Are There Differences in Screening MMSE Interview Durations Between Subjects Who Maintain Severity and Subjects Who Drop Below Screening Inclusionary MMSE Thresholds at Baseline? An Exploratory Analysis.
Alan Kott, Xingmei Wang, David S. Miller- Psychiatry and Mental health
- Cellular and Molecular Neuroscience
- Geriatrics and Gerontology
- Neurology (clinical)
- Developmental Neuroscience
- Health Policy
- Epidemiology
Abstract
Background
We have previously identified a large proportion of subjects (> 10%) entering clinical trials in early Alzheimer’s disease to possibly suffer from Baseline inflation (Kott, 2023). Without useful predictive markers, the identification of these subjects happens only when these subjects are already randomized into the trial. Our prior analyses showed an almost linear relationship between MMSE interview duration and MMSE total score (Kott, 2022). To be able to identify possible subjects with MMSE score inflation at Screening, we wanted to test the hypothesis that those subjects with likely inflated MMSE scores had significantly longer scale administration times, possibly corresponding to increased symptom severity below the MMSE inclusionary threshold.
Method
Data from 6,922 subjects were pooled from 6 early AD clinical trials. We restricted the analysis to interviews between 3 and 30 minutes long, shorter interviews are technically not possible and longer interviews likely indicate user error. This decreased the sample to 6,717 subjects. Subjects were dichotomized into subjects meeting Screening MMSE criteria at Baseline and subjects who dropped below the Screening MMSE criteria at Baseline. A generalized linear model was fitted to the data with the Screening MMSE duration used as the predicted variable and type of subject (drops below criteria yes/no) as the predictor variable correcting for the MMSE score at Screening.
Result
937 subjects no longer met MMSE inclusion criteria at Baseline. For those subjects, MMSE interview duration increased significantly (∼40 seconds).
Conclusion
Our post‐hoc analysis found significantly longer Screening MMSE interview durations in subjects who no longer met MMSE inclusion cutoffs at Baseline. The increase was consistent with our hypothesis and likely reflects that these subjects had their Screening MMSE scores inflated in order to meet inclusion criteria. We suggest including prolonged MMSE durations in analytical programs as a marker of possible MMSE score inflation alongside other markers, like MMSE vs ADAS‐Cog discordance. We plan additional analyses to better understand this problem.