China

In routine practice, clinicians in China use validated screens, most commonly the Chinese or Mandarin versions of the Montreal Cognitive Assessment (MoCA) and the Mini-Mental State Examination (MMSE). It is often combined with simple add-ons like a Clock-Drawing task to sample executive and visuospatial skills. Recent large Chinese datasets now provide age- and education-adjusted normative values for MoCA, which improves triage accuracy and helps avoid misclassification in people with lower schooling. In memory-clinic settings, these brief tools determine who needs fuller neuropsychological testing, imaging, or biomarker work-ups.

Computed tomography (CT) and magnetic resonance imaging (MRI) are standard parts of the dementia work-up in tertiary hospitals and are used to exclude structural causes and support etiologic diagnosis. Amyloid-positron emission tomography (PET) is available in major centres and has been shown in Chinese memory-clinic cohorts to change diagnoses and treatment plans, particularly in early-onset or diagnostically uncertain cases. However, amyloid-PET is not reimbursed by China’s basic medical insurance and remains costly, so its routine use is limited outside large urban hospitals and research networks. In general, CT and MRI are widely obtainable in tertiary care, but amyloid-PET utilization is still constrained by coverage and price.

Clinical genetic testing is targeted rather than routine. Apolipoprotein E (APOE) genotyping and monogenic panels (APP, PSEN1/PSEN2) are typically ordered for early-onset, familial, or atypical presentations in academic centres or via reference laboratories. China maintains organized research efforts and registries focused on familial and early-onset Alzheimer’s disease, and multi-centre reports describe the mutation spectrum in hundreds of Chinese families. In everyday clinical pathways, however, genetics is not a population screen and is generally confined to selected cases where results will change counselling or management.

Cerebrospinal fluid (CSF) biomarkers (Aβ42/40, total-tau, phospho-tau) are used in tertiary memory clinics to clarify difficult or atypical cases. Recent work from Chinese centres shows that adding CSF improves diagnostic confidence and classification when routine clinical and imaging data are inconclusive. In parallel, blood-based biomarkers, such as notably plasma p-tau217 (and related ratios such as p-tau217/Aβ42), panels including Aβ42, p-tau181, and NfL, are showing high diagnostic and prognostic performance in Chinese research studies. Several studies have established reference intervals and evaluated accuracy against PET and CSF. These blood tests are moving from research toward staged clinical adoption but are not yet a nationwide routine, whereas CSF testing is already embedded at many tertiary sites.

China has rapidly expanded dementia-capable services in tertiary hospitals, building out memory clinics and dedicated inpatient units that bring neurology, psychiatry, geriatrics, neuropsychology, and rehabilitation under one roof. Still, coverage of such services is strongest in major cities, with referral upward still common from smaller cities and counties. Moreover, the government’s new National Action Plan for Coping with Dementia in the Elderly (2024–2030) tasks 15 ministries with standardizing diagnosis and treatment, scaling dementia care units within elder-care institutions, and training the workforce through 2030. These steps are expected to push capacity beyond top hospitals over the next few years.

Most clinic visits and the older Alzheimer’s disease drugs (donepezil, rivastigmine, galantamine, memantine) are covered to some extent by China’s basic medical insurance, based on what each province implements from National Reimbursement Drug List. However, the new anti-amyloid drugs are not yet broadly reimbursed, so people typically pay out-of-pocket unless a hospital or insurer has a special deal with the manufacturer. On top of the drug price, people often need repeated MRI scans and infusion visits which is often not covered by the insurance and makes a significant additional cost.

Family care remains the backbone of day-to-day support with more than 84% of people living with dementia being taken care of by their families. At the same time, national NGOs and hospital programs provide helplines, caregiver education, and periodic respite or skills-training pilots. However, no unified national cash allowance exists specifically for dementia caregiving. Also, the availability of day care, home-based support, and navigation services varies by city and program maturity. One cross-section research shows that caregiver burden in China is strongly shaped by several intersecting factors, including the cognitive status of the person living with dementia, post-acute care (PAC) availability, levels of social support, and the accessibility of community-based services. When cognitive impairment is more severe and community services are limited or fragmented, families carry a heavier load, often without adequate respite, training, or professional guidance. Strengthening formal social support networks, expanding high-quality home-care and day-care services, and scaling PAC facilities can significantly reduce this burden by distributing care responsibilities across trained providers rather than leaving them almost entirely to families. Alzheimer’s Disease Chinese (ADC) and the China Association for Alzheimer’s Disease (CAAD) help fill gaps with public education, carer resources, and professional training.

In January 2025, the Chinese government launched the “National Action Plan for Coping with Dementia in the Elderly (2024–2030)” (sometimes called the National Action Plan on Response to Dementia), a major policy document co-led by the State Council of the People’s Republic of China and multiple ministries. This plan places dementia firmly within the broader Healthy China 2030 strategy, recognizing dementia as not simply a clinical condition but a large-scale public-health and elder-care challenge. The document sets the ambition of building a full lifecycle system that covers prevention, population-level screening and early intervention, standardized diagnosis and treatment, rehabilitation and care expansion. It also involves workforce training, and creation of dementia-friendly environments. For example, one element proposes that 50% of elderly-care institutions with more than 100 beds and sufficient capability will establish dedicated dementia units, while approximately 15 million dementia-care personnel will be trained by 2030.
The Action Plan arises in the context of China’s rapidly aging population, rising life-expectancy, and a corresponding rise in dementia prevalence. It was estimated that China had more than 16 million people living with dementia by 2021, accounting for nearly 30 % of the global total. The plan emphasizes that screening and early intervention will help “effectively control the growth rate” of dementia prevalence, while improved access to care and standardization are central to narrowing the urban-rural and regional service gap. The significance of this plan is that dementia is elevated from a disease-specific clinical policy into the realm of health system planning, elder care policy, workforce development and social inclusion.

Alongside the Action Plan, China has been pursuing a Long-Term Care Insurance (LTCI) Pilot Programme that is the backbone for financing and organizing non-acute (home, community, institutional) care for older people living with disabilities and chronic conditions, including dementia. The LTCI initiative began with fifteen pilot cities selected around 2016 and expanded over subsequent years. The LTCI system is intended to cover eligibility benefits and provider payment mechanisms. Evaluations of the pilots show promising outcomes but substantial heterogeneity. Across cities there are very different eligibility thresholds, service‐coverage scopes, financing channels and quality of care. For example, some cities only covered people living with severe disability, while others began to include mild or moderate cognitive impairment. The LTCI system is seen as the principal mechanism through which the dementia Action Plan’s goals for expanding home and community services and “care beyond hospitals” will be funded and organized.
Looking forward, the strategy foresees the national‐scale rollout of LTCI beyond pilot cities, harmonization of assessment tools and benefit design, integration with the dementia Action Plan. In that case, the services for people living with dementia are explicitly supported through LTCI. It will also strengthen provider supply of community and home care, while also building data monitoring and quality assurance systems. These efforts aim to ensure equitable access across regions, improve service quality, and move from ad hoc to systematized long-term dementia care support.

Large Chinese study populations are advancing blood-based biomarkers, especially plasma p-tau and Aβ ratios, and evaluating clinical utility against PET and CSF, while tertiary clinics continue to refine how CSF and amyloid-PET change diagnosis and management in complex cases. At the same time, teams are producing Chinese normative data and education-adjusted cut-offs for MoCA, improving fairness and scalability of early detection beyond top hospitals.
One research conducted by Shenzhen Healthy Ageing Research (SHARE) project, tested whether machine learning (ML) models combining selected AD8 items with routinely collected health and demographic data could outperform the traditional AD8 ≥2 threshold for detecting cognitive impairment. Using five ML algorithms and feature selection via permutation importance, researchers identified key predictors, including two specific AD8 items (judgment problems and difficulty with financial transactions), sex, age, body mass index, waist circumference, platelets, serum uric acid, white blood cells, heart rate, carcinoembryonic antigen, and abnormal electrocardiography (ECG). The multilayer perceptron (MLP) achieved the highest accuracy (AUC 0.83), substantially exceeding the baseline AD8 score (AUC 0.65), with AdaBoost, SVM, and GBDT also performing well. SHAP analysis showed female sex, older age, and lower waist circumference were the strongest contributors to cognitive impairment predictions. The findings demonstrate that integrating brief informant items with simple health parameters and ML techniques can greatly improve community dementia screening efficiency in China, offering a practical approach for early identification while minimizing resource demands.
China Health and Retirement Longitudinal Study (CHARLS) participates in the international project the Health and Retirement Study (HRS), which collects nationally representative, longitudinal, and harmonized data that align with global standards to enable cross-national research on aging. These studies integrate rich cognitive assessments, biomarker collection, and repeated health and socioeconomic interviews, offering a multidimensional picture of aging in China that is directly comparable to data from the United States and other countries. CHARLS and related harmonized datasets, accessible through the NIA-supported Gateway to Global Aging, are essential for understanding cognitive aging patterns, dementia risk, and the social and economic determinants of late-life health within China’s rapidly aging population.