Regression-based normative scores for the Montreal Cognitive Assessment (MoCA) in an Asian population

This study aimed to calculate stratified normative scores of the Montreal Cognitive Assessment (MoCA) in an adult population in Singapore, accounting for key demographic influences. Demographic data and MoCA scores of 1,103 healthy adults (aged 21 to 97) were obtained from a community health study conducted in central Singapore. Factors associated with MoCA scores were identified using multiple linear regression and β coefficients were used to estimate normative MoCA scores across strata. Model performance was assessed using five-fold cross-validation. Normative reference scores were calculated and stratified by age group, education level, and ethnicity to reflect typical MoCA performance across demographic groups. The final regression model had an adjusted R² of 0.284 (p < 0.001), with age group (β = -0.325 to -2.312) and education level (β = 1.783 to 4.206) accounting for the majority of the explained variance (R² = 0.271). Ethnicity also remained a significant factor in the model, with lower scores observed among Malay (β = -1.248) and Indian (β =-0.795) participants compared to Chinese. Among the 64 demographic combinations of age group, education level and ethnicity, the lowest normative score (20.0) was derived for Malay individuals aged ≥ 75 years with no formal education. MoCA scores varied systematically with age, education level, and ethnicity in the study population. The resulting stratified reference scores provide clinicians and researchers a useful context for interpreting individual MoCA performance relative to demographically similar peers in Singapore's adult population. However, these reference scores are not diagnostic thresholds and should be interpreted with caution until validated against clinically diagnosed cognitive impairment.