Available with limitations

Connect observational evidence to transparent, reproducible methods.

Observational analysis depends on the cohort definition and the assumptions a person wrote down. ValidTrials can ingest a FHIR bundle, store a hashed patient key, and run a declared association or causal contrast with a propensity model, an inverse-probability estimate, a standardized difference and an E-value. Small counts are hidden. Federation across separate deployments is not available: this is one node.

What this capability does

Hashed identity

The source patient id is not stored.

Cohorts

Assumptions must be written. A short note is refused.

Declared design

The label association or causal is the analyst’s declaration. The software does not choose it.

E-value

The formula is VanderWeele and Ding. It is a sensitivity number, not a proof.

Small cells

A count below 10 is hidden, and the total is hidden if any cell is hidden.

One node

There is no multi-organization network across deployments.

How it works

Define the cohort

Write the assumptions.

Check the data

See what was dropped.

Run

Use the declared design.

Read the sensitivity

The E-value does not replace a limitation.

Connected modules and services

Questions

Is a causal label a causal proof?

No. It is what the analyst declared.

Can two hospitals compute together without sharing rows?

Not across separate deployments. Aggregates on this one node still require an administrator’s consent and still hide small cells.

Available with limitations means the described software behavior exists. A professional service is work a person does under a separate scope. Nothing here is a certification.

See the workflow on your protocol.

The conversation is about your study design and what the software can and cannot do today.