A statistical shortcut increasingly used to make the case for a new class of Alzheimer's drugs can produce overstated claims about how those drugs actually work, according to a research letter from the Brown University School of Public Health published in JAMA Neurology.
The technique at issue is called quantile aggregation. It sorts trial participants into groups, averages the results within each group, and then looks for patterns across those group averages. The Brown team examined what happens when that method is applied to the relationship between cognition and amyloid, the protein that accumulates in the brains of people with Alzheimer's disease. The approach was first published in an analysis of donanemab, Eli Lilly and Company's Alzheimer's drug.
Why this matters comes down to what averaging does to individual differences. Senior author Sarah Ackley, an assistant professor of epidemiology at Brown who runs the Computational Epidemiology Lab, noted that many researchers believe reducing amyloid buildup could slow the memory loss and cognitive decline of the disease, which is why amyloid has become a central target for newer drugs. If a method smooths away the variation between patients, an apparent link between clearing amyloid and preserving thinking skills can look tighter and more convincing than the underlying trial data support.
For patients and families weighing these medications, the finding is a reminder that the strength of the evidence depends heavily on how trial results are analyzed, not just on the headline numbers. The letter does not challenge whether the drugs remove amyloid; it questions whether this particular analytic technique can reliably show that removing amyloid is what drives any cognitive benefit.