The questions people
actually ask.
Placebo replacement, therapeutic-area coverage, and defensibility for trial teams. Where the data lives, what partners earn, and how governance holds for the institutions that supply it.
Method, evidence, and defensibility
01Do you replace the placebo arm?
No. Today we build hybrid and augmented external control arms that reduce and strengthen the control arm, not fully replace the placebo arm, which regulators still generally require in confirmatory trials. Full registrational replacement is our long-term goal and is indication- and regulator-specific.
02Which therapeutic areas do you cover?
The method is therapeutic-area agnostic. Psychiatry is where we benchmarked first. The strongest regulatory precedent for external control arms is in oncology and rare disease. We extend to neurology, immunology, and cardiometabolic where the disease course and endpoints support it. See the Therapeutic Areas explorer for details and a maturity tag per area.
03How do you keep endpoints defensible?
Every derived endpoint carries item-level traceability back to its source record, and is validated against assessed ground truth with clinical partners. Your biostatistics team can inspect the provenance of every data point.
04Where does the data come from, and is it compliant?
From governed partnerships with health systems. Data is de-identified at source and processed inside the partner's governance boundary. It never leaves their servers. We operate within HIPAA, GDPR, and DPDP frameworks.
05Are you a CRO?
No. We are a specialized infrastructure layer for external control arms. We work alongside your CRO and your biostatistics team, not in place of them.
06How is this different from Unlearn, Medidata, or Flatiron?
Those companies validate the category. Our angle is serving sponsors they overlook, being therapeutic-area agnostic, and building from partner data that never leaves the institution rather than from a single fixed dataset. In short: Unlearn focuses on digital twins and covariate adjustment, Flatiron on oncology real-world data, and Medidata offers a synthetic control arm within a broader trial-technology suite.
07What regulatory guidance do you align to?
FDA's 2023 draft guidance on externally controlled trials, FDA's January 2025 draft guidance on AI in regulatory decision-making (a seven-step credibility framework), and EMA precedent qualifying AI methods such as PROCOVA.
08What does an engagement look like?
Protocol intake, feasibility mapping against available real-world data, cohort assembly, validation against ground truth, and a documentation hand-off your team can defend. Typical timeline to be confirmed.
09How much can this save?
Reducing the control arm by roughly half can cut control cost and shorten recruitment. Use the estimator to see illustrative figures for your phase and area. Numbers are illustrative, not a quote.
10Do you handle the regulatory submission?
We deliver regulatory-ready documentation and evidence packages aligned to external-control guidance. Your regulatory team leads the submission.
11What if my indication has subjective endpoints?
That is our origin. Deriving validated scales from unstructured clinical narratives, as in psychiatry, is the hardest version of this problem, and it is what we built first.
12Can you support single-arm and hybrid designs?
Yes. External controls for single-arm studies, and augmented or hybrid control arms for randomized designs.
Data partnerships and governance
01Does our data ever leave our servers?
No. Data is de-identified at source and structured in place inside your governance boundary. Sapien works through a governed research key; raw records never leave your environment.
02What do we get?
A new revenue stream from data you already hold, a trial-ready query layer over your own data, retained academic publishing rights, and a stronger position as a trial site, without building or staffing the infrastructure.
03What does Sapien get?
Commercial rights to build external control arms and cohorts for sponsors from the structured data. We never sell raw records.
04Is this HIPAA, GDPR, and DPDP compliant?
Yes. Governance, access, and provenance are defined with your team before any data moves, inside your compliance framework.
05How much work is it for our team?
Minimal. We bring the data engineering and regulatory science. Your clinicians keep treating patients.
06Who owns the data?
You do. You remain the steward of your data and a named contributor to the evidence it generates.
07What about patient privacy?
Strict de-identification at source. Every output can be traced to a source record for audit, but identifiers are removed before any processing.
08Has this model worked elsewhere?
Yes. Leading systems already run data-in-place partnerships with commercial data companies. The Mayo Clinic and nference relationship is a public industry precedent, not a Sapien partnership. We bring the same model across therapeutic areas, starting with psychiatry.
09What is the commercial arrangement?
A partnership model: your data structured and de-identified in place; you keep publishing rights and a query layer; we gain commercial rights. Specifics are defined per partnership.