STATISTICS IN MEDICINE Statist. Med. 2000; 19:1607–1616 Modelling Mini Mental State Examination changes in Alzheimer’s disease Marta S. Mendiondo 1;*;† , J. Wesson Ashford 2 , Richard J. Kryscio 3 and Frederick A. Schmitt 4 1 Center on Aging, 101 Sanders-Brown Building; University of Kentucky; Lexington; KY 40536; U.S.A. 2 Departments of Psychiatry and Neurology; Sanders Brown Center on Aging; 901 S. Limestone; Lexington; KY 40536; U.S.A. 3 Department of Statistics; Room 817 P. O. T.; University of Kentucky; Lexington; KY 40536-0027; U.S.A. 4 Departments of Neurology; Psychiatry and Psychology; Sanders Brown Center on Aging; 901 S. Limestone; Lexington; KY 40536; U.S.A. SUMMARY The Mini Mental State Examination (MMSE) is widely used to measure dementia severity in Alzheimer’s disease patients. While changes over time in the MMSE due to dementia have been studied, the relationship between MMSE scores and the duration of disease course is less well understood. Using the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) data, we modelled change in MMSE as a function of time for this population. For this purpose we used the interval between consecutive MMSE assessments as the time factor. We also investigated the impact of sex, education and age at testing on the resulting model. Analyses showed that Alzheimer’s disease progression over time (ADP) can be modelled using a cubic or a logarithmic function of MMSE score. From these curves ADP can be obtained as a function of MMSE. These models demonstrate that there are dierent rates of change for various ranges of the MMSE. Additional analyses suggest that patient factors aect rates of ADP, younger patients and more educated patients progress more rapidly, while sex has little impact on ADP. Such estimations of disease course are useful when comparing dierent populations for both clinical and research purposes. Copyright ? 2000 John Wiley & Sons, Ltd. INTRODUCTION Assessment of illness severity is critical to the diagnosis, treatment and research of Alzheimer’s disease (AD). Several clinical staging methods have been developed for estimating dementia sever- ity in AD patients. However, these broad categorical approaches can suer from observer bias. Consequently, structured mental status measures are often used in conjunction with clinical ratings for more objective measurement of dementia severity. One of the original eorts in this regard indicated that scales incorporating cognitive and descriptive items correlate with the underlying * Correspondence to: Marta S. Mendiondo, Center on Aging, 101 Sanders-Brown Building, University of Kentucky, Lexington, KY 40536, U.S.A. † E-mail: marta@ms.uky.edu Copyright ? 2000 John Wiley & Sons, Ltd.