Introduction: Identifying health conditions of persons with cognitive impairment (PCI) in the community and exploring their implications for caregiving experience are vital for effective allocation of healthcare resources. This study examined distinct PCI health profiles among community-dwelling PCI and their association with caregiving burden and benefits. Methods: Latent profile analysis and multivariable regression were applied to dyadic data from 266 PCI and their caregivers in Singapore. Results: Three PCI health profiles were identified: less impaired (40% of PCI), moderately impaired (30%), and severely impaired (30%). Caregivers for severely impaired PCI were more likely to report a higher level of caregiving burden, and caregivers for moderately impaired PCI were more likely to report a higher level of caregiving benefits, compared to caregivers for less impaired PCI. Conclusion: The findings captured heterogeneity in health status among PCI in the community. Tailored interventions, based on PCI health profiles, should be designed to reduce caregiving burden and increase caregiving benefits.

Globally, approximately 75% of persons with dementia remain undiagnosed [1]. Accordingly, a majority of persons with cognitive impairment (PCI) and their family caregivers live in the community without receiving sufficient attention from the healthcare system [2]. Identifying health disparities among PCI in the community is thus important to capture the most vulnerable PCI for effective allocation of limited healthcare resources.

Intuitively, there are multiple dimensions of PCI health (e.g., cognitive, functional, or mental) that may interact in different ways. For instance, some PCI with cognitive impairment may not have difficulties in activities of daily living (ADLs) or mood problems, while others suffer from multimorbidity with functional and behavioral limitations [3]. However, studies on this topic rarely focus on combinations of different health status among PCI, making it difficult to comprehensively capture PCI at high health risk in the community. To address this gap, this study applies a clustering approach to identify distinct health profiles among PCI, using multiple instruments that measured cognitive, functional, behavioral, and mood impairments of PCI.

Identifying health profiles among PCI opens the possibility to explore whether PCI health impacts their caregivers. Caregiving has been portrayed in a negative light, in association with poorer physical and mental health and higher mortality among caregivers [4, 5]. In particular, caregiver stress process theory proposes that care recipients’ poor health is a primary stressor that impairs health and well-being of caregivers [6, 7]. Relevant studies thus probed the association between PCI health and a higher level of caregiving burden [8], including deteriorating health, as well as financial adversity and disruption in social life [9]. However, little is known about whether PCI health is associated with positive aspects of caregiving, although studies increasingly pay attention to whether and to what extent caregiving helps promote physical and mental well-being of caregivers [10, 11]. Since caregiving burden and benefits are not mutually exclusive [12], caring for a PCI in poor health may also provide a sense of meaning and satisfaction to caregivers [13] while allowing caregivers to be physically active [14].

In sum, the present study aimed to identify (1) distinct health profiles of community-dwelling PCI using multiple health instruments and (2) the association of PCI health profiles with caregiving burden and benefits. The study was conducted in Singapore, an Asian city-state with one of the world’s fastest aging populations; by 2050, about 40% of Singaporeans will be aged 60 or above [15]. A study conducted in 2015 estimated that about one in ten older Singaporeans had dementia [16], and the numbers will rise due to increasing life expectancy and population aging. Under these circumstances, exploring health profiles among PCI and their implications for family caregivers is important in Singapore, where family caregiving is prevalent and pivotal, presumably due to a national drive toward aging in place [15] under a culture of filial piety embedded in the Confucian cultural legacy [17].

Participants and Data Collection

Dyadic data on 266 pairs of PCI and their caregivers were utilized from the “Caring for persons with dementia and their caregivers in the community: Towards a sustainable community based dementia care system (COGNITION)” study. The COGNITION study aimed to understand the health and well-being of PCI and their caregivers in the community in Singapore [18]. In 2018, the study team visited 9,828 households in the community and administered a 10-item screener [19] to 3,589 older adults, aged 60 years and above. The screener, called AD8Plus, includes eight questions from the Eight-item Interview to Differentiate Aging and Dementia (e.g., repeats the same things over and over; forgets the correct month or year) and two questions (e.g., three-item recall and copying two intersecting pentagons) from the Mini-Mental State Examination (MMSE). Of those, 323 older adults, who scored 8 or less on the screener, were considered to have cognitive impairment and invited to the study. An eligible caregiver should be a family member or a friend aged 21 years or older who was most involved in providing direct care to PCI or ensuring the provision of care to PCI. The final sample consisted of 266 PCI and their caregivers, who provided written informed consent for study participation (online suppl. Fig. 1; for all online suppl. material, see https://doi.org/10.1159/000530606 for the sample flowchart). If PCI were unable to respond due to health reasons, consent was obtained from PCI’s legal representative or next of kin.

Measures

PCI Health Indicators

Eight health indicators derived PCI health profiles. First, (1) chronic conditions were a sum of medical ailments, such as cancer, diabetes, and high blood pressure, that PCI have ever been diagnosed with by medical professionals. The questions were answered by caregivers, with a possible range from 0 to 19. Second, two instruments measured cognitive impairment of PCI: (2) AD8Plus, a brief cognitive test for early dementia [19], and (3) the full 30-item battery of MMSE, which comprehensively assesses the orientation, attention, memory, language, and visual-spatial skills [20]. For both instruments, originally, the higher the score, the better the cognitive function. However, to match with other indicators, the scores were reverse-coded so that higher scores reflected worse cognitive function.

Third, the Revised Memory and Behavior Problems Checklist (RMBPC) [21] assessed problems in (4) memory, (5) behavior, and (6) mood of PCI from the perspective of caregivers. Caregivers were asked how frequently they had encountered three types of problems in their PCI during the past week: memory (7 items; e.g., “asking the same question over and over” and “forgetting what day it is”), behavior (8 items; e.g., “talking loudly and rapidly” and “threats to hurt others”), and mood (9 items; e.g., “appear sad and depressed” and “talking about feeling lonely”). There were five response categories: “never occurred at all = 0”; “not in the past week = 1”; “1–2 times per week = 2”; “3–6 times per week = 3”; “daily or more often = 4.” For each domain, a summated score was constructed. A higher score indicated more severe memory, mood, and behavior problems.

Lastly, (7) Activities of Daily Living (ADLs) and (8) instrumental ADLs (IADLs) difficulties measured functional limitations. ADLs assessed PCI’s difficulties in seven basic ADLs such as eating and getting dressed; IADLs evaluated PCI’s difficulties in seven types of IADLs, including shopping, preparing meals, and taking medicine. PCI responded to these questions; if PCI were unable to answer due to health reasons (about 50% of the PCI), their caregivers answered them as a proxy. Both measures had three response categories: “completely unable,” “with some help,” and “without help.” The former two responses were collapsed into one, and the total number of difficulties for ADLs and IADLs were separately counted.

Caregiving Outcomes: Burden and Benefits of Caregiving

The Zarit Burden Interview (ZBI) [22] and the Short-Positive Aspects of Caregiving (S-PAC) scale [23] measured burden and benefits of caregiving, respectively. The ZBI has been widely used to measure caregiving burden and validated in Singapore [24]. The 22-item instrument had five response categories: “never = 0,” “rarely = 1,” “sometimes = 2,” “quite frequently = 3,” and “nearly always = 4.” The summated score was used, with a possible range from 0 to 88. The higher the score, the higher the burden.

The seven-item S-PAC scale evaluated perceived benefits of caregiving, asking to what extent caregiving contributes to self-affirmation and outlook on life among caregivers. Each item had five response categories: “disagree a lot = 1,” “disagree a little = 2,” “neither agree nor disagree = 3,” “agree a little = 4,” and “agree a lot = 5.” The summated score was used, with a possible range from 5 to 35. The higher the score, the higher the caregiving benefits.

Covariates: PCI and Caregiver Characteristics

PCI characteristics comprised age (range: 61–103 years), gender (female = 1; male = 0), marital status (married = 1; widowed/separated/divorced/never married = 0), and dementia diagnosis status (formally diagnosed with dementia = 1, the rest = 0). Caregiver characteristics included age (range: 23–93 years), female (female = 1; male = 0), minority ethnicity (Malay, Indian, and other nationalities = 1; Chinese = 0), married (married = 1; non-married = 0), highest completed education (no formal education = 1; primary school = 2; secondary = 3; postsecondary and tertiary = 4), working status (working full-time or part-time = 1; not working or never worked = 0), financial adequacy (usually inadequate = 1; occasionally adequate = 2; adequate = 3; more than adequate = 4), poor self-rated health (poor or fair health = 1; good, very good, or excellent health = 0), coresident caregiver (caregiver live in the same household with PCI = 1; caregiver did not live with PCI = 0), and long-term caregiver, of 5 years or more (helping PCI for memory problems for more than 5 years = 1; the rest = 0).

Analytic Strategy

Latent profile analysis (LPA) identified health profiles among PCI. LPA unveils unobserved profiles within a population, based on individuals’ response patterns to a set of continuous indicators [25]. The core part of LPA is to determine an optimal number of profiles. The models with different numbers of profiles were fitted, and model fit indices were compared. The model with the lowest information criteria values and statistically significant likelihood tests [26] was deemed to fit the data better than others. The entropy index above 0.8, and a proportion of the smallest class above 5% were further considered for reliability and precision of profile distinction. Individuals were then assigned to their most likely profile based on the highest posterior probabilities. Using this information, multivariable regression tested the association of PCI health profiles with caregiving burden and benefits, taking other covariates into account. Missing values, about 14% in the eight health indicators, were handled by the robust maximum likelihood estimator in LPA. Multiple imputations with predictive mean matching imputed about 5% of missing values in the regression model [27].

Sample Characteristics

Table 1 provides sample characteristics. The average age of PCI was 81 years old; 57% of them were females, 47% were married, and 37% were formally diagnosed with dementia. Caregivers were on average 63 years old; 59% of them were females, 10% were ethnic minorities, 65% were married, 44% were working, 31% suffered from poor health, 85% lived with their PCI, and 17% provided care for memory issues for more than 5 years.

Table 1.

Variables in the analysis

VariableMean (SD)/%Range
PCI health indicators 
 Chronic conditions 2.42 (1.84) 0–19 
 Cognitive function 
  AD8Plus 5.22 (2.63) 2–10 
  MMSE 14.66 (8.18) 0–30 
 RMBPC 
  Memory problems 9.05 (6.92) 0–28 
  Behavior problems 2.56 (4.09) 0–32 
  Mood problems 3.81 (5.43) 0–36 
 Functional limitations 
  ADL difficulties 2.18 (2.73) 0–7 
  IADL difficulties 4.17 (2.77) 0–7 
Caregiving experience 
 Caregiving burden 23.90 (13.58) 0–88 
 Caregiving benefits 26.24 (5.77) 0–35 
PCI characteristics 
 Age 81.39 (8.60) 61–103 
 Female 57.1  
 Married 47.4  
 Dementia diagnosed 36.8  
Caregiver characteristics 
 Age 62.85 (13.79) 23–93 
 Female 59.4  
 Ethnic minority 9.8  
 Married 64.7  
 Education 2.73 (0.99) 1–4 
 Working 44.4  
 Financial adequacy 2.37 (0.84) 1–4 
 Poor self-rated health 31.2  
 Coresident caregiver 85.0  
 Long-term caregiver, of 5 years or more 16.5  
VariableMean (SD)/%Range
PCI health indicators 
 Chronic conditions 2.42 (1.84) 0–19 
 Cognitive function 
  AD8Plus 5.22 (2.63) 2–10 
  MMSE 14.66 (8.18) 0–30 
 RMBPC 
  Memory problems 9.05 (6.92) 0–28 
  Behavior problems 2.56 (4.09) 0–32 
  Mood problems 3.81 (5.43) 0–36 
 Functional limitations 
  ADL difficulties 2.18 (2.73) 0–7 
  IADL difficulties 4.17 (2.77) 0–7 
Caregiving experience 
 Caregiving burden 23.90 (13.58) 0–88 
 Caregiving benefits 26.24 (5.77) 0–35 
PCI characteristics 
 Age 81.39 (8.60) 61–103 
 Female 57.1  
 Married 47.4  
 Dementia diagnosed 36.8  
Caregiver characteristics 
 Age 62.85 (13.79) 23–93 
 Female 59.4  
 Ethnic minority 9.8  
 Married 64.7  
 Education 2.73 (0.99) 1–4 
 Working 44.4  
 Financial adequacy 2.37 (0.84) 1–4 
 Poor self-rated health 31.2  
 Coresident caregiver 85.0  
 Long-term caregiver, of 5 years or more 16.5  

N = 266; listwise deletion applied (N = 256 for caregiving benefits; N = 263 for financial adequacy and MMSE).

PCI, persons with cognitive impairment; AD8, Eight-item Interview to Differentiate Aging and Dementia; MMSE, Mini-Mental State Examination; RMBPC, Revised Memory and Behavior Problems Checklist; ADL, Activities of Daily Living; IADLs, Instrumental Activities of Daily Living.

Three PCI Health Profiles

LPA model fit indices appear in Table 2. The information criteria continued to drop when an additional profile was added. However, the smallest class only occupied 2% of the total sample for models with 5 and 6 profiles. According to the likelihood test, the model with 4 profiles was not superior to the model with 3 profiles. Therefore, a 3-profile model was selected.

Table 2.

Model fit indices for latent profiles

ProfilesICLikelihood testEntropySmallest class, %
AICBICaBICp (VLMR)p (BLRT)
5,930 5,987 5,936 NA NA   
5,398 5,488 5,409 <0.001 <0.001 0.970 32.3 
5,197 5,319 5,211 0.01 <0.001 0.933 29.7 
5,044 5,198 5,062 0.34 <0.001 0.940 6.4 
4,929 5,116 4,951 0.26 <0.001 0.946 2.3 
4,830 5,049 4,856 0.17 <0.001 0.954 2.3 
ProfilesICLikelihood testEntropySmallest class, %
AICBICaBICp (VLMR)p (BLRT)
5,930 5,987 5,936 NA NA   
5,398 5,488 5,409 <0.001 <0.001 0.970 32.3 
5,197 5,319 5,211 0.01 <0.001 0.933 29.7 
5,044 5,198 5,062 0.34 <0.001 0.940 6.4 
4,929 5,116 4,951 0.26 <0.001 0.946 2.3 
4,830 5,049 4,856 0.17 <0.001 0.954 2.3 

N = 266.

AIC, Akaike Information Criterion; BIC, Bayesian Information Criterion; aBIC, adjusted Bayesian Information Criterion; p (VLMR), p value from the Vuong-Lo-Mendell-Rubin test; p (BLRT), p value from the Bootstrapped Likelihood Ratio Test; IC, Information Criteria.

Figure 1 illustrates the way the three profiles were distinguished. The x-axis indicates eight health indicators; the y-axis indicates mean scores for each of the standardized indicators by profile. The line in the middle is zero. Across indicators, the lower the scores, the better the health status. Numerical details are shown in online supplementary Table 1.

Fig. 1.

Three health profiles of PCI; N = 266. AD8, Eight-item Interview to Differentiate Aging and Dementia; MMSE, Mini-Mental State Examination; RMBPC, Revised Memory and Behavior Problems Checklist; ADL, Activities of Daily Living; IADLs, Instrumental Activities of Daily Living.

Fig. 1.

Three health profiles of PCI; N = 266. AD8, Eight-item Interview to Differentiate Aging and Dementia; MMSE, Mini-Mental State Examination; RMBPC, Revised Memory and Behavior Problems Checklist; ADL, Activities of Daily Living; IADLs, Instrumental Activities of Daily Living.

Close modal

The three profiles were labeled as “less impaired,” “moderately impaired,” and “severely impaired.” The less impaired profile comprised 40% of PCI with below-average cognitive impairment and memory/behavior/mood problems and fewer functional limitations than others. The moderately impaired profile included 30% of PCI who reported average cognitive impairment and memory/behavior/mood problems, a below-average ADL score but a slightly above-average IADL score. The severely impaired profile represented 30% of PCI with a high level of cognitive impairment; slightly above-average memory, behavior, and mood problems; and a greater number of ADL and IADL difficulties.

Online supplementary Table 2 shows the distribution of caregiving experience and PCI and caregiver characteristics by PCI health profiles. Caregivers of less impaired PCI were more likely to report a lower level of caregiving burden and benefits, compared to severely impaired PCI and moderately impaired PCI. Older, female, and unmarried PCI and PCI with a formal diagnosis of dementia were more likely to have the severely impaired profile. Caregivers of severely impaired PCI were more likely to be highly educated and care for their PCI for more than 5 years.

PCI Health Profiles and Caregiving Burden and Benefits

Table 3 presents results from multivariable regression, testing the association of PCI health profiles with caregiving burden and benefits. Model 1 reports that the severely impaired profile, compared to the less impaired profile, was associated with a higher level of caregiving burden. Model 2 shows that the moderately impaired profile, relative to the less impaired profile, was associated with a higher level of caregiving benefits.

Table 3.

Association of PCI health profiles with caregiving burden and benefits: results from linear regression models

Model 1:Outcome:caregiving burdenModel 2:Outcome:caregiving benefit
β95% CIβ95% CI
PCI health profiles 
 Less impaired Reference Reference 
 Moderately impaired 3.39 [−0.62, 7.39] 2.34* [0.48, 4.21] 
 Severely impaired 5.43** [1.42, 9.44] 1.10 [−0.78, 2.98] 
PCI characteristics 
 Age 0.10 [−0.08, 0.28] −0.02 [−0.10, 0.06] 
 Female 1.27 [−2.16, 4.70] 1.55 [−0.23, 3.32] 
 Married 4.68* [0.31, 9.06] 0.36 [−1.50, 2.23] 
 Dementia diagnosed 7.37*** [3.88, 10.86] −0.98 [−2.49, 0.54] 
Caregiver characteristics 
 Age −0.02 [−0.17, 0.13] −0.02 [−0.10, 0.06] 
 Female 2.16 [−1.06, 5.38] −0.08 [−1.57, 1.42] 
 Ethnic minority −2.47 [−8.77, 3.83] 3.65** [1.23, 6.08] 
 Marital status −1.13 [−5.18, 2.93] 0.64 [−1.10, 2.37] 
 Education 1.72 [−0.28, 3.72] 0.45 [−0.48, 1.38] 
 Working −1.06 [−4.83, 2.70] 0.30 [−1.30, 1.90] 
 Financial adequacy −3.52*** [−5.59, −1.45] 1.10* [0.13, 2.08] 
 Poor self-rated health 3.90* [0.53, 7.27] 0.73 [−0.81, 2.27] 
 Coresident caregiver 1.83 [−2.79, 6.45] −0.96 [−3.20, 1.28] 
 Long-term caregiver, of 5 years or more −3.83 [−8.38, 0.71] −0.90 [−2.94, 1.13] 
Model 1:Outcome:caregiving burdenModel 2:Outcome:caregiving benefit
β95% CIβ95% CI
PCI health profiles 
 Less impaired Reference Reference 
 Moderately impaired 3.39 [−0.62, 7.39] 2.34* [0.48, 4.21] 
 Severely impaired 5.43** [1.42, 9.44] 1.10 [−0.78, 2.98] 
PCI characteristics 
 Age 0.10 [−0.08, 0.28] −0.02 [−0.10, 0.06] 
 Female 1.27 [−2.16, 4.70] 1.55 [−0.23, 3.32] 
 Married 4.68* [0.31, 9.06] 0.36 [−1.50, 2.23] 
 Dementia diagnosed 7.37*** [3.88, 10.86] −0.98 [−2.49, 0.54] 
Caregiver characteristics 
 Age −0.02 [−0.17, 0.13] −0.02 [−0.10, 0.06] 
 Female 2.16 [−1.06, 5.38] −0.08 [−1.57, 1.42] 
 Ethnic minority −2.47 [−8.77, 3.83] 3.65** [1.23, 6.08] 
 Marital status −1.13 [−5.18, 2.93] 0.64 [−1.10, 2.37] 
 Education 1.72 [−0.28, 3.72] 0.45 [−0.48, 1.38] 
 Working −1.06 [−4.83, 2.70] 0.30 [−1.30, 1.90] 
 Financial adequacy −3.52*** [−5.59, −1.45] 1.10* [0.13, 2.08] 
 Poor self-rated health 3.90* [0.53, 7.27] 0.73 [−0.81, 2.27] 
 Coresident caregiver 1.83 [−2.79, 6.45] −0.96 [−3.20, 1.28] 
 Long-term caregiver, of 5 years or more −3.83 [−8.38, 0.71] −0.90 [−2.94, 1.13] 

N = 266; results from 10 imputed data sets.

PCI, persons with cognitive impairment.

*p < 0.05, **p < 0.01, ***p < 0.001.

Regarding covariates, caregivers with married PCI and PCI formally diagnosed with dementia were more likely to report higher caregiving burden, whereas caregivers who reported financial adequacy were less likely to report higher caregiving burden. Caregivers of ethnic minorities and with financial adequacy were more likely to report higher caregiving benefits.

Heterogeneity in health status among PCI in the community may make it challenging to allocate healthcare resources effectively. This study provided the first systematic investigation of distinct health profiles among community-dwelling PCI and their associations with caregiving burden and benefits in Singapore.

Applying LCA to several multidimensional health indicators, this study identified three distinct PCI health profiles: less impaired (40%), moderately impaired (30%), and severely impaired (30%). The findings imply that 30% of persons with severe cognitive impairment in the community were more likely to suffer as well from other physical, functional, behavioral, and mood impairment, which calls for targeted interventions. The findings also add to the literature that categorizes older adults into different profiles using multiple health indicators [3, 28, 29]. Specifically, capturing moderately impaired PCI, this study complements the local study that identified two distinct health profiles – health at risk and relatively healthy profiles – from 2,444 community-dwelling older Singaporeans [28]. However, the profile distinction did not show a unique combination of different health indicators, such as cognitive impairment without physical difficulties [3].

This study further examined the association of PCI health profiles with burden and benefits of caregiving. In line with stress process theory, poor PCI health was related to an increased level of caregiving burden: Caregivers with less impaired PCI reported the lowest burden, while those with severely impaired PCI profile reported the highest burden. On the other hand, the relationship between PCI health and caregiving benefits was nonlinear. Compared to caregivers with less impaired PCI, caregivers with moderately impaired PCI, rather than caregivers with less impaired PCI, reported a significantly higher level of benefits. The results largely correspond to a recent study showing a nonlinear association between caregiving demands and caregiver mental well-being in Canada: a moderate level of demands benefits caregiver mental well-being, whereas a higher level of demands harms [30]. In short, this study speaks to the complex nature of caregiving experiences, which may empower but also strain family caregivers [9, 10, 11].

Overall, this study showed that the more severe the cognitive impairment, the more likely the PCI were to have memory/behavior/mood problems and functional difficulties. Yet, only about 50% of the severely impaired PCI had been formally diagnosed with dementia (online suppl. Table 2) and thus might not have been captured by healthcare services. Therefore, this study points to the need for cognitive screening of older adults in the community. In Singapore, this can be done by partnering with community organizations such as eldercare centers or local polyclinics, using a validated, brief screening tool such as AD8Plus [19].

Thereafter, tailored strategies, based on PCI health profiles, should be developed to address PCI health problems and improve caregiving experiences. For less impaired PCI and their caregivers, it is important to provide education and resources for caregivers to understand and cope with mild cognitive impairment and its progression [31]. Next, moderately impaired PCI may begin to suffer from IADL difficulties as shown in Figure 1. Interventions should aim to address not only cognitive impairment but also difficulties in IADLs such as preparing meals and taking medications. Supporting this group is essential for long-term and sustainable caregiving as caregivers in this group are more likely to experience caregiving benefits. Last but not least, interventions should focus on the severely impaired PCI with behavioral problems and ADL difficulties. These may include community-based support programs involving physicians, nurses, and social workers who routinely visit PCI and their caregivers in the community to provide curated care [32]. In addition, intensive support, such as peer support groups, cognitive behavioral therapy, and psychoeducational programs, should be offered to caregivers who are at risk for physical and mental health decline, as well as social isolation. Identifying and addressing the specific care needs of severely impaired PCI may also help alleviate caregiver strain [18].

The study had some limitations. First, data were collected from a specific community in Singapore with an above-average proportion of older adults and may not be representative of the national population, as well as other country contexts. Second, the data were cross-sectional and did not address causal research questions. Third, while the study addressed multiple dimensions of PCI health, it did not cover all dimensions of PCI health due to a lack of relevant measures.

As the incidence of cognitive impairment in the community grows hand in hand with population aging, it will be increasingly important to identify health profiles of PCIs in the community so that limited healthcare and social care resources can be appropriately targeted at those who need it most. Future studies should continue to investigate distinct profiles of PCI health in the community and their implications for caregiving experiences in different demographic and cultural contexts.

The Caring for persons with dementia and their caregivers in the community: Towards a sustainable community based dementia care system (COGNITION) study received approval from the Institutional Review Board at the National University of Singapore (approval number H-17-013). Written informed consent was obtained from both persons with cognitive impairment (PCI) and their caregivers in the COGNITION study prior to survey administration. If PCI were unable to respond due to health reasons, consent was obtained from PCI’s legal representative or next of kin.

The authors have no conflicts of interest to declare.

Caring for persons with dementia and their caregivers in the community: Towards a sustainable community based dementia care system (COGNITION) study was supported by the National Innovation Challenge on Active and Confident Ageing Grant (award No.: MOH/NIC/COG05/2017). The funding organizations had no role in the design or conduct of the study; collection, management, analysis, or interpretation of the results; or preparation or approval of the manuscript.

Pildoo Sung designed the study, conducted data analysis, and wrote the manuscript. Jeremy Lim-Soh contributed to the writing and critically reviewed the manuscript. Angelique Chan, the principal investigator of the COGNITION study, managed the project and collected the data.

The complete COGNITION study dataset is not publicly available due to legal and ethical reasons. The data used in this study are available upon reasonable request. Further inquiries can be directed to the corresponding author.

1.
Gauthier
S
,
Webster
C
,
Servaes
S
,
Morais
JA
,
Rosa-Neto
P
World alzheimer report 2022: life after diagnosis–navigating treatment, care and support
London, England
Alzheimer’s Disease International
2022
.
2.
Morrisby
C
,
Joosten
A
,
Ciccarelli
M
.
Do services meet the needs of people with dementia and carers living in the community? A scoping review of the international literature
.
Int Psychogeriatr
.
2018
;
30
(
1
):
5
14
.
3.
Lafortune
L
,
Béland
F
,
Bergman
H
,
Ankri
J
.
Health status transitions in community-living elderly with complex care needs: a latent class approach
.
BMC Geriatr
.
2009
;
9
(
1
):
6
.
4.
Pinquart
M
,
Sörensen
S
.
Associations of stressors and uplifts of caregiving with caregiver burden and depressive mood: a meta-analysis
.
J Gerontol B Psychol Sci Soc Sci
.
2003
;
58
(
2
):
112
28
.
5.
Schulz
R
,
Beach
SR
,
Czaja
SJ
,
Martire
LM
,
Monin
JK
.
Family caregiving for older adults
.
Annu Rev Psychol
.
2020
;
71
(
1
):
635
59
.
6.
Pearlin
LI
,
Mullan
JT
,
Semple
SJ
,
Skaff
MM
.
Caregiving and the stress process: an overview of concepts and their measures
.
Gerontologist
.
1990
;
30
(
5
):
583
94
.
7.
Son
J
,
Erno
A
,
Shea
DG
,
Femia
EE
,
Zarit
SH
,
Stephens
MAP
.
The caregiver stress process and health outcomes
.
J Aging Health
.
2007
;
19
(
6
):
871
87
.
8.
Kim
H
,
Chang
M
,
Rose
K
,
Kim
S
.
Predictors of caregiver burden in caregivers of individuals with dementia
.
J Adv Nurs
.
2012
;
68
(
4
):
846
55
.
9.
Adelman
RD
,
Tmanova
LL
,
Delgado
D
,
Dion
S
,
Lachs
MS
.
Caregiver burden: a clinical review
.
JAMA
.
2014
;
311
(
10
):
1052
60
.
10.
Tarlow
BJ
,
Wisniewski
SR
,
Belle
SH
,
Rubert
M
,
Ory
MG
,
Gallagher-Thompson
D
.
Positive aspects of caregiving: contributions of the REACH project to the development of new measures for Alzheimer’s caregiving
.
Res Aging
.
2004
;
26
(
4
):
429
53
.
11.
Roth
DL
,
Fredman
L
,
Haley
WE
.
Informal caregiving and its impact on health: a reappraisal from population-based studies
.
Gerontologist
.
2015
;
55
(
2
):
309
19
.
12.
Sung
P
,
Goh
VS
,
Azman
ND
,
Visaria
A
,
Malhotra
R
.
Types of caregiving experience and their association with caregiver depressive symptoms and quality of life
.
J Aging Health
.
2022
34
4–5
591
601
.
13.
Li
Q
,
Loke
AY
.
The positive aspects of caregiving for cancer patients: a critical review of the literature and directions for future research
.
Psychooncology
.
2013
;
22
(
11
):
2399
407
.
14.
Mehri
N
,
Kinney
J
,
Brown
S
,
Rajabi Rostami
M
.
Informal caregiving and all-cause mortality: a meta-analysis of longitudinal population-based studies
.
J Appl Gerontol
.
2021
;
40
(
7
):
781
91
.
15.
Malhotra
R
,
Bautista
MAC
,
Müller
AM
,
Aw
S
,
Koh
GCH
,
Theng
Y-L
.
The aging of a young nation: population aging in Singapore
.
Gerontologist
.
2018
;
59
(
3
):
401
10
.
16.
Subramaniam
M
,
Chong
SA
,
Vaingankar
JA
,
Abdin
E
,
Chua
BY
,
Chua
HC
.
Prevalence of dementia in people aged 60 years and above: results from the WiSE study
.
J Alzheimers Dis
.
2015
;
45
(
4
):
1127
38
.
17.
Lee
Y-R
,
Sung
K-T
.
Cultural influences on caregiving burden: cases of Koreans and Americans
.
Int J Aging Hum Dev
.
1998
;
46
(
2
):
125
41
.
18.
Sung
P
,
Chan
AWM
.
Care needs profiles and their association with caregiver strain among community-dwelling older adults with cognitive impairment
.
Dement Geriatr Cogn Disord
.
2022
;
51
(
4
):
340
7
.
19.
Tew
C-W
,
Ng
T-P
,
Cheong
C-Y
,
Yap
P
.
A brief dementia test with subjective and objective measures
.
Dement Geriatr Cogn Dis Extra
.
2015
;
5
(
3
):
341
9
.
20.
Creavin
ST
,
Wisniewski
S
,
Noel-Storr
AH
,
Trevelyan
CM
,
Hampton
T
,
Rayment
D
.
Mini-Mental State Examination (MMSE) for the detection of dementia in clinically unevaluated people aged 65 and over in community and primary care populations
.
Cochrane Database Syst Rev
.
2016
2016
1
CD011145
.
21.
Teri
L
,
Truax
P
,
Logsdon
R
,
Uomoto
J
,
Zarit
S
,
Vitaliano
PP
.
Assessment of behavioral problems in dementia: the revised memory and behavior problems checklist
.
Psychol Aging
.
1992
;
7
(
4
):
622
31
.
22.
Zarit
SH
,
Reever
KE
,
Bach-Peterson
J
.
Relatives of the impaired elderly: correlates of feelings of burden
.
Gerontologist
.
1980
;
20
(
6
):
649
55
.
23.
Siow
JYM
,
Chan
A
,
Østbye
T
,
Cheng
GH-L
,
Malhotra
R
.
Validity and reliability of the Positive Aspects of Caregiving (PAC) scale and development of its shorter version (S-PAC) among family caregivers of older adults
.
Gerontologist
.
2017
;
57
(
4
):
e75
84
.
24.
Seng
BK
,
Luo
N
,
Ng
WY
,
Lim
J
,
Chionh
HL
,
Goh
J
.
Validity and reliability of the Zarit Burden Interview in assessing caregiving burden
.
Ann Acad Med Singap
.
2010
;
39
(
10
):
758
63
.
25.
Spurk
D
,
Hirschi
A
,
Wang
M
,
Valero
D
,
Kauffeld
S
.
Latent profile analysis: a review and “how to” guide of its application within vocational behavior research
.
J Vocat Behav
.
2020
;
120
:
103445
.
26.
Nylund-Gibson
K
,
Choi
AY
.
Ten frequently asked questions about latent class analysis
.
Transl Issues Psychol Sci
.
2018
;
4
(
4
):
440
61
.
27.
Morris
TP
,
White
IR
,
Royston
P
.
Tuning multiple imputation by predictive mean matching and local residual draws
.
BMC Med Res Methodol
.
2014
;
14
(
1
):
75
13
.
28.
Ng
CWL
,
Luo
N
,
Heng
BH
.
Health status profiles in community-dwelling elderly using self-reported health indicators: a latent class analysis
.
Qual Life Res
.
2014
;
23
(
10
):
2889
98
.
29.
Yap
KH
,
Warren
N
,
Allotey
P
,
Reidpath
DD
.
Chronic disease profiles of subjective memory complaints: a latent class analysis of older people in a rural Malaysian community
.
Aging Ment Health
.
2020
;
24
(
5
):
709
16
.
30.
Bierman
A
,
Lee
Y
,
Penning
MJ
.
Mental health benefits and detriments of caregiving demands: a nonlinear association in the Canadian longitudinal study on aging
.
J Aging Health
.
2022
089826432211252
.
31.
Gauthier
S
,
Reisberg
B
,
Zaudig
M
,
Petersen
RC
,
Ritchie
K
,
Broich
K
.
Mild cognitive impairment
.
Lancet
.
2006
;
367
(
9518
):
1262
70
.
32.
Backhouse
A
,
Ukoumunne
OC
,
Richards
DA
,
McCabe
R
,
Watkins
R
,
Dickens
C
.
The effectiveness of community-based coordinating interventions in dementia care: a meta-analysis and subgroup analysis of intervention components
.
BMC Health Serv Res
.
2017
;
17
(
1
):
717
.