The Societal Cost of Dementia in Singapore: Results from the WiSE Study

Abstract

Background:

There is currently limited evidence on the economic burden that dementia exerts on multi-ethnic Asian populations.

Objective:

The present study aimed to estimate the economic cost of dementia in Singapore.

Methods:

We used data from the Well-being of the Singapore Elderly study, a nationally representative survey of the older Singapore Resident population aged 60 years and above. Generalized linear modeling was used to estimate factors associated with costs.

Results:

The total cost of dementia in 2013 was estimated at S$532 million (95% CI, S$361 million to S$701 million) while the annual cost per person was estimated at S$10,245 per year (95% CI, S$6,954 to S$12,495). Apart from dementia, higher total societal cost were also significantly associated with older age, Indian ethnicity, and those who were diagnosed with heart problems, stroke, diabetes or depression, whereas being divorced/separated, lower education, and those who were diagnosed with hypertension were significantly associated with lower total societal cost.

Conclusion:

The study provides a rich body of information on healthcare utilization and cost of dementia, which is essential for future planning of services for the elderly population.

Keywords

Cost of illness dementia elderly health services societal cost

INTRODUCTION

Dementia is a clinical syndrome caused by a number of progressive organic brain disorders that affect memory, thinking, behavior, and the ability to perform everyday activities [1]. It is well established that dementia is a leading contributor to disability and need for care among older people [2] resulting in increased cost to society [3]. As a result of the “greying of the world” and the increasing number of people suffering from dementia, cost of care of people with dementia has an enormous economic impact on the health care and social services systems [4], thus significantly affecting every health system in the world [2].

According to Alzheimer’s Disease International (ADI), the number of people living with dementia worldwide in 2013 was estimated at 44.3 million, and is estimated to reach 75.6 million by 2030 and 135.5 million by 2050 [1]. The total worldwide cost of dementia, including medical expenditures and informal care costs, was estimated at US$604 billion in 2010. The authors suggested that if all potential background factors remained unchanged, and one only considers the increase in the number of people with dementia forecast, this figure will increase by 85% by 2030 [2]. Wimo and colleagues argued that this amount corresponds to 1% of the world’s gross domestic product (GDP) and mentioned that if dementia care was a country, it would be the world’s 21st largest economy, ranking between Poland and Saudi Arabia [2].

A number of cost-of-illness (COI) studies have been conducted to estimate the economic impact of dementia to governments, health, and social care systems. COI studies quantify the economic value of resources lost due to the disease or consumed in its prevention, treatment, and care [5]. In the United Kingdom, the findings of research commissioned by the Alzheimer’s Society estimated the annual societal cost of dementia to be £26.3 billion while the annual per-capita cost was estimated to be £32,250 [6]. In the United States, the total cost of dementia in 2010 was estimated at US$215 billion while the annual cost per person was estimated to be US$56,290 [7]. For the Asia Pacific region, it is estimated that the total costs associated with dementia are US$185 billion [8]. These figures are likely to increase as the number of people with dementia grow, and societal demands in countries with emerging economies and large populations in the region including India, China, and Indonesia continue to mount [1].

Singapore is an island city-state in South-East Asia, with a multi-ethnic population of about 5.5 million in 2014, of which 3.9 million were Singapore citizens and Permanent Residents. The resident population comprises Chinese (74.3%), Malays (13.3%), Indians (9.1%), and other ethnic groups (3.3%) [9]. With increasing life expectancy and low fertility rates, the Singapore resident population continues to age. There are more citizens in the older age groups today as the ‘post-war baby boomers’ enter their silver years. The proportion of citizens aged 65 years and above increased from 11.7% in 2013 to 12.4% in2014 [10].

Singapore’s healthcare philosophy comprises a dual healthcare delivery system with a public and private sector where affordability of healthcare is ensured with the 3Ms: Medisave, a national mandatory healthcare saving plan; Medishield, a national low cost medical insurance scheme for catastrophic illness; and Medifund, where subsidies are provided for needy Singaporeans through a national fund. There is also a special financing system scheme for the elderly population, Eldershield, which provides basic financial protection to those who need long-term care. The healthcare philosophy places a significant emphasis on individual responsibility and the need for co-payment for services provided. General Practitioners (GPs) provide 80% of the primary healthcare services, and doctors in government polyclinics provide the remaining 20%. Public hospitals (referred to as restructured hospitals) provide about 80% of the tertiary care in Singapore, while the remaining care is provided by private hospitals. Patients seeking treatment in public hospitals may apply for a range of subsidies on their total bill; the extent of subsidy received is subjected to guidelines set by the government to allocate limited resources to those who need them most.

There is currently limited evidence on the economic burden that dementia exerts on this society. The present study aimed to estimate the economic cost of dementia in Singapore.

METHODS

Study design

We used data from the Well-being of the Singapore Elderly (WiSE) study. The WiSE study was a nationally representative survey of the older Singapore Resident population aged 60 years and above. The sample was derived using a disproportionate stratified sampling design. We oversampled residents aged 75 years and above, and those of Malay and Indian ethnicity in order to ensure that sufficient sample size for these population subgroups could be achieved to improve the reliability of our estimates. The study was approved by the institutional ethics review boards of participating institutions, which include the National Healthcare Group, Domain Specific Review Board (DSRB), and the SingHealth Centralised Institutional Review Board (CIRB). Written informed consent was obtained from all respondents. In the case where respondents were unable to provide informed consent, written informed consent was taken from their legally acceptable representative/next of kin. The respondents were randomly selected from a national registry that maintains the names, other socio-demographic details such as age, gender, and ethnicity, and household addresses of all residents in Singapore. The respondents were approached at the household address provided by the registry. Face-to-face interviews were conducted by professional survey interviewers using an online Computer Assisted Personal Interviewing (CAPI) application and data were collected in real-time. The study method has been described in detail elsewhere [11].

Measures

Dementia diagnosis was established according to 10/66 criteria [12]. The 10/66 dementia diagnosis algorithm [13] requires: a structured clinical mental state interview, the Geriatric Mental State (GMS), which applies the Automated Geriatric Examination for Computer Assisted Taxonomy (AGECAT) computer algorithm [14]; a cognitive test battery comprising the Community Screening Instrument for Dementia (CSI’D’) with the global cognitive score (COGSCORE) [15] and the modified Consortium to Establish a Registry for Alzheimer’s Dementia (CERAD) 10 word list learning task with delayed recall [16] and; an informant interview, the CSI’D’ with informant score (RELSCORE) [15] for evidence of cognitive and functional decline. The dementia diagnosis was given to respondents scoring above a cut-point of predicted probability of dementia derived from the logistic regression equation developed in the 10/66 international pilot study, using coefficients from the GMS, CSI-D, and 10 word list learning tasks [13]. Its algorithm has been cross-culturally validated elsewhere [12 , 17] and in our sample [11]. Data on other health conditions including hypertension, transient ischemic attacks (TIA), depression, heart problems (myocardial infarction, cardiac failure, and valvular heart disease), stroke and diabetes as well as socio-demographic data including age, gender, ethnicity, marital status, education, and employment were also collected.

Resource utilization and costing

Information on resource utilization was obtained from respondents and their informants using an adapted version of the Client Service Receipt Inventory (CSRI) [18]. The informant was defined as ‘a person who knows the older person best’. The instrument asked whether respondents had used specific services during the 3-month period prior to the interview. In addition to this, information on informal care arrangements (in terms of time spent with the respondents in assisting them in daily activities), paid care, and change in work status due to care provision was also collected. The costs were calculated from a broad societal perspective that consists of healthcare and social care costs.

Healthcare costs

Four components of healthcare cost derived from direct medical care, intermediate and long-term care (ILTC), traveling related to medical appointments or care, and indirect medical care were included in the analysis. Direct medical care refers to care provided by health professionals in the public or private sector, outpatient, or inpatient setting. This included care from the government primary care doctors (polyclinic doctor), restructured hospital doctors, other restructured hospital health workers (e.g., physiotherapist, nurse, medical social workers (MSWs)), private hospital/clinic doctors, other private health workers (e.g. physiotherapist, nurse), government and private dentists, traditional healers, admissions to government/restructured or private hospitals (hospital admissions), accident & emergency (A&E) department visits, use of medication, and ILTC (e.g.,care provided in day care centers, respite care, and nursing homes). The indirect medical care cost was calculated and included the respondent’s traveling time in order to receive the services (traveling) and time spent by family members or friends in accompanying respondents while using the services (accompanying).

Generally, we estimated the specific health care costs by multiplying each service unit (i.e., consultations per minute, visits per day) by their own unit cost price. For the estimation of annual costs, the 3-month values were multiplied by four. Due to variations and paucity of data relating to unit cost from local sources, we used an alternative approach, i.e., extrapolation through application of UK unit costs was employed [19], to estimate the unit cost for selected direct medical care (consultations with the primary care doctor, restructured hospital doctor, and other restructured hospital health workers) and ILTC services (day care center, respite care, and nursing home). We assumed that the relationship between UK and Singapore unit costs for these services were fixed and the ratio of costs between two countries has remained unchanged over the years [19]. This approach consisted of the following steps: a) determination of the reference unit costs (UK) for each specific service, b) generation of ratios for inpatient and outpatient settings between the reference country (UK) and Singapore using data from the WHO-CHOICE (WHO-CHOosing Interventions that are Cost-Effective) database [20], and c) application of these ratios to the unit cost of each selected services in the reference country in order to generate country-specific unit costs for Singapore (Table 1). Reliable sources of UK unit costs as a reference country, such as the Unit Cost of Health and Social Care 2013 [21] and the NHS Reference Costs 2012-13 publications [22] were used to identify and match the appropriate unit costs to the reported services. These publications aim to provide robust information about the costs of key health and social care services provided in the UK. For other direct medical care including private health care doctors, other private health care workers, dentists, traditional healers, A&E, and medication; the average out-of-pocket reported expenses amount was used instead of applying the ratios to the UK unit cost as they were deemed more representative of the Singapore population. For hospital admissions, we used the unit cost per bed day according to the WHO-CHOICE database based on 2008 prices. The figures were converted into local currency units (S$) and inflated to 2013 prices using the Consumer Price Index (CPI) (Department of Statistics, 2013) [23]. The average national wage rates according to the International Labor Organization and the Global Wage Database figures [24] were used to calculate the indirect cost of carers’ time spent in accompanying the respondents to utilize the services, accounting only for traveling time. The latest available information on the mean monthly wages per gender was in 2011 prices and therefore they were converted in 2013 prices using the CPI for Singapore [23].

Social care services

Social care included unpaid care (informal care) and paid care. Informal care refers to care provided by family members or friends that can be further classified into help provided with activities of daily living (ADL), instrumental activities of daily living (IADL), and supervision [25]. The former includes help with dressing, feeding, looking after one’s appearance, toileting, and bathing. IADL involves help with communication and transportation for the past 24 hours, prior to the interview. It is assumed that this kind of care is provided every day during the 3-month period. Paid care was classified into care during the day on an occasional, regular or constant basis and care during the night on a “sleep-in” or “waking” basis. It is assumed that occasional, regular, and constant paid care was provided in 1, 4, and 7 days per week, respectively. Similarly, it is assumed that paid care at night is 8 hours, 7 days per week for those who need it. The opportunity cost approach was used as the primary method of valuing unpaid (informal) care. This approach estimates the value of the informal caregiver’s wages forgone as a result of spending time on providing informal care. Under this method, the average national wages by occupation levels in Singapore were used as a proxy for valuing care provided by family members or friends. To account for the fact that many caregivers are elderly and out of work force, we imputed their wages using average wages from employed persons with similar demographic characteristics such as age, gender and education and then scaled down the imputed wages by multiplying with labor force participation rate in the same demographic groups. This approach assumes that many caregivers would not work even if they were not providing caregiving services. This method has been used previously [7]. For paid care, due to lack of information on the amount of money paid for home care from the survey, the average per hour wage of a semi-skilled worker was used as unit cost.

Statistical analysis

Statistical analyses were carried out using the SAS software version 9.2 (SAS Institute Inc., Cary, NC, USA) and STATA version 13.0. To ensure that the survey findings were representative of the Singapore elderly population, all estimates were analyzed using survey weights to adjust for oversampling, non-response, and poststratification according to age and ethnicity of the Singapore elderly population. Considering the data as mostly complete, more sophisticated methods such as multiple imputations were not applied. Missing data in categorical and continuous variables were imputed using the last value carried forward and mean imputation methods, respectively. Mean and standard errors were calculated for continuous variables, and frequencies and percentages for categorical variables. To estimate the annual total societal costs of dementia, we estimated the increase in cost attributable to dementia or excess costs over and above the base rate using a series of regression analyses. The base rate consists of the costs that are generated annually by every person, on average [26]. In the first analysis, the total societal cost (as the dependent variable) was regressed on dementia without controlling for covariates. In the second analysis, the effect of dementia on the total societal cost was controlled for socio-demographic variables including age, gender, ethnicity, marital status, education, and employment; and other chronic conditions including hypertension, TIA, depression, heart problems (myocardial infarction, cardiac failure, and valvular heart disease), stroke, and diabetes. We have repeated these analyses for all other cost components (direct medical care, ILTC, traveling, indirect care, total healthcare, ADL, IADL, unpaid care, paid care, total healthcare, and total social care). Given that the distribution of costs in our sample was skewed (asymmetrically distributed) with many zeros and both assumptions of normality and homoscedasticity of residuals were violated, a generalized linear modeling (GLM) was used to model the effects of dementia on costs after controlling for covariates. To use GLMs, the distribution of cost data and a function that specifies the relationship between the mean and the linear specification of the covariates (the so called link function) must be specified [27]. As recommended by Manning and Mullahy [28], the appropriateness of alternative error distributions (Gaussian, Poisson, gamma, or inverse Gaussian) was examined using modified Park’s test [29, 30]. Results of 11 costs components supported the use of GLM model with gamma family and log link, another one component did not converge with the GLM model. Therefore, two-part model with a logistic model in the first part and GLM model with log-link and gamma distribution in the second part were used to estimate the adjusted costs for this component after controlling for socio-demographic variables and other chronic conditions. All statistically significant differences were evaluated at the 0.05 level using 2-sided tests.

RESULTS

Socio-demographic characteristics of the sample

A total of 2,565 respondents completed the study giving a response rate of 65.6% . Informants of 2,421 respondents completed the requisite interviews and data of these respondents and informants were included in this study. The sample comprised 57% female and 43% male respondents. The majority of the respondents were aged between 60–74 years (74.8%), of Chinese ethnicity (82.6%), and currently married (65.4%). The prevalence of dementia in this study was 10% (Table 2).

Annual contacts and costs per person with dementia

Table 3 gives the annual contacts and costs of individual components per person among the total sample and by dementia group. Overall, the most commonly utilized direct medical care was payments for medication (82%) followed by primary care visits (polyclinic doctor) (43.1%), with an average number of 2.5 contacts per year. Approximately 29% had consulted restructured hospital doctors, and 26.5% had consulted private hospital/clinic doctors. For social care component, the most commonly utilized care component of informal care was paid care during daytime (7.9%), with mean contact of 22.9 days. The costs of hospital admissions were the largest cost drivers in those with dementia (S$7,178). We found that the dementia cases seemed to incur higher costs in all the components than non-dementia cases except for private hospital/clinic, dentist, and traditional healer (Table 3). The annual healthcare costs and the social care services costs per person attributable to dementia with and without adjustment for socio-demographic and comorbid conditions are shown in Table 4. We found that the unadjusted total annual societal care costs per person attributable to dementia were S$27,331 (95% CI, S$20,200 to S$34,461). That is, those diagnosed with dementia would be expected to incur S$27,331 more in total annual societal care costs than those without dementia. After adjustment for socio-demographic and other comorbid health conditions, the annual total costs of dementia were S$10,245 (95% CI, S$6,954 to S$12,495). The total annual cost of dementia in Singapore among elderly aged 60 years and over is estimated at S$532 million and range from S$361 million to S$701 million per year (Table 4). In this multivariate model, other covariates that are significantly associated with higher total societal cost include those aged 75–84 years and 85 years and over (versus aged 60 to 75 years), Indian and Other ethnicity (versus Chinese), and those who reported that a doctor had ever diagnosed them with heart problems, stroke, diabetes or depression, whereas being divorced/separated (versus married/cohabiting), did not complete primary or completed primary education (versus tertiary), and those with hypertension were associated with lower total societal cost. There was a significant positive interaction effect of dementia and hypertension on total societal cost (Table 5).

DISCUSSION

This is the first study to provide a comprehensive estimate of the societal cost attributed to dementia in the Singapore elderly population based on a nationally representative sample. We found that the total cost of dementia for the population of Singapore was S$532 million (US$409 million, using 2013 exchange rate of 1US$ = 1.3S$) per year, suggesting a significant economic burden. Our total annual societal costs estimates were considerably lower than those projected by the ADI which has estimated that the total costs of dementia were US$1.7 billion [8]. However, it should be borne in mind that direct comparisons to the previous estimates is not quite appropriate due to the differing methods employed by researchers. For example, the cost estimated by the ADI was not adjusted for the cost of coexisting conditions. If we had estimated costs without adjusting for covariates, our total cost of dementia would be tripled to S$1.4 billion per year (US$1.3 billion, using 2013 exchange rate of 1US$ = 1.3S$) and would be slightly lower than that estimated by ADI [8]. However, if interaction between dementia and other chronic conditions was allowed in multivariate analysis, it was found that the adjusted total cost of dementia alone would be much lower (S$397 billion per year = US$305 million, using 2013 exchange rate of 1US$ = 1.3S$). The current study also differed from the ADI study in terms of diagnostic criterion for dementia and the range of cost components used. The prevalence of dementia used in ADI report was derived from projected population figures from different reports and analyses [8]. For example, according to ADI report in 2006, the dementia prevalence was estimated based on age-gender rates using meta-analysis together with United Nations demographic data [31] whereas our results were based on a single national representative sample. Thus, comparisons of our estimates with previous studies should be treated with considerable caution.

We found that the annual cost of dementia per person in 2013 was estimated at S$10,245 (US$7,881, using 2013 exchange rate of 1US$ = 1.3S$). Our cost estimate was within the range of worldwide estimated cost of dementia; from US$6,827 in upper middle income countries to US$32,869 in countries with high incomes based on 42 studies reviewed worldwide [2]. Our estimate was slightly lower than the mean worldwide cost of dementia (US$16,986) although variations have been observed across countries [2]. It was also lower than other Western studies. For example, Jonsson and Wimo [3] reviewed 16 studies in European countries and reported that the median estimate of total annual cost per person with dementia was estimated to be € 28,000 [3]. In the United Kingdom, a recent report commissioned by the Alzheimer’s Society for Dementia UK revealed that the average annual cost of dementia was £32,250 per person [6]. Meanwhile in the United States, the total annual cost per person was US$56,290 and US$42,689 depending on the method used to value informal care [7].

We found that the total costs were split unequally between healthcare and social care costs. The main costs driver was from the social care component, which contributed 76% of the societal costs. The costs of unpaid care represented 60% of social care costs. This trend was similarly found in low- and middle-income countries where the proportion of informal care costs tends to pre-dominate the total costs [2]. This trend was also noted previously in the UK where even though the direct health (£4.9 billion) and social care costs (£11.2 billion) of dementia were high, they were dwarfed by the indirect costs associated with unpaid care and support provided by family members and other carers (£11.6 billion) [6]. The finding that the main cost driver was the social care component is also consistent with the findings reported from other high income countries. Wimo et al. [2] reported that in high income countries, the direct costs of social care accounted for nearly half of all costs. In the study by Hurd et al. [7] in United States, the main component of the cost attributable to dementia was the cost for institutional and home-based long-term care rather than the cost of medical services; the sum of the costs for nursing home care and formal and informal home care represented 75% to 84% of attributable costs.

The findings of the present study suggest that informal care such as care provided by family members or other carers in the community remains an important component of the dementia care system in Singapore. As reported by Konig et al. [32], the cost of care for dementia patients living in the community tends to be higher than for patients living in nursing homes. They found the mean annual costs of care for community-dwelling patients was higher than for patients living in nursing homes and the costs increased substantially if time spent on supervision of patients was included in the analysis. One reason for care provided in the community being more expensive is that it is rendered on an individual basis whereas in nursing homes caregivers care for more than one patient simultaneously. However, this results in a significant burden of care on family and friends, imposing significant costs on caregivers, leading to concerns about the sustainability of that care in the long term [33]. As such, increasing empowerment of the caregivers may be an attractive alternative. For example, in United States, the North Dakota Dementia Care Services Program, whose primary goal was to increase unpaid caregivers’ competence and confidence, indicated that program participation could accomplish the additional goal of prolonging care at home and reducing the use of potentially avoidable medical services, resulting in significant savings even after program costs were accounted for [34].

Apart from dementia, higher total societal cost was also significantly associated with heart problems, stroke, diabetes, and depression. However, we found that the incremental costs attributable to dementia at the individual-level exceed those of common mental disorders. Dementia costs more than two times as much as heart problems, stroke, diabetes, and depression. These findings are consistent with results from other international population studies in low-income and middle-income countries that found that dementia was consistently the leading contributor to disability [35] and dependence [36]. The results of this study also found that those of older age, Indian and other ethnicity were significantly associated with increased societal costs, whereas being divorced/separated, lower education (some, but did not complete primary or completed primary), and hypertension were significantly associated with decreased total societal cost. We also found that there was a significant positive interaction effect of hypertension and dementia on total societal costs, indicating that the average total costs among those who had both dementia and hypertension were significantly higher than those without this condition. This is an important finding from a public health perspective. It shows that effective intervention in older people is complicated by ageism, complex multimorbidity, exacerbated by user fees, inadequate income security, and social protection [37].

Our study has several limitations. First, the study had a 34.4% non-response which may have affected the generalization of the study findings if no effort had been made to reduce the bias. Recent studies have advocated the use of non-response adjustment weight when dealing with unit non-response in survey data analysis [38, 39]. We also noted that there is a large effect of weighting due to disproportionate sampling design employed in this study. According to Tracy & Carkin [39], the disproportionate stratified sampling is significantly associated with design effects; therefore, sample data must be weighted to remedy the design effects and/or possible selection effects due to disproportionate sampling. Therefore, in order to control for oversampling as well as non-response biases, we have computed survey weights that incorporate sampling weight, non-response weight, and post-stratification weight, which was then used to weight the sample back to the population to ensure that the survey findings were representative of the Singapore elderly population. This approach has been recommended when analyzing complex survey data [40]. We also would like to highlight that this response rate is considered acceptable for psychiatric epidemiology survey that involve lengthy questionnaires and interviews. Second, the cost calculations were based on self-reported service utilization, which may underestimate the service use and the corresponding costs. The recall period of 3-month was longer compared to other studies. However, the CSRI is a well-validated scale that has been widely used in 10/66 population-based studies in the elderly population. A previous study has shown that the scale had good agreement with GP case records for reporting of number of GP visits [41]. Third, several assumptions were made while calculating the unit costs of healthcare and social care which was largely done due to lack of information on the unit costs in Singapore. However, this is a widely accepted practice used for performing economic cost estimations [19 , 43].

Our study has several strengths including a large sample size, the inclusion of healthcare and social care costs, as well as the use of well-validated diagnostic instruments to diagnose dementia. In the present study, we used a bottom-up approach to estimate the costs attributable to dementia. Bottom-up methods have the advantage of providing greater detail in relation to specific cost elements, and the same study can be extended to capture further information.

This study provides evidence of the significant societal cost of dementia. The study also provides a rich body of information on the health services and cost of dementia that is essential for policy analysis and formulation as well as the basis for future planning for services of elderly population in Singapore.

Footnotes

ACKNOWLEDGMENTS

The study was funded by the Ministry of Health, Singapore and the Singapore Millennium Foundation of the Temasek Trust.

Authors’ disclosures available online ().

References

Alzheimer’s Disease International (2014) Dementia and risk reduction: An analysis of protective and modifiable factors, World Alzheimer Report. London, Alzheimer’s Disease International.

Wimo

, Jonsson

, Bond

, Prince

, Winblad

(2013) The worldwide economic impact of dementia 2010. Alzheimers Dement 9, 1–11.

Jonsson

, Wimo

(2009) The cost of dementia in Europe: A review of the evidence, and methodological considerations. Pharmacoeconomics 27, 391–403.

Wimo

, Winblad

, Jonsson

(2007) An estimate of the total worldwide societal costs of dementia in 2005. Alzheimers Dement 3, 81–91.

Rice

(1966) Department of Health, Education, and Welfare. Estimating the Cost-of-Illness, Public Health Service. Washington, DC.

Alzheimer’s Disease International (2014) Second edition - Overview. World Alzheimer Report. Dementia UK, London, Alzheimer’s Disease International.

Hurd

, Martorell

, Langa

(2013) Monetary costs of dementia in the United States. N Engl J Med 369, 489–490.

Alzheimer’s Disease International (2014) World Alzheimer Report. Dementia in the Asia Pacific Region. London, Alzheimer’s Disease International.

Department of Statistics (2014) Yearbook of Statistics Singapore, 2014, Department of Statistics, Ministry of Trade & Industry, Republic of Singapore.

10.

Population in, Brief (2014) National Population and Talent Division, Prime Minister’s Office, Singapore, http://www.nptd.gov.sg/portals/0/homepage/highlights/population-in-brief-2014.pdf. Accessed on February 29, 2015.

11.

Subramaniam

, Chong

, Vaingankar

, Abdin

, Chua

, Eng

, Heng

, Hia

, Huang

, Jeyagurunathana

, Kua

, Lee

, Mahendran

, Magadi

, Malladi

, McCrone

, Pang

, Picco

, Sagayadevan

, Sambasivam

, Seng

, Seow

, Shafie

, Shahwan

, Tan

, Yap

, Zhang

, Ng

, Prince

(2015) Prevalence of dementia in people aged 60 years and above: Results from the WiSE study. J Alzheimers Dis 45, 1127–1138.

12.

Prince

, Ferri

, Acosta

, Albanese

, Arizaga

, Dewey

, Gavrilova

, Guerra

, Huang

, Jacob

, Krishnamoorthy

, McKeigue

, Rodriguez

, Salas

, Sosa

, Sousa

, Stewart

, Uwakwe

(2007) The protocols for the 10/66 dementia research group population-based research programme. BMC Public Health 7, 165.

13.

Prince

, Acosta

, Chiu

, Scazufca

, Varghese

(2003) Dementia Research Group. Dementia diagnosis in developing countries: A cross-cultural validation study. Lancet 361, 909–917.

14.

Copeland

, Dewey

, Griffiths-Jones

(1986) A computerized psychiatric diagnostic system and case nomenclature for elderly subjects: GMS and AGECAT. Psychol Med 16, 89–99.

15.

Hall

, Hendrie

, Brittain

, Norton

(1993) The development of a dementia screening interview in two distinct languages. Int J Methods Psychiatry Res 3, 1–28.

16.

Ganguli

, Chandra

, Gilby

, Ratcliff

, Sharma

, Pandav

, Seaberg

, Belle

(1996) Cognitive test performance in a community-based nondemented elderly sample in rural India: The Indo-U.S. Cross-National Dementia Epidemiology Study. Int Psychogeriatr 8, 507–524.

17.

Prince

, de Rodriguez

, Noriega

, Lopez

, Acosta

, Albanese

, Arizaga

, Copeland

, Dewey

, Ferri

, Guerra

, Huang

, Jacob

, Krishnamoorthy

, McKeigue

, Sousa

, Stewart

, Salas

, Sosa

, Uwakwa

, 10/66 Dementia research group (2008) The 10/66 Dementia Research Group’s fully operationalised DSM-IV dementia computerized diagnostic algorithm, compared with the 10/66 dementia algorithm and a clinician diagnosis: A population validation study. BMC Public Health 8, 219.

18.

Beecham

, Knapp

(1992) Costing psychiatric interventions. Measuring Mental Health Needs. Thornicroft

, Brewin

, Wing

, eds. Gaskell, London, pp.163–183.

19.

Liu

(2013) Economic costs of dementia in low and middle income countries. Thesis (Ph. D). King’s College London (University of London).

20.

World Health Organization. Cost effectiveness and strategic planning (WHO-CHOICE). http://www.who.int/choice/en/.

21.

Curtis

(2013) Unit Costs of Health and Social Care 2013. Personal Social Services Research Unit, University of Kent.

22.

Department of Health (2013) NHS Reference Costs 2012–13 London, Department of Health.

23.

Department of Statistics. Singapore Statistics –Consumer Price Index. Department of Statistics, Ministry of Trade & Industry, Republic of Singapore 2013. Available at: http://www.singstat.gov.sg/statistics/browse_by_theme/prices.html.

24.

International Labour Organisation –Global Wage Database 2013, http://www.ilo.org/travail/areasofwork/wages-and-income/WCMS_142568/lang–en/index.htm.

25.

Vaingankar

, Chong

, Abdin

, Picco

, Jeyagurunathan

, Zhang

, Sambasivam

, Chua

, Ng

, Prince

, Subramaniam

(2016) Care participation and burden among informal caregivers of older adults with care needs and associations with dementia. Int Psychogeriatr 28, 221–231.

26.

Cuijpers

, Smit

, Oostenbrink

, de Graaf

, ten Have

, Beekman

(2007) Economic cost of minor depression: A population-based study. Acta Psychiatr Scan 115, 229–236.

27.

Glick

, Doshi

, Sonnad

, Polsky

(2007) Economic Evaluation in Clinical Trials. Oxford University Press, New York, pp. 89–113.

28.

Manning

, Mullahy

(2001) Estimating log models: To transform or not to transform?. J Health Econ 20, 461–494.

29.

Park

(1966) Estimation with heteroscedastic error. Econometrica 24, 888.

30.

Barber

, Thompson

(2004) Multiple regression of cost data: Use of generalised linear models. J Health Serv Res Policy 9, 197–204.

31.

Alzheimer’s Disease International (2006) Dementia in the Asia Pacific Region: The epidemic is here. Asia Pacific Members of Alzheimer’s Disease International.

32.

Konig

, Leicht

, Brettschneider

, Bachmann

, Bickel

, Fuchs

, Jessen

, Köhler

, Luppa

, Mösch

, Pentzek

, Werle

, Weyerer

, Wiese

, Scherer

, Maier

, Riedel-Heller

, AgeCoDe Study Group (2014) The cost of dementia from the societal perspective: Is care provided in the community really cheaper than nursing home care?. J Am Med Dir Assoc 15, 117–126.

33.

Connolly

, Gillespie

, O’Shea

, Cahill

, Pierce

(2014) Estimating the economic and social costs of dementia in Ireland. Dementia 13, 5–22.

34.

Klug

, Halaas

, Peterson

(2014) North Dakota Assistance Program for Dementia Caregivers Lowered utilization, produced savings, and increased empowerment. Health Aff 34, 605–612.

35.

Sousa

, Ferri

, Acosta

, Albanese

, Guerra

, Huang

, Jacob

, Jotheeswaran

, Rodriguez

, Pichardo

, Rodriguez

, Salas

, Sosa

, Williams

, Zuniga

, Prince

(2009) Contribution of chronic diseases to disability in elderly people in countries with low and middle incomes: A 10/66 Dementia Research Group population-based survey. Lancet 374, 1821–1830.

36.

Sousa

, Ferri

, Acosta

, Guerra

, Huang

, Jacob

, Jotheeswaran

, Hernandez

, Liu

, Pichardo

, Rodriguez

, Salas

, Sosa

, Williams

, Zuniga

, Prince

(2010) The contribution of chronic diseases to the prevalence of dependence among older people in Latin America, China and India: A 10/66 Dementia Research Group population-based survey. BMC Geriatr 10, 53.

37.

Albanese

, Liu

, Acosta

, Guerra

, Huang

, Jacob

, Jimenez-Velazquez

, Llibre Rodriguez

, Salas

, Sosa

, Uwakwe

, Williams

, Borges

, Jotheeswaran

, Klibanski

, McCrone

, Ferri

, Prince

(2011) Equity in the delivery of community healthcare to older people: Findings from 10/66 Dementia Research Group cross-sectional surveys in Latin America, China, India and Nigeria. BMC Health Serv Res 11, 153.

38.

Carkinm

, Tracy

(2015) Adjusting for unit non-response in surveys through weighting. Crime Delinquency 61, 143–158.

39.

Tracy

, Carkinm

(2014) Adjusting for design effects in disproportionate stratified sampling designs through weighting. Crime Delinquency 60, 306–325.

40.

Heeringa

, West

, Berglund

(2010) Applied survey data analysis. Taylor and Francis Group, LLC.

41.

Patel

, Rendu

, Moran

, Leese

, Mann

, Knapp

(2005) A comparison of two methods of collecting economic data in primary care. Fam Pract 22, 323–327.

42.

Evans

, Edejer

, Adam

, Lim

(2005) Methods to assess the costs and health effects of interventions for improving health in developing countries. BMJ 331, 1137–1140.

43.

Ortegon

, Lim

, Chisholm

, Mendis

(2012) Cost effectiveness of strategies to combat cardiovascular disease, diabetes, and tobacco use in sub-Saharan Africa and South East Asia: Mathematical modelling study. BMJ 344, e607.