Back

The anatomy of regression-to-the-mean in simulated epilepsy trials

Goldenholz, D. M.; Bhansali, R. M.; Kaptchuk, T. J.; Westover, M. B.

2026-07-31 neurology
10.64898/2026.07.27.26359051 medRxiv
Show abstract

Regression to the mean (RTM) can inflate apparent placebo response in epilepsy trials, but its mechanisms are often conflated. Using CHOCOLATES, we simulated 1,000,000 patients with 36 months of daily seizure counts and simulated placebo trials: 2-month baselines followed by 3-month test periods without treatment effects. Transient worsening (RTM type 1), stricter eligibility thresholds (RTM type 2), reduced sensitivity, and false alarms (RTM type 3) each increased RTM and apparent response. These findings show that placebo-arm improvement can arise from temporary illness, natural variability, measurement error, or mixtures thereof, informing epilepsy trial design and endpoint interpretation.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.