Comment from Gillian Sapia
AnonymousSupportIndividual
Summary: The commenter, a caregiver for a child with an ultra-rare disease, proposes a new clinical trial framework for ultra-rare diseases that focuses on individual "N-of-1" trajectories rather than group averages. They argue that because ultra-rare diseases are highly heterogeneous, trials should measure meaningful deviations from a patient's own baseline using real-world data and personalized outcomes.
I have this amazing idea... based off things FDA has released... ultra rare is in a gray area but I really think this is possible. I’m working on an idea I can’t stop thinking about: a different way to run trials in ultra-rare disease.
Ultra-rare isn’t just “low numbers.” It’s feasibility. It’s heterogeneity. It’s patients whose disease expression and progression don’t fit a single composite endpoint and whose most meaningful outcomes are not the same as the person sitting next to them in the study. What if we stopped forcing ultra-rare patients into one endpoint hierarchy and instead built the trial around what ultra-rare actually looks like?
Here’s the concept: Instead of “N-of-1” meaning one patient receiving one treatment, imagine a full trial where EVERY participant is an N-of-1.
In a trial of 42 patients, you’d run 42 structured N-of-1 baselines then combine them.
How it would work:
Each patient completes a pre-treatment baseline period using structured real-world data (RWE) tracking (think: Citizen Health AI-supported longitudinal data capture). Each patient pre-specifies 1–3 measurable outcomes that are meaningful for THEIR phenotype (not a generic composite that washes out signal). Treatment effect is analyzed within each person: change from their own baseline trajectory (slope / progression / event frequency / symptom burden). Then we aggregate standardized within-patient effect sizes across the trial to see overall benefit without pretending heterogeneity doesn’t exist. Why this matters: Composite endpoints can hide real benefit in ultra-rare disease because progression isn’t uniform. One patient may be highly progressive; another may be stable but cognitively impacted; another may have seizures as the defining feature. Forcing them into the same endpoint “stack” risks missing what’s actually clinically meaningful.
This approach reframes the primary question from:
“Did the group average improve?”
to:
“How many individuals showed meaningful deviation from their expected trajectory?”
Ultra-rare demands rigor but it also demands honesty about biology, variability, and what families are actually living.
I’m going to share a real example next: my daughter Penelope’s slope data (before vs on treatment) across outcomes that matter as her caregiver. I’ll add those visuals.
