The economic burden of dementia in China, 1990–2030: implications for health policy

Abstract Objective To quantify and predict the economic burden of dementia in China for the periods 1990–2010 and 2020–2030, respectively, and discuss the potential implications for national public health policy. Methods Using a societal, prevalence-based, gross cost-of-illness approach and data from multiple sources, we estimated or predicted total annual economic costs of dementia in China. We included direct medical costs in outpatient and inpatient settings, direct non-medical costs – e.g. the costs of transportation – and indirect costs due to loss of productivity. We excluded comorbidity-related costs. Findings The estimated total annual costs of dementia in China increased from 0.9 billion United States dollars (US$) in 1990 to US$ 47.2 billion in 2010 and were predicted to reach US$ 69.0 billion in 2020 and US$ 114.2 billion in 2030. The costs of informal care accounted for 94.4%, 92.9% and 81.3% of the total estimated costs in 1990, 2000 and 2010, respectively. In China, population ageing and the increasing prevalence of dementia were the main drivers for the increasing predicted costs of dementia between 2010 and 2020, and population ageing was the major factor contributing to the growth of dementia costs between 2020 and 2030. Conclusion In China, demographic and epidemiological transitions have driven the growth observed in the economic costs of dementia since the 1990s. If the future costs of dementia are to be reduced, China needs a nationwide dementia action plan to develop an integrated health and social care system and to promote primary and secondary prevention.


Introduction
According to the 2013 Alzheimer's Disease International report, about 44.4 million people were living with dementia in 2013 and this number is expected to reach an estimated 75.6 million by 2030. 1 In China, which has the largest population of people with dementia, the prevalence of dementia appears to have increased steadily between 1990 and 2010. 2,3 However, this trend might be partly attributed to temporal variations in the methods used to estimate such prevalence. 4 The results of a national survey in 2008-2009 indicated that dementia was more common in rural areas than in urban settings. 5 Given the rapid growth of the elderly population in China, 6 dementia is expected to pose tremendous challenges to the national health-care system and to the sustainable development of the national economy.
Most cost-of-illness studies for dementia have been carried out in high-income countries such as Sweden, the United Kingdom of Great Britain and Northern Ireland and the United States of America. [7][8][9][10][11] The economic costs of dementia in China -which have yet to be investigated in detail -are likely to differ, both in magnitude and type, from those in such distant high-income countries.
In this study, we sought to estimate and predict the costs of dementia in China for the periods 1990-2010 and 2020-2030, respectively. It was hoped that, by quantifying the economic costs of dementia, Chinese policy-makers would be motivated to develop a nationwide action plan, prioritize policies on dementia-related care and research and reduce the economic and societal burdens of dementia in China.

Methods
In this cost-of-illness study, we used a prevalence-based, bottom-up approach to quantify or predict the costs of dementia in China between 1990 and 2030, from a societal perspective. We categorized all the costs into three classes: 12,13 (i) direct medical costs, that is goods and service costs related to the diagnosis and treatment of inpatients and outpatients with dementia; (ii) direct non-medical costs, that is transport costs and costs related to formal care in nursing homes or informal care at home; and (iii) indirect costs resulting from dementiaattributable loss of productivity.

Data sources
We used multiple data sources for all estimates. We used age-specific prevalence of dementia in China, for the period 1990-2010, derived from a comprehensive systematic review. 3 From the electronic health records of the facilities, we collected cost data for patients with diagnosed dementia who were admitted either to the Shandong Centre for Mental Health -the only provincial psychiatric hospital in the eastern province of Shandong -between 1 January 2005 and 31 March 2014 or to the Daizhuang Psychiatric Hospital -one of the oldest psychiatric hospitals in China and also in Shandong province -between 1 January 2012 and 30 September 2014. The routine electronic health records include sociodemographic data and data on clinical diagnosis and disease classification, itemized costs, e.g. for drugs, examinations and beds. In each of the two study facilities, dementia was diagnosed and defined according to the International statistical classification of diseases and related health problems, 10th revision. 14

We
Objective To quantify and predict the economic burden of dementia in China for the periods 1990-2010 and 2020-2030, respectively, and discuss the potential implications for national public health policy. Methods Using a societal, prevalence-based, gross cost-of-illness approach and data from multiple sources, we estimated or predicted total annual economic costs of dementia in China. We included direct medical costs in outpatient and inpatient settings, direct non-medical costs -e.g. the costs of transportation -and indirect costs due to loss of productivity. We excluded comorbidity-related costs. Findings The estimated total annual costs of dementia in China increased from 0.9 billion United States dollars (US$) in 1990 to US$ 47.2 billion in 2010 and were predicted to reach US$ 69.0 billion in 2020 and US$ 114.2 billion in 2030. The costs of informal care accounted for 94.4%, 92.9% and 81.3% of the total estimated costs in 1990, 2000 and 2010, respectively. In China, population ageing and the increasing prevalence of dementia were the main drivers for the increasing predicted costs of dementia between 2010 and 2020, and population ageing was the major factor contributing to the growth of dementia costs between 2020 and 2030. Conclusion In China, demographic and epidemiological transitions have driven the growth observed in the economic costs of dementia since the 1990s. If the future costs of dementia are to be reduced, China needs a nationwide dementia action plan to develop an integrated health and social care system and to promote primary and secondary prevention.

Research
Economic burden of dementia in China Junfang Xu et al. excluded 26 patients with dementia who were diagnosed as having other chronic conditions that needed treatment, e.g. anxiety, diabetes or hypertension, leaving data from the records of 146 patients with dementia in our analysis. We also searched the China National Knowledge Infrastructure, PubMed and Wanfang bibliographic databases for studies, on the use of health resource by people with dementia in China, published between 1 January 1990 and 31 July 2015. The search terms included "Alzheimer's disease", "China", "cost burden", "dementia", "economic burden", "formal care" and "informal care". We obtained costs for outpatient visits and transportation from a published study. 15 Data on demographics and wages came from the China Statistical Yearbook 2015. 16 The United Nations population projections for China 17 and predictions of the prevalence of dementia based on data from a systematic review 3 were used to estimate the total numbers of people in China who would have dementia in 2020 and 2030.

Cost estimates
For our estimates we included costs for hospitalization, formal care, informal care, loss of productivity, outpatient care and transportation. Our estimates of hospitalization costs, which included the costs of all medicines, clinical examinations, specialist consultations and bed care, were based on the mean values of the costs recorded in the electronic health records of the two study facilities and the values given in a published    18,19 Informal care costs were estimated assuming that a carer, who might otherwise be earning the national mean salary, 13 spent a mean of 6.3 hours per day 20 giving care to each dementia case in informal care at home. We assumed that, during the observational periods, 86% of Chinese dementia cases were receiving informal care at home and that 4.9% of such cases would seek formal care. 13,21 We estimated the numbers of disability-adjusted life-years (DALYs) lost because of dementia using each case's dementia severity score, assessed using the Global Deterioration Scale, 22 and different weights for each of seven levels of severity. 23 Costs of the productivity lost because of dementiarelated disability were then estimated. For these estimations, we assumed that each person aged at least 60 years had a mean productivity weight of 0.1 and we used an annual discount rate of 3.5% to adjust the costs to 2015 values. 24 Outpatient care costs included the costs of treatments and specialist consultations received in clinics or at home. 15 Transportation costs comprised the costs travelling to and from medical centres. 15 All the estimated costs were converted to United States dollar (US$) values in January 2015, when US$ 1 was equivalent to about 6.2 Chinese yuan.

Statistical analysis
We estimated the numbers of people with dementia in China in 1990, 2000 and 2010 by multiplying the age-specific prevalence of dementia 3 by the corresponding numbers of people in each age group in the population. For our predictions for 2020 and 2030, we used age-specific prevalence derived using a regression model and the relevant data from a comprehensive review. 3 We estimated the total annual costs of dementia by multiplying the mean costs per patient by the total number of patients with dementia. Total annual costs for 2020 and 2030 were projected using a dynamic general disequilibrium model 25 and assuming that the use of health resources by a dementia case was constant while the elderly population grew and the age-specific prevalence of dementia varied over time. We employed the Laspeyres decomposition method to estimate the relative contributions made by the ageing population and changes in the age-specific prevalence of dementia to the predicted future costs of dementia. 26 When the relevant data on medical costs for particular years were missing, we assumed that those costs would have increased by the same amount as the per-capita gross domestic product (GDP).
We conducted multiple sensitivity analyses to assess the impact of variations in the key input parameters on our primary estimates. Specifically, we estimated the total costs of dementia by (i) using prevalence data derived from a different systematic review 2 that yielded more conservative estimates of the prevalence of dementia than those that we used 3 for our primary estimates; (ii) assuming that medical costs would increase 5% every year; (iii) using the minimum and maximum values -instead of the overall mean -for the hours spent on informal care; 27 (iv) using the means of the minimum and maximum values recorded, in China's 22 provinces, five autonomous regions and four municipalities, for an informal carer's wages -instead of the overall national mean value; (v) assuming that dementia cases aged at least 60 years had a mean productivity weight of 0.50 27 instead of 0.1; (vi) assuming that 70% or 99% of people with dementia -instead of 86% -would live at home; 7 and (vii) assuming that 60.4% or 0.4% of dementia patients -instead of 4.9% -would seek professional care. 21

Results
The total number of people with dementia in China was estimated to be about 3.5 million in 1990, 5.1 million in 2000 and 9.6 million in 2010 ( Table 1). The overall prevalence of dementia among people aged at least 60 years was projected to increase from 5.8% in 2020 to 6.7% in 2030 ( Table 2). The total number of people with dementia in China was projected to reach 14.1 million by 2020 and 23.3 million by 2030.
Between 1990 and 2010, the mean annual costs of formal and informal care for each dementia case increased more than 60-fold and more than 18-fold, respectively (Table 3). Over the same period, the national total annual costs of dementia increased more than 50-fold, from about US$ 0.9 billion in 1990 to approximately US$ 47.2 billion, or about 0.7% of China's GDP in 2010.
The total costs of dementia were projected to reach US$ 69.0 billion in 2020 and US$ 114.2 billion in 2030 (Fig. 1). The ageing population and increases in dementia prevalence appeared to be the major driving factors for the high costs of dementia in 2010-2020 and the ageing population also appeared to be the dominant force behind the growth of dementia costs between 2020 and 2030 (Fig. 2). According to our estimates, the costs of informal care for dementia accounted for 94.4%, the total costs of dementia care in China in 1990, decreasing to 92.9% in 2000 and 81.3% in 2010 (Fig. 3). In contrast, the costs of formal care accounted for just 2.7% of the total costs in 1990, increasing to 2.8% in 2000 and 12.8% in 2010. Hospitalization costs accounted for only about 1% of total costs in each year between 1990 and 2010.
The results of our sensitivity analyses indicated that variations in informal care hours, prevalence of dementia and productivity weighting had a substantial impact on estimates of the total costs of dementia, whereas changes in medical prices, wages and the proportions of patients living at home or seeking professional care had relatively little impact on such estimates (Table 4).

Increasing costs of dementia
In China, from an age of 60 years, the prevalence of dementia almost doubles every five years and, as elsewhere, 28,29 about half of those who have survived to an age of at least 90 years are affected by dementia. According to our estimates, the economic burden of dementia in China will increase substantially over the next few decades and this increase will be driven primarily by population ageing and the increasing prevalence of dementia. Our estimates of the total costs of dementia in China for 2010 and 2030 represent about 7.8% of the estimated US$ 604 billion global cost in 2010 and 10% of the forecasted US$ 1110 billion global cost in 2030. 1,7 Our primary estimates of dementia costs are probably underestimates as they ignore the effects, on the costs of care, of home visits by professional careers, the predicted increases in the prevalence of diabetes, hypertension and other risk factors for dementia 30 and the predicted increases in the use of medical services and devices.
Our analyses indicated that, in China, informal care accounted for more than 80% of the total dementia costs in 2010. This proportion is consistent with the global trend revealed by a systematic review 31 but higher than the proportion, of about 60%, reported for low-and middle-income countries by the World Health Organization. 32 In China, the care of an elderly family member is traditionally perceived as an act of filial piety and dementia cases are therefore generally kept out of nursing homes. 15,33 In rural areas there is also a lack of facilities and professional carers for dementia cases. 34 There is no consensus on how to estimate the costs of informal care -with regard to care hours, costs per hour or types of caregiver -or the potential indirect costs, e.g. of lost productivity, incurred by informal caregivers. 7,13,20 As in previous estimates of the global costs of dementia, 7,13 we assumed that informal caregivers would be earning the national mean wage if they were not caring for dementia cases. Our estimates of the costs for hospitalization due to dementia indicated that such costs increased more slowly between 2000 and 2010 than between 1990 and 2000. It seems likely that hospitalization costs became more stable in 2003, when the Chinese government launched a programme of health-care reform designed to make medical treatment more accessible. 35 Our data also revealed a substantial increase in the proportion of the total costs of dementia care attributed to formal care between 2000 and 2010. In countries with rapid economic growth, the responsiveness of the quantity demanded for a service to a change in the income of the people demanding that service, that is, the   36,37 it remains to be seen whether increasing wealth in China will have much effect on the uptake of formal care for dementia cases. However, it seems possible that China's one-child policy, which was implemented in the late 1970s and only phased out from 2015, will leave too few adults in the 2030s to give sustainable informal care to all of the dementia cases. 38,39 Implications for health policy In China, public health policy needs to be tailored to address the economic burden posed by dementia. There needs to be greater focus on developing an integrated health and social-care system, including improvements in the efficiency of dementia care and improved health education and financial and social support for dementia cases and their caregivers. Dementia care might be integrated with the national programme for critical illness insurance. Informal caregivers could be taught knowledge and skills relevant to dementia care. 40 Central or local governments could adopt preferential tax policies or offer other financial incentives to encourage nongovernmental organizations to participate more in dementia-related care services and education. Policy-makers also need to establish a strategic action plan designed to promote the primary and secondary prevention of dementia. In the past few decades, epidemiological studies have identified several modifiable risk fac-tors for dementia, e.g. cardiovascular disease, that can be targeted for primary prevention. 41 Current intervention strategies against cardiovascular disease are likely to be effective in delaying dementia onset. The declining incidence of dementia in some high-income countries may be attributable to improvements in the control of other risk factors such as diabetes, hypertension and smoking. [42][43][44][45] In addition, in settings where effective medical and social-care interventions are available, screening for the early detection of dementia may be cost-effective. 46,47 Geriatricians in clinical settings should be alert to the first symptoms and signs of the dementia syndrome. Early interventions for dementia may delay entry into nursing homes and reduce the overall costs of care. 48

Strengths and limitations
Our estimates were based on data from multiple sources. The cost data on the use of health resources by dementia cases were from itemized routine hospital records, prevalence data were from a comprehensive systematic review 3 and United Nations population projections and data from the National Bureau of Statistics of China represented the most authoritative sources for demographic and income data. Our study had several limitations. First, given the considerable variations in economy, cultures, health-care systems, social welfare and traditions across China, cost data from two health facilities in Shandong province are unlikely to be nationally representative. We partly addressed this concern by weighting our estimates and performing multiple sensitivity analyses. The values for income per capita, health expenditure and hospital costs for Shandong province are similar to the mean national values. 49 Second, the growth rate of medical prices is generally faster than that of the GDP values we used to fill gaps in the data on dementia costs. Given the increasing awareness and use of health services for dementia, it also seems likely that the costs of medical services will increase more rapidly in the coming decades than by 5% annually, that is, by the rate we used in our sensitivity analysis. Third, as people with dementia often suffer from other chronic health conditions, 50,51 comorbidities might reduce the accuracy of our estimates of dementia costs. Fourth, data on the proportion of patients with

Research
Economic burden of dementia in China Junfang Xu et al.
dementia seeking professional care were very limited. Finally, while we used a dynamic model in our projection to account for the ageing population and changes in prevalence of dementia over time, alterations in other factors -e.g. use of medical devices, prevalence of risk factors for dementia and the hiring of professional caregivers -may have reduced the accuracy of our predictions. If, in the future, additional data become available, alternative approaches such as micro-simulation may provide better cost estimates. 52,53 In conclusion, demographic and epidemiological transitions in the past two decades have driven substantial growth in the economic costs of dementia in China. This trend is likely to continue over the next two decades.
Given the huge economic burden of dementia, policy-makers in China are advised to make dementia a national health priority and to develop a strategic nationwide action plan. Failure to take appropriate action now will allow the economic burden of dementia to grow even further and could, in the long-term, cause dysfunction throughout China's entire health-care system. ■ Competing interests: AW received grants and personal fees from Eli-Lilly, grants from Nutricia, Ingenix and Envivo and personal fees from Lundbeck. AW owns the copyright on the Resource Use in Dementia instrument, for which he receives royalties.