelihood of Mini-COG impairment (adjusted OR 1.06 per year, p < 0.001). Education level had a strong graded association, where participants with only primary education had more than twice the odds of Mini-COG impairment compared to tertiary education (adjusted OR 2.25, p = 0.007). Gender was not significantly associated with Mini-COG impairment (adjusted OR for males 1.21, p = 0.26). None of the vascular risk factors showed statistically significant associations with Mini-COG impairment after adjustment: hypertension (adjusted OR 1.15, p = 0.44), hypercholesterolaemia (adjusted OR 1.08, p = 0.64), diabetes mellitus (adjusted OR 0.91, p = 0.54), previous stroke (adjusted OR 1.43, p = 0.38), kidney disease (adjusted OR 1.12, p = 0.55) and smoking (adjusted OR 0.78, p = 0.42).

4. Discussion

This study reports the final results of a large-scale community dementia screening initiative in Brunei, expanding upon the previously reported pilot phase [13]. Senior citizen activity centres and community venues provided access to socially engaged older adults who may not regularly attend healthcare services. Workplace screening at government ministries captured a younger high-risk population who would not typically present for cognitive concerns. Clinical settings, such as dialysis centres and outpatient clinics enabled screening of individuals with established vascular comorbidities during routine medical visits. This diverse range of recruitment settings demonstrates that opportunistic cognitive screening can be embedded within public health activities, medical follow-ups and non-healthcare environments.

Similar community-based screening initiatives reported comparable feasibility and acceptability. A primary-care-based screening programme in India found that over 40% of screened individuals had possible dementia, and detection of vascular risk factors facilitated integrated the management of both cognitive and non-communicable disease risk [15]. Similar case-finding programmes in Singapore and other Southeast Asian settings have likewise demonstrated that brief cognitive tools can be utilised outside memory services [16,17]. Our findings add to this evidence base by showing the screening can be delivered across a wider range of venues, from religious and social gathering places to workplace health promotion events.

Findings from this study reveal a high prevalence of vascular risk factors and cognitive symptoms among participants. This was consistent with the pilot study findings and national non-communicable disease prevalence data from Brunei [3,13], and underscores the importance of considering dementia risk reduction messaging in existing vascular disease management programmes. While there were a significant number of participants with cognitive symptoms, only 17.3% felt they should have their memory checked. This suggests that many individuals may not recognize the significance of these symptoms or may be reluctant to seek formal assessment. This supports the need for proactive screening to identify individuals who may benefit from further evaluation but would not otherwise present to health services.

Age and education level were the strongest predictors of suspected cognitive impairment on the Mini-COG, while vascular risk factors showed no independent associations in this study. Each additional year of age was associated with 6% higher odds of impairment on the Mini-COG, consistent with the known relationship between advancing age and cognitive decline. The association with education level indicates the protective effect of higher educational attainment, through improvement in cognitive reserve [18,19]. The lack of association of vascular risk factors was unexpected; longitudinal studies consistently demonstrate that midlife hypertension, diabetes and other vascular risk factors predict later dementia risk, even when cross-sectional associations may be weak [20,21]. This may be partly attributable to the cross-sectional nature of this study and the reliance on self-reported comorbidities. Undiagnosed or unrecognized hypertension, diabetes and hyperlipidaemia are common in community populations, particularly among individuals who do not undergo regular health screening. Such non-differential misclassification may also attenuate observed associations between vascular risk factors and cognitive outcomes.

In addition, the Mini-COG is a brief screening instrument that primarily assesses memory recall and basic visuospatial function. It does not comprehensively evaluate executive function, processing speed, or attention, which are domains commonly affected in vascular cognitive impairment. Thus, the absence of independent associations in this study may also partly reflect the measurement limitations of the screening tool, rather than a true lack of a relationship between vascular risk burden and cognitive function.

While universal population screening for dementia in low-prevalence community settings may have limited benefit, a case-finding approach targeting individuals with risk factors, symptoms or specific healthcare encounters is more appropriate and cost-effective [22]. Our experience supports a pragmatic model, where cognitive screening is embedded within existing health contacts and community platforms. Opportunistic screening opportunities include non-communicable disease clinics, workplace health screening programmes and community events. In each of these settings, screening can be combined with brief education on brain-health lifestyles and dementia risk reduction [23]. The screening must be coupled with referral pathways for confirmatory assessment. The screening questionnaire was designed to be easily adapted into a self-referral form. Participants could take their questionnaire to their healthcare provider, which summarised their risk factors, symptoms and Mini-COG score as a basis for clinical discussion. In Brunei, this involves primary care physicians for initial evaluation and risk factor management, with onward referrals to specialist geriatric or neurology services for comprehensive cognitive assessment and diagnosis if indicated.

Strengths of this study include the large sample size, diverse recruitment settings and systematic collection of data on demographics, risk factors, symptoms and objective cognitive screening. Limitations include the cross-sectional design precluding conclusions about causality or prognostic significance of screening findings. Risk factors and symptoms were self-reported for practicality, which may lead to recall bias or underreporting. The Mini-COG is a brief screening tool and may not capture subtle or domain-specific cognitive deficits, particularly executive function that is commonly affected in vascular cognitive impairment. The convenience sampling approach may limit generalisability, as individuals who are home-bound, institutionalised, socially isolated, or less likely to attend organised community events or healthcare venues may have been underrepresented.

Areas for future research include longitudinal follow-up of screened participants to determine the prognostic value of the Mini-COG and subjective symptoms for subsequent dementia diagnosis. Comparative studies across Southeast Asian countries could explore how education systems and health system structures influence screening outcomes and pathways to care. Finally, research on culturally appropriate post-screening interventions, such as cognitive stimulation, lifestyle modification programmes and caregiver support, may complement early detection efforts.

5. Conclusions

Community-based dementia screening using a brief questionnaire and Mini-COG was feasible across multiple settings in Brunei. There was a high prevalence of vascular risk factors and cognitive symptoms. Age and educational attainment were stronger predictors of abnormal Mini-COG results compared to vascular risk factors in this study. These findings support opportunistic screening in healthcare and community settings, and emphasise the need for public education on dementia symptoms and risk reduction.