This study compared two methods for detecting reliable cognitive decline — classical test theory (CTT) and item response theory (IRT) Reliable Change Indices (RCIs) — in 1,136 cognitively unimpaired older adults with elevated brain amyloid enrolled in the A4 clinical trial.
While CTT- and IRT-based RCIs were strongly correlated overall (r = .83), among the 92 participants flagged as declining by at least one method, 44 (47.8%) received discordant classifications — meaning nearly half were labeled differently depending on which method was used.
- Sample limited to cognitively unimpaired, amyloid-positive older adults, reducing generalizability to broader populations. - Only the MMSE was used, a tool with known ceiling effects; findings may not extend to other screening instruments. - No gold-standard outcome (e.g., clinical progression) was used to determine which RCI method was more accurate.
Clinicians using repeated MMSE screening to detect early cognitive decline should be aware that CTT-based RCIs may misclassify nearly half of individuals near the ceiling — consider IRT-based methods or instruments with better precision in high-functioning older adults.
Explore related topics
