Prevalence and risk factors for Alzheimer’s disease and related dementias in Ghana: evidence from a cross-sectional population-based study

Abstract

Alzheimer’s disease and Related Dementia (ADRD) is a growing concern for low- and middle-income countries. Yet, studies on the prevalence and risk factors for dementia in sub-Saharan Africa are limited. This study estimated the prevalence and identified the risk factors for ADRD in Ghana. A cross-sectional design involving 384 participants aged 60 years or more completed the Brief Community Screening Instrument for Dementia (CSID) with an additional set of questions on risk factors. The prevalence of dementia was 16% (62/384). Age (AOR = 1.031 (95% CI 1.002-1.061, p = 0.035), education (AOR = 0.689 (95% CI 0.517-0.919, p = 0.011), and employment (AOR = 0.320 (95% CI 0.149-0.685, p = 0.004) were associated with dementia. Also, ‘experience of stressful life event from age 16-64’ (AOR = 1.325 (95% CI 1.034-1.698, p = 0.024), ‘experience of stressful life event from age 65+’ (AOR = 1.258 (95% CI 1.024-1.546, p = 0.042), and ‘activities of daily living’ (AOR = 0.925 (95% CI 0.868-0.986, p = 0.037) identified as risk factors of dementia. The burden of ADRD was high. Urgent actions are needed to address this problem. There is a possibility that the burden of dementia may have been overestimated because of the study instrument. Findings should be interpreted in light of this limitation.

Keywords:

• Lifestyle
• sub-Saharan Africa
• brief community screening instrument for dementia
• stressful live events
• activities of daily living
• cognitive impairment

Reviewing Editor:

• Joe Thomas, Institute of Health and Management, Australia

Subjects:

• Health & Society
• Health Conditions
• Public Health Policy and Practice

Introduction

Alzheimer’s disease and Related Dementia (ADRD) is a growing concern for low- and middle-income countries (LMICs). With the population aging, the number of people living with dementia is expected to grow rapidly in LMICs, particularly in sub-Saharan Africa (SSA). Projections indicate that by 2050, approximately 161 million persons who are 60 years or older will be residents of SSA (Guerchet et al., 2017). This is likely to impose an increasing burden on health systems throughout SSA, as the country’s population ages, accompanied by rapid increases in the prevalence of vascular and other non-communicable disease risk factors for dementia (Olayinka & Mbuyi, 2014). The prevalence and incidence of dementia increase with age. Yet, there is only a handful of studies on the prevalence and risk factors for dementia in SSA that could inform context-specific interventions.

Ghana, a developing low and middle-income country in SSA, has seen rapid increases in the aging population over the past two decades, mainly due to the declining fertility rates and increasing life expectancy (Nyame et al., 2019). However, Alzheimer’s Disease and Related Dementia (ADRD) are largely understudied in Ghana. The latest World Health Organization (WHO) data published in 2018 indicated that deaths related to ADRD in Ghana reached 2,672 or 1.33% of total deaths (World Health Organization, 2018). The only study on prevalence conducted by the Kintampo Dementia Initiative (KDI), showed a standardized prevalence of 6.6% (95% CI-3.6–6.8), for all ages (Nyame et al., 2019). Regional estimates from population and hospital-based studies in SSA estimate the prevalence to range from 2.29-21.6% (Olayinka & Mbuyi, 2014). Similarly, another recent review with a meta-analysis of dementia in SSA (no studies from Ghana) noted a prevalence of 5.0% (Ojagbemi et al., 2021).

Estimating the population prevalence of ADRD in SSA, and identifying associated risk factors, has been identified as a public health priority (Guerchet et al., 2017). Although age and sex seem to be consistently associated with dementia in SSA studies, the evidence around other risk factors such as smoking, alcohol use, physical inactivity, the experience of stressful life events, vascular events, and others has not been thoroughly investigated at the country level. For example, a longitudinal cohort study conducted by Gerritsen et al. (2017) in a Swedish sample reported that the experience of multiple negative life events at baseline predicted a higher risk for dementia. Similarly, unhealthy lifestyles such as excessive alcohol consumption, and inadequate participation in physical and social activities are associated with an increased risk of dementia (Sabia et al., 2018; Dhana et al., 2020; Nguyen et al., 2024). These risk factors are suggested to impact the life course similar to other chronic conditions of late life. However, the evidence is mostly based on studies carried out in high-income countries other than SSA. This study thus aimed to explore the association between these factors and dementia risk in the Ghanaian elderly population.

Accurate and timely assessment and diagnosis of dementia are essential in providing optimal clinical care and management to patients. However, the assessment and diagnosis of dementia in both primary care and at the population level are complex and can be difficult (Ferri & Jacob, 2017; Social Care Institute of Excellence, 2020). Part of the complexity and difficulty in diagnosing dementia stems from the fact that the symptoms of the condition could be mistaken as evidence of normal aging (Slavin et al., 2013; Ferri & Jacob, 2017). In low-resource settings, this difficulty is further compounded by limited access to healthcare, low health literacy, the stigma associated with the condition, and the limited capacity to screen and diagnose dementia (Maestre, 2012; Orikiriza, 2020).

There are many tests/tools available which can be administered to screen for the existence of dementia. Usually, screening entails cognitive testing for a history of cognitive and functional decline (Prince et al., 2011). Most of the existing tools for screening for dementia have limited utility in low-to-middle-income countries. For example, the Mini-Mental State Examination tool which is widely used in high-income countries is prone to educational and cultural bias (Ng et al., 2007). Additionally, the General Practitioner Assessment of Cognition (GPCOG), the Memory Impairment Screen (MIS), and the Mini-Cog have all been deemed not suited for settings with limited education such as Sub-Saharan Africa (Prince et al., 2011).

On the other hand, the Community Screening Instrument for Dementia (CSI-D) has been extensively used and validated across many settings including sub-Saharan Africa. Both the longer and shorter versions have proven to be reliable and valid (Prince et al., 2011). CSI-D has been used extensively in several sub-Saharan African countries including Nigeria (Yusuf et al., 2011), Tanzania (Paddick et al., 2014), South Africa (De Jager et al., 2017), and Uganda (Mubangizi et al., 2020). The strengths of this instrument are that it is not literacy dependent, and secondly, the inclusion of the informant interview provides insight into the cognitive and functional abilities of the subject thereby leading to a high probability of predicting the presence of dementia (Yusuf et al., 2011).

Given the lack of evidence, the need for more research in Ghana and SSA is obvious. The present study aimed to conduct a population-based study to estimate the prevalence and key modifiable and non-modifiable risk factors for ADRD in Ghana. Our findings will guide the health systems in Ghana and other countries in the SSA region with a similar healthcare system to accelerate the attention to these people and their families in the utilization of formal and informal health services for better health outcomes.

Materials and methods

Study design

This was a quantitative community-based cross-sectional study. The study was carried out in six (6) randomly selected rural communities in the Juaben Municipal Assembly in the Ashanti Region and the Adentan Municipal Assembly in the Greater Accra Region. A multi-stage sampling method was employed by first randomly selecting two districts, followed by a random selection of three communities in each of the municipalities using a table of random numbers. Finally, in each community that was selected, a door-to-door survey was conducted in a random direction by trained research assistants in the Department of Health Promotion and Disability Studies of the Kwame Nkrumah University of Science and Technology. The first house to be visited was determined randomly by spinning a pen at the center of the community and the first house in the direction where the pen pointed became the starting point of the sample recruitment until the number of participants allocated to each community was reached.

The study was conducted among persons aged 60 years or older based on available evidence, which suggests that ADRD is more common among this subpopulation (World Health Organization (WHO), 2018). In addition to the primary study participants, close relatives of the primary respondents were also interviewed as part of the survey. The addition of the close relations of individuals suspected to have dementia enhances the validity of the data-gathering instrument used (Mubangizi et al., 2020; Prince et al., 2011). A minimum sample size of 384 was calculated based on an expected prevalence of 6.6.% and a precision of ± 5% at a 95% confidence level. The Committee on Human Research, Publications, and Ethics, an institutional review board of the Kwame Nkrumah University of Science and Technology, Ghana approved the study protocol. The ethical approval reference number is CHRPE/AP/057/21.

Table 1. Socio-demographic profile of informants.

Variable value	Mean (SD)	Minimum/maximum value
Age (in completed years)	38 (16.28)	11 − 82
	Frequency (384)	Percentage (%)
Highest Educational level
None	32	8.33
Primary school	64	16.67
Junior high/Middle school	164	42.71
Secondary school	77	20.05
Tertiary	47	12.24
Relationship to primary participant
Spouse	41	10.68
Sibling	18	4.69
Child	148	38.54
Grandchild	66	17.19
Other relatives	58	15.10
Non-biological relation	53	13.80

Instrumentation

The study utilized the brief Community Screening Instrument for Dementia (CSID) which is a validated and reliable tool for estimating the prevalence of Alzheimer’s disease and related dementia (Mubangizi et al., 2020; Prince et al., 2011). The tool consists of two parts, a cognitive scale, and an informant scale. The former is for the primary respondent who is suspected to have dementia while the other part (informant scale) is for a close relation of the primary respondent. The addition of the informant scale is meant to validate the scores obtained from the cognitive scale for the index participant. The cognitive scale for the primary respondent was made up of seven items with a minimum score of 0 and a maximum score of 9. The rating score for the cognitive scale was as follows: 0-4 indicates probable dementia; 5-6 indicates possible dementia and 7-9 is deemed as normal. For the participants who scored 5-6, the researchers interviewed their close relations. In the end, the primary respondents were classified as screened positive for dementia if they had a CSID total score of ≤ 4. We included additional questions on correlates of dementia such as socio-demographic factors, family history, lifestyle, activities of daily living, and experience of stressful life events. Activities of daily living were assessed with eleven (11) items that were adopted and modified from a previous study (Gureje et al., 2006) with a minimum and maximum total score of 0 and 33 respectively. The Likert scale adopted ranged from 0 (unable to do) to 3 (can do without difficulty). A lower score indicated difficulty in performing activities of daily living. Experience of stressful life events was investigated for three different life periods: 0-15 years; 16-64 years; and 65 years and above. We adopted and modified the set of psychosocial risk factors previously described by Persson & Skoog (1996). For the experience of stressful life events from 0-15 years and 16-64 years, there were five (5) items each with a Yes/No response while for the experience of stressful life events from 65+, there were eight (8) items with a Yes/No response. A higher score indicated experience of more stressful life events. Data for this study were collected from May to June 2021.

Calculation of prevalence

Due to the potential of the brief CSID instrument to yield a higher prevalence of dementia over the true prevalence of dementia, we utilized the back estimation method previously described in a related subject (De Jager et al., 2017; Mubangizi et al., 2020) to estimate the true prevalence of dementia. We relied on the screened positive data from our study and the previously published sensitivity (0.95) and specificity (0.90) values of the brief CSID tool. These sensitivity and specificity values have been reported by Prince et al. (2011) and Mubangizi et al. (2020). Based on the sensitivity and specificity values of the brief CSID tool, we calculated the number of true positives, test positives, and test negatives. Test positives are participants who screened positive using the study instrument while test negative were participants without the outcome. These data were inputted into the following formula together with the established sensitivity and positivity values to estimate the true positives. Finally, the number of true positives was then used to estimate the true prevalence rate of dementia. The number of true positives was derived using: $$\mathit{\text{TP}}=\frac{\mathit{\text{testp}}-\left(\mathit{\text{testn}}\times \frac{1-\mathit{\text{SP}}}{\mathit{\text{SP}}}\right)}{1-\left(\frac{1-\mathit{\text{SP}}}{\mathit{\text{SP}}}\times \frac{1-\mathit{\text{SE}}}{\mathit{\text{SE}}}\right)}$$

Data analysis

All data analyses were performed using STATA software version 17. The dependent variable for the analysis was the presence/absence of dementia while the independent variables were the participants’ socio-demographic characteristics, lifestyle factors, and experience of stressful life events across the life course. All statistical tests were considered significant if the probability value was less than 0.05 (P-value <0.05).

Results

Socio-demographics of primary participants

The mean age of primary participants was 72 years (SD = 10.22). The majority of participants were females (69%), and Christianity was the dominant religion that participants disclosed affiliation with (87%). The majority of participants had ‘Junior high school’ as their highest educational attainment (40%), with 33% having no formal education. As regards marital status, the majority of participants (45%) were widows/widowers; 40% were married. The majority of participants (58%) were unemployed. There was a participant response rate of 100%. Further details of participants’ characteristics are presented in Table 1. The mean age of the informant participants was 38 years (SD= 16.28). The majority of the informants were also females (69%). More than a third (42.71%) of the informants had ‘Junior high school’ as their highest level of educational attainment. Most of the informants were biologically related to the primary participants. The profiles of the informants are presented in Table 2.

Table 2. Socio-demographic profile of participants.

Variable	Mean (SD)	Minimum value/ Maximum value
Age (in completed years)	72 (10.22)	60–130
	Frequency (n = 384)	Percentage (%)
Marital Status
Married	153	40
Single	3	0.8
Cohabiting	1	0.2
Divorced	55	14
Widowed	171	45
Highest Educational Level
No formal education	127	33
Primary school	58	15
Junior high school	152	40
Senior high school	14	4
Tertiary	33	9
Employment Status
Unemployed	222	58
Employed	162	42
Religious Affiliation
Christianity	335	87
Islam	33	9
African Traditional Religion	8	2
Other	8	2

Prevalence of dementia

In the study sample of 384 participants who were screened, 13 (3.39%) of them recorded a CSID cognitive score of ≤ 4. Out of the 384 respondents, another 74 (19.27%) had a CSID cognitive score of 5 or 6. For these 74 participants, the informant score was subtracted from the corresponding primary participants’ CSID cognitive score. The resultant data showed that 44/74 had a CSID total score of ≤ 4. Therefore, the total number of participants who screened positive for probable dementia based on the combined CSID cognitive score and the CSID total score was 57 (15%) obtained by the addition of 13 and 44. We then used the back estimation method already described to estimate the true prevalence. From the estimation, the true positives were 62 thereby giving a prevalence rate of 16% (62/384) with a positive predictive value (PPV) of 62% and a negative predictive value (NPV) of 99% (Table 3). Participants who fall within the category of ‘possible dementia’ were 8% (30/384), while the remaining participants (77%/297) were identified as ‘normal’. As regards the regional distribution of probable dementia, it was observed that the condition was more prevalent in the Ashanti region (17%) than in the Greater Accra region (12%). Table 4 has some additional information on the regional distribution of the condition.

Table 3. Estimates of prevalence based on sensitivity & specificity calculations.

Categories	Frequency	Percentage
Apparent prevalence	57	15
True prevalence	62	16
Sensitivity	–	95
Specificity	–	90
Positive predictive value	–	62
Negative predictive valued	–	99

Table 4. Lifestyle and psychosocial risk factors.

Variable	Mean (SD)	Minimum value/ Maximum value
ADL	29.59 (4.35)	0-33
Experience of stressful life events (0-15)	1.46 (1.18)	0-4
Experience of stressful life events (16-64)	2.25 (1.27)	0-5
Experience of stressful life events (65+)	2.13 (1.76)	0-8
Average no. of days in a week spent doing moderate exercise	3.31 (2.22)	0-7
Average minutes spent doing moderate exercise in a day	89.04 (83.98)	10-510
Average number of days in a week spent walking	3.92 (2.29)	1-7
Average minutes spent walking in a day	75.83 (80.04)	10-600
	Frequency (n = 384)	Percentage (%)
Smoking status
Never Smoked	320	83.33
Ever smoked	55	14.32
Currently smoking	9	2.34
Alcohol consumption
Yes	103	26.82
No	281	73.18
History of stroke
Yes	43	11.20
No	341	88.80
Experience of head injury
Yes	93	24.22
No	291	75.78

Lifestyle and psychosocial risk factors

The findings indicated that most of the participants were able to perform the activities of daily living such as bathing, dressing, reaching, and toileting (29.59 ± 4.35). The experience of stressful life events from age 0-15 was minimal (1.46 ± 1.18) compared with the experience of stressful life events from age 16-64 and 65+. Further, most of the participants (83.33%) had never smoked while nearly a fourth had a history of head injury (24.22%). The lifestyle risk factors are presented in Table 4.

Factors associated with dementia

Socio-demographics and probable dementia

Socio-demographic variables showing significant association with probable dementia in a binary logistic regression model were index participants’ age, education, religion, and employment; sex and marital status did not show significant association. When these variables (i.e. showing significant association) were further examined in a multivariable logistic regression model, age (p = 0.035; AOR = 1.031; 95% CI 1.002 – 1.061), education (p = 0.011; AOR = 0.689; 95% CI 0.517 – 0.919) and employment (p = 0.004; AOR = 0.320; 95% CI 0.149-0.685) retained a significant association with probable dementia. Whereas age had a positive association with probable dementia, education, and employment both had a negative association with probable dementia. Additional information is presented in Table 5.

Table 5. Socio-demographics and probable dementia.

Variables	Probable Dementia
	Screened Prevalence %	Crude p-value	Adjusted p-value [Beta Coef.]**	Adjusted Odds Ratio [95% CI]**
Age		0.000	0.035 [0.030]*	1.031 [1.002-1.061]
Education
None	25.2	0.000	0.011 [−0.371]*	0.689 [0.517-0.919]
Primary school	13.8
Junior High School	9.9
Senior High School	0.0
Tertiary	2.0
Religion
Christianity	13.1	0.007	0.060 [−0.412]	0.662 [0.431-1.016]
Islam	21.2
African Traditional	37.5
Non-religious	37.5
Employment
Unemployed	21.2	0.000	0.004 [−1.145]*	0.320 [0.149-0.685]
Employed	6.2
Sex
Male	12.5	0.385	–	–
Female	15.9
Marital Status
Cohabiting	0.0	0.409	–	–
Single	33.3
Divorced	12.7
Widowed	18.1
Married	11.8

** Explanatory variables included in a multivariable logistic regression model were age, education, religion, and employment.

Lifestyle factors and probable dementia

Lifestyle factors that showed significant association with probable dementia in a binary logistic regression analysis were; ‘experience of stressful life events from age 16-64’, ‘experience of stressful life events from age 65+’, and ‘activities of daily living’. Each of the aforementioned variables was examined further in a multivariable logistic regression model adjusting for socio-demographic variables – age, education, religion, and employment. All three lifestyle variables retained statistically significant association, though ‘experience of stressful life event from age 65+’ had a relatively weak association. Both ‘experience of stressful life event from age 16-64’ and ‘experience of stressful life event from age 65+’ showed a positive association with probable dementia (p = 0.024; AOR = 1.325; 95% CI 1.034-1.698 and p = 0.042; AOR = 1.258; 95% CI 1.024-1.546 respectively), while ‘activities of daily living had a negative association with probable dementia (p = 0.037; AOR = 0.925; CI 0.868-0.986).

On the other hand, variables including smoking status, alcohol behavior, physical activity, history of stroke, and an experience of head injuries did not show any statistically significant association with probable dementia in a binary logistic regression analysis. Additional information is presented in Table 6.

Table 6. Lifestyle factors and probable dementia.

Variables	Probable Dementia
	Crude p-value	Adjusted p-value [Beta Coef.]⸸⸸	Adjusted Odds Ratio [95% CI]⸸⸸
Experience of stressful life events from age 16-64	0.003	0.024 [0.282]	1.325 [1.034-1.698]
Experience of stressful life events from age 65+	0.000	0.042 [0.212]	1.258 [1.024-1.546]
Activities of daily living	0.000	0.037 [−0.065]	0.925 [0.868-0.986]
Smoking status	0.314	–	–
Alcohol behavior	0.427	–	–
Physical activity (vigorous exercise, moderate exercise)	0.619 0.185	–	–
History of Stroke	0.791	–	–
Experience of head injuries	0.788	–	–

^⸸⸸ Explanatory variables were adjusted for age, education, religion, and employment.

Discussion

Prevalence of dementia

In this study of 384 older adults from Ghana, we estimated the population prevalence of probable dementia to be 16% based on the sensitivity and specificity analysis described above while the apparent prevalence was 15%. In Europe, the prevalence of dementia has been estimated at 5.05%, lower than that found in our study (Niu et al., 2017). Similarly, about 5.0% of Americans aged 65 -74 are living with dementia, a rate that is lower than what we found in the current study (Alzheimer’s Association, 2024). The disparities in the prevalence rate between the current study and those reported in Europe and America could be attributable to diagnostic tools deployed coupled with variations in the age ranges of the study population.

Furthermore, in Uganda (Mubangizi et al., 2020) and South Africa (De Jager et al., 2017) where a similar screening tool was used, the prevalence of dementia based on the back estimation method in our study is higher than the apparent prevalence. This discrepancy could be a result of the variations in the actual prevalence of dementia in the screened population in the different countries and or the sensitivity and specificity values used in the estimation. The estimated probable prevalence in this study is higher than the 6.38% noted in the report by Alzheimer’s Disease International in 2017 (Guerchet et al., 2017) and the 5.0% reported in the recent meta-analysis of studies in SSA (Ojagbemi et al., 2021). Recently, studies in Uganda (Mubangizi et al., 2020) and Tanzania (Longdon et al., 2013) using a similar screening tool and with a similar age distribution, recorded a prevalence of 20% and 15.4% respectively. The higher prevalence of dementia in the current study and previous studies in Uganda (Mubangizi et al., 2020) and Tanzania (Longdon et al., 2013) could be a result of the methodology used – door-to-door screening rather than hospital-based screening and detection. This is because when the Diagnostic and Statistical Manual of Mental Disorders (4^th edition) tool was applied to the same sample in the Tanzanian study, the prevalence of probable dementia dropped to 6.4%. It has been noted that the CSID screening tool yields high prevalence estimates compared with structured clinical interviews explaining the high prevalence in our study and those in Uganda and Tanzania. However, with an NPV value of nearly 100%, the probability that a participant with normal cognitive functioning was screened as positive in the current study was very minimal. Secondly, the choice of the brief CSID was influenced by its suitability for use by non-specialists in a low-resource setting like Ghana (Prince et al., 2011; Mubangizi et al., 2020).

Socio-demographic factors and dementia

Consistent with other findings in the literature, our study suggested that age is an important factor associated with dementia both at the bivariate and multivariate levels. The odds of having dementia increased with advancing age in this study population. This finding is supported by a similar study conducted in Rwanda which found that probable dementia positively correlated with age even after adjusting for possible confounders (Mubangizi et al., 2020). One of the plausible explanations is that aging is accompanied by changes in brain health including activation of inflammation and weakening of the neurons in areas related to memory (Longdon et al., 2013). Also, our findings showed that formal education had an inverse association with probable dementia. Again, the association between fewer years of schooling and the experience of dementia is well established. In one study, the investigators found that the risk of dementia was twice as high among participants with fewer years of schooling compared to those with more years of schooling (Wu et al., 2011). From the literature, it has been established that formal education or staying longer in school is protective against the development of dementia (Sharp & Gatz, 2011; Langa et al., 2017). One of the plausible explanations for this association is that education helps the brain to develop more synapses that may boost a person’s cognitive reserve thereby helping to prevent dementia. Besides, highly educated people tend to have a healthier lifestyle than those with less education. For instance, people with more education may be aware of the dangers associated with smoking, lack of exercise, and poor nutrition and thus may be motivated to adopt healthier choices.

Additionally, the current study observed that the risk of dementia was associated with being unemployed compared to being employed. Several studies have established a correlation between being unemployed or retired and the risk of dementia (Mythily et al., 2015; Kivimäki et al., 2021). The protection offered by being employed could be attributed to the cognitive stimulation offered by the execution of job-related activities particularly mentally stimulating occupations. Furthermore, the financial returns from being employed and the meaning, purpose, and satisfaction derived from the same may also insulate people from stress or societal pressures that can be damaging to an individual’s mental health and well-being.

The current study did not find any statistically significant association between religious affiliation and dementia. This finding is at variance with a study conducted by Lin and his colleagues in a Taiwanese sample where a strong correlation was established (Lin et al., 2015). The investigators found that participants with affiliation to Christianity showed a decreased dementia risk (adjusted odds ratio [AOR] = 0.46, 95% confidence interval [CI] = 0.25–0.87) compared with those without any religious affiliation. One of the possible explanations for this association is that involvement in religious activities promotes active and stimulating social engagement which are vital in slowing cognitive decline. The lack of statistical significance in our study could thus be attributed to possible variations in their level of active engagement in the activities of their particular religion (Holtzman et al., 2004; Moxey et al., 2011).

Lifestyle factors and dementia

From the results, the primary participants who experienced notable stress from age 16 to 64 years were 1.3 times more likely to have probable dementia. Similarly, index participants who indicated having experienced notable stress from age 65 years were 1.2 times more likely to have probable dementia. Stress and its association with dementia have been reported by several authors (Radford et al., 2017; Sindhi et al., 2017; Nabe-Nielsen et al., 2020). A recent systematic review and meta-analysis by Severs et al. found that the experience of traumatic events such as war trauma and childhood trauma was associated with an increased risk of dementia. Conversely, in a longitudinal study involving the adult population of Sweden, the investigators did not find an association between participants’ experience of stressful life events and dementia (Sundström et al., 2014). As Sundström et al. (2014) have previously noted, the contradictory findings in the literature regarding the association between stressful life events and dementia risk to a certain degree might be due to differences in the study methodologies used. Another possible reason for this difference is the variation in individual responses to stressful life events. One individual might consider a certain event as extremely stressful whereas another might perceive it as rather trivial thereby subsequently producing differential outcomes in a person’s life.

As regards index participants’ ‘activities of daily living’ (ADL) and its association with probable dementia, it can be observed from the results that a participant was less likely to have probable dementia, for any given unit increase in the ADL of that participant. Thus, ADL was useful in reducing the likelihood of a participant having probable dementia. In a recent study conducted in Uganda in which authors explored the potential association between an ADL-related variable ‘physical activity’ and probable dementia, authors reported an inverse significant association between the two variables which corroborates the findings of our study (Mubangizi et al., 2020).

Similar to the findings of the study conducted in Uganda our study did not establish a significant association between the variables smoking, alcohol use, and vascular accidents on one hand, and probable dementia on the other hand (Mubangizi et al., 2020). There are a couple of reasons that could explain why these variables did not present as risk factors in this study despite their known association with dementia. First, it is possible that the sample size of 384 was not sufficiently large to detect the existence of a possible association between these known risk factors and dementia. Second, the lack of association between these factors and dementia could be attributed to the intensity of these factors in the study population. For example, a previous study established a higher risk of dementia in people who abstain from alcohol or consume >14 units/week compared with long-term moderate consumption of 1-14 units/week (Sabia et al., 2018). Concerning the link between smoking and the risk of dementia, previous studies have suggested that an epigenetic mechanism may be involved (Kawakami et al., 2023). For example, Kawakami et al. (2023) observed that the dementia risk associated with smoking was higher in middle-aged people (<65 years) compared with older people (≥65 years). These findings confirm the complex relationships between these risk factors and the risk of developing dementia.

Limitations of the study

The current study is subject to a few limitations that need to be acknowledged. First, the screening tool (CSID) employed in this study has the potential to overestimate the actual prevalence of dementia due to its ability to pick out early dementia or mild cognitive impairment. Despite this, the brief CSID has proven to be a valid and reliable screening tool for dementia (De Jager et al., 2017; Mubangizi et al., 2020). Second, the present study relied on self-reported data which is subject to the possibility of recall bias especially when participants are asked to recollect historical information. Finally, other important risk factors for dementia such as obesity, diabetes, hearing loss, and hypertension were not examined in this study considering the challenges one could potentially encounter in investigating all possible risk factors in a single study.

Despite these limitations, the present study provides valuable population-based information for planning and designing supportive interventions, policy directions, and research agendas for Alzheimer’s disease and related dementia in Ghana and elsewhere in SSA.

Conclusion

We found a higher prevalence of dementia in six rural communities in Ghana than what has been previously reported. Although the prevalence rate is somewhat comparable to similar studies elsewhere in SSA, the relatively high rate should be a major source of concern for the public health community considering that the disease burden is projected to increase in the foreseeable future. Older age, lower education, being unemployed, decreased functioning in activities of daily living, and experience of stressful life events were found to be associated with dementia risk. Although there is currently limited evidence to support a cause-effect relationship between any preventative strategy and the development of dementia, we recommend that improving access to educational and employment opportunities will help address some of the identified risk factors. In addition, we recommend that staying physically active and maintaining connections with family and friends is likely to contribute to alleviating decreased functioning in activities of daily and the experience of stressful life events respectively.

In terms of directions for future research, we suggest that similar studies should be carried out in other parts of the country to ascertain the overall burden of dementia in Ghana. Secondly, we recommend that future studies in Ghana and elsewhere in SSA should explore other known risk factors of dementia such as diabetes, hypertension, hyperlipidemia, depression, epilepsy, and genetic polymorphisms which were not examined in this study to ascertain their relevance in the sub-Saharan population.

Ethics approval and consent to participate

Ethical clearance was obtained from the Committee on Human Research, Publication, and Ethics, an institutional review board at Kwame Nkrumah University of Science and Technology, Ghana. Informed consent was obtained from all participants and the study was implemented per relevant guidelines/regulations, including the Declaration of Helsinki.

Acknowledgments

We wish to thank the study respondents for their willing participation in this study. Finally, our sincere gratitude goes to all the research assistants at the Department of Health Promotion and Disability Studies at the Kwame Nkrumah University of Science and Technology, Kumasi who assisted with the data collection.

Disclosure statement

The authors declare that they have no competing interests.

Data availability statement

The data that support the findings of this study are available from the corresponding author, [PO], upon reasonable request.

Additional information

Funding

This study was funded by grant D43-TW007267 from the Fogarty International Center, US National Institutes of Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Notes on contributors

Paul Okyere

Paul Okyere, Ph.D., MSc, is a Senior Lecturer in the Health Promotion and Disability Studies department at the Kwame Nkrumah University of Science and Technology (KNUST), Ghana. He studies the influence of behavioral, cultural, and social determinants of health on individual and population health outcomes. His interests particularly revolve around safety promotion, injury prevention, chronic noncommunicable diseases, and maternal and reproductive health issues. He is currently the coordinator of the Ph.D. program in the School of Public Health at KNUST, where his department is located.

Emmanuel Appiah-Brempong

Emmanuel Appiah-Brempong, Ph.D., is a Senior Lecturer of Public Health, researcher and development consultant with both national and international experience. He holds a Ph.D. in Public Health and has a research focus on Health Promotion &amp; Disease Prevention. Emmanuel has advanced experience with both quantitative and qualitative research approaches and does research in areas including non-communicable disease (NCD) prevention, health impact assessment (HIA), health behavior modelling, social &amp; behavior change communication (SBCC), WASH and health, and mental health promotion. He has particular interest in intervention research, especially cluster-randomized controlled trials.

Arti Singh

Arti Singh, Ph.D., is a Senior Lecturer at the Department of Epidemiology and Biostatistics at Kwame Nkrumah University of Science and Technology. She is a clinician by training but also has a PhD in Epidemiology. Her research interests include non-communicable diseases, qualitative methodology, and community engagement. She is motivated to make a personal contribution to global health research.

Suparna Qanungo

Suparna Qanungo, Ph.D., is an Associate Professor in the College of Nursing at the Medical University of South Carolina. Dr. Qanungo’s research interest lies in global health, community-based supportive care for cancer and aging. Her focus is to develop and implement interventions for underserved communities to improve health outcomes, access to care and quality of life, locally, and globally.

Peter Donkor

Peter Donkor is a Professor/Consultant in Oral and Maxillofacial Surgery at the Kwame Nkrumah University of Science and Technology (KNUST) and Komfo Anokye Teaching Hospital (KATH) Ghana. His research interests include cleft lip and palate, oral cancer, injury control, global surgery, and medical education.

Charles Mock

Charles Mock, Ph.D., is a Professor Emeritus of Surgery and Epidemiology at the University of Washington, USA. Charles has extensive international experience in global health research particularly in the area of injury prevention and control. His research interests are in epidemiology of injuries, especially in developing countries; surgical and trauma outcomes; and treatment of injuries.