The Mini-Mental State Examination (MMSE) has been commonly used to measure cognitive change over time. The aim of present study was to investigate the normative rates of change for the MMSE across test-retest intervals.

We administered the Korean MMSE (K-MMSE) to 1055 community-dwelling middle aged and older adults three times over 6 years. Based on the Christensen's health screening criteria (1991), 234 middle-aged healthy adults (104 men, 130 women; mean age 55.95±6.20 years; age range 45-64 years; mean education 7.63±4.06 years) and 505 healthy elderly (200 men, 305 women; mean age 71.00±4.62 years; age range 65-79 years; mean education 5.61±5.12 years) were selected for the statistical analysis. Reliable change indices were computed using two different statistical methods, the Reliable Change Index adjusted for practice effects (RCI_PE; Chelune et al., 1993) and the Standardized Regression-Based Change Index (SRBCI; McSweeny et al., 1993).

For the middle-aged healthy adult group, the 90% confidence intervals of the RCI_PE and SRBCI were the same such as ±4 in 2-year, ±5 in 4-year, and ±6 in 6-year test-retest intervals. For the healthy elderly group, the 90% confidence intervals of the RCI_PE were -5 and +4 in 2-year interval and -7 and +5 in 4- & 6-year intervals. The 90% confidence intervals of the SRB change index were ±4 in 2-year interval and ±6 in 4- & 6-year intervals.

The result provides the normative data of the reliable change scores for the K-MMSE for the middle-aged and older adults. It shows that the reliable change indices were varied across different age groups as well as test-retest intervals.