P30 Trend in dementia incidence in China 2002–2014: population-based longitudinal study

Background Globally, some 50 million people had Alzheimer’s disease and related dementias in 2018. China accounts for approximately a quarter of cases. It is estimated that the cost of dementia in China was approximately USD 50 billion in 2010. China’s rapid ageing process is occurring at an earlier stage of economic development than other countries, posing great demands on society, particularly health and care services. Accurate estimation of the recent time trend in the incidence of dementia in China is critical for tackling future challenges and formulating public health policies. There is evidence from several high-income countries of a recent downward trend in dementia incidence rates. It is currently unclear, in view of its less-developed status, whether the time trend is also downward in China. The aim of this study is to estimate the recent time trend in dementia incidence in China and to explore how much the effect of changes in risk factors on this trend. Methods The Chinese Longitudinal Healthy Longevity Study (CLHLS) provides five waves of data (2002–2014). CLHLS wave 3 to wave 8 provided information about health status and quality of life of 47,584 people aged over 60 years in 23 provinces, municipalities and autonomous regions of China. In order to maintain representativeness, refreshment participants were recruited to study periodically. Dementia was ascertained at each wave using standardized cognitive and functional impairment criteria. Competing risks of mortality and non-random dropout were accounted for in analysing the temporal trend in dementia incidence. Joint model was fitted to estimate the temporal trends in dementia incidence in China. R package JM was employed to fitted the joint model. Results Age-sex adjusted dementia incidence (2002–2014) increased at an annual rate of 2.5% (95% CI 2.4%-2.6%). Changes in potential risk factors accounted for about 20% of the time effect in dementia incidence. Conclusion Time trends in dementia incidence in China was upward recently. To our knowledge, this is the first study to estimate the trend in dementia incidence in a middle-income country. This longitudinal study used joint modelling with time-to-event data to explore the time trends in dementia incidence for 12 years in China. The first limitation of this study was not possess data from eight provinces in remote areas in China. The second potential limitation is caused by the recall bias and measurements bias over time. The third limitation is that we assumed a constant rate of change of the dementia incidence rate which based on the rate of change was uniform across the Chinese population, but the population structure changed between 2002 to 2014.