BACKGROUND

Subjective questionnaires used for the diagnosis of pre-mild cognitive impairment (pre-MCI) and conventional criteria for MCI might mainly result in false-positive diagnostic errors. The integrated criteria based on demographically adjusted total and process Z scores on neuropsychological tests have been validated to be sensitive for measuring pre-MCI, MCI, and MCI subtypes. However, the underrepresentativity of Chinese populations and the complexity in some tests limit the use of the established Z scores in the elderly Chinese population.

OBJECTIVE

The aim of this study was to develop a useful Z score calculator to assess individual cognitive performance after adjustment of the scores on each of the neuropsychological tests according to sex, age, and education and to establish preliminary norms for the objective assessment of cognition function in elderly Chinese individuals.

METHODS

The neuropsychological test battery consists of measures of performance on different cognitive domains, including episodic memory, language, executive function, processing speed, and attention. Data were obtained from 220 clinically cognitively normal Chinese volunteers aged 60 years or older from the cohort of the Shanghai study of health promotion among frail elderly individuals. Regression models were used to investigate the impact of age, education, and sex on test scores. Z scores were estimated for the different scores by subtracting the predicted mean from the raw score and dividing it by the root mean square error term for any given linear regression model.

RESULTS

Preliminary analyses indicated that age, sex, or level of education significantly affected test scores. A series of linear regression models was constructed for all instruments, i.e.: the Trail-Making Test A and B, to assess executive function or attention; the Boston Naming Test and animal list generation, to assess language; delayed free correct responses and the Hopkins Verbal Learning Test-Revised (HVLT-R) recognition, as well as three process scores, i.e., intrusion errors, learning slope, and retroactive interference, from the HVLT-R, to assess memory, by adjusting for the covariates of age, sex, and level of education concurrently. The Z scores on all neuropsychological tests were estimated based on the corresponding regression coefficients.

CONCLUSIONS

We constructed a multivariable regression-based normative Z score method for the measurement of cognition among older Chinese individuals in the community. The normative score will be useful for the accurate diagnosis of different subtypes of pre-MCI and MCI after preliminary rapid screening in the community.