To compare and validate neurocognitive tests in the Harmonized Cognitive Assessment Protocol (HCAP) for the China Health and Retirement Longitudinal Study (CHARLS), and to identify appropriate tests to be administered in future waves of CHARLS.

We recruited 825 individuals from the CHARLS sample and 766 subjects from hospitals in six provinces and cities in China. All participants were administered the HCAP-neurocognitive tests, and their informants were interviewed regarding the respondents’ functional status. Trained clinicians administered the Clinical Dementia Rating scale (CDR) to assess the respondents’ cognitive status independently.

The testing protocol took an average of 58 minutes to complete. Refusal rates for tests of general cognition, episodic memory, and language were less than 10%. All neurocognitive test scores significantly correlated with the CDR global score (correlation coefficients ranged from 0.139 to 0.641). The Mini-Mental State Examination (MMSE), the Health and Retirement Study (HRS) - telephone interview for cognitive status (TICS), community screening instrument for dementia (CSI-D) for respondent, episodic memory and language tests each accounted for more than 20% of the variance in global CDR score (p < 0.001) in bivariate tests. In the CHARLS subsample, age and education were associated with neuropsychological performance across most cognitive domains, and with functional status.

A brief set of the CHARLS-HCAP neurocognitive tests are feasible and valid to be used in the CHARLS sample and hospital samples. It could be applied in the future waves of the CHARLS study, and it allows estimating the prevalence of dementia in China through the population-based CHARLS.

The prevalence and factors associated with delays in help seeking for people with dementia in China are unknown.

Within 1,010 consecutively registered participants in the Clinical Pathway for Alzheimer's Disease in China (CPAD) study (NCT01779310), 576 persons with dementia (PWDs) and their informants reported the estimated time from symptom onset to first medical visit seeking diagnosis. Univariate analysis of general linear model was used to examine the potential factors associated with the delayed diagnosis seeking.

The median duration from the first noticeable symptom to the first visit seeking diagnosis or treatment was 1.77 years. Individuals with a positive family history of dementia had longer duration (p = 0.05). Compared with other types of dementia, people with vascular dementia (VaD) were referred for diagnosis earliest, and the sequence for such delays was: VaD < Alzheimer's disease (AD) < frontotemporal dementia (FTD) (p < 0.001). Subtypes of dementia (p < 0.001), family history (p = 0.01), and education level (p = 0.03) were associated with the increased delay in help seeking.

In China, seeking diagnosis for PWDs is delayed for approximately 2 years, even in well-established memory clinics. Clinical features, family history, and less education may impede help seeking in dementia care.