Is a trial stalled for lack of a control arm
Build rare-disease control arms from real-world records.
Why it stalls
Rare and severe disease trials often cannot randomize a control, so they run single-arm. That leaves nothing to compare against, and sourcing natural-history data becomes the bottleneck of the whole program.
One hospital's records are not enough. These patients move between hospitals, and a single institution holds only one segment of the path.
Our corpus links multiple hospitals' records under one patient: 610 patients hold records from two or more hospitals, followed for 10.1 years on average.
What can be built
If the current corpus holds it, we confirm now. If not, we collect it.
External control arm
Real-world records of patients with the disease who did not receive the treatment. The comparison axis for a single-arm trial.
Natural history cohort
Longitudinal records following the disease itself, not a treatment window. Mean 10.1 years, median 6.7.
New cohort construction
If the patients you need are not in the corpus, we define the criteria and collect them on patient consent, starting from an up-front contract covering issuing costs.
Our part and yours
Finding this out after signing is too late, so it goes here.
- We build the cohort and the analysis documents on top of it. Filing with regulators and responding to review are yours.
- We cannot guarantee a regulator will accept the evidence. Our job ends at documenting exactly what was counted and how.
- Patient-level data does not leave even under contract. What leaves is aggregates and documents.
Common questions
- What if the disease we need is not in the current corpus?
- Give us the criteria and we first measure what the current corpus holds. If it falls short, we scope the collection cost and move to a construction contract.
- Does it meet regulator-grade audit trail requirements?
- The output states what was matched, on which spelling, against which denominator, and what was removed. Requirements beyond that are scoped at contract.
- Can a control arm stand without patient-level data?
- It depends on the use. Some questions hold at the aggregate level and some do not. The fastest way to find out is a meeting with your own protocol.
