Life sciences

Evidence infrastructure for rare disease

We build the platforms that turn what families observe every day into data that holds up to a regulator, a payer or a sponsor.

See the case
From daily life to defensible data
  1. At home

    The caregiver records what they see, in the app or on WhatsApp.

  2. Consented

    Every use of the data has its own versioned consent.

  3. Traceable

    Nothing is deleted: every correction is recorded.

  4. Research-ready

    De-identified, coded extracts, exportable as FHIR.

The problem

The bottleneck is not the molecule, it is the evidence

For most of the 7,000+ rare diseases there is no natural-history baseline to compare against. Populations are small and scattered across the world. Outcome instruments rarely capture what a caregiver sees at home, day after day. And the durability of a one-time therapy has to be shown over years, in the gaps between clinic visits that happen twice a year.

The missing data does exist: it lives in notebooks, WhatsApp groups and the memory of families. What does not exist is the infrastructure that makes it structured, consented, traceable and interoperable. That is what we build.

What we build

Six layers, from the notebook to the extract

Longitudinal capture

Daily entry by the caregiver: sleep, mood, feeding, seizures, therapies, milestones. Designed so an exhausted family can do it in under a minute.

Granular consent

Separate, versioned consents, with re-consent and a record of every change. Consent is data, not a checkbox.

Immutable trail

Append-only audit log: errors are corrected by adding, never by mutating history. Built to ALCOA+ principles.

De-identification and disclosure control

k-anonymity, suppression of small groups, and a documented re-identification risk assessment.

Interoperability

Common data elements, ORPHA and HPO coding, FHIR R4 export and research-ready extracts. Data built to be joined with other people's data.

AI with human review

Agents that structure the caregiver's account and reduce the burden of recording. They never replace clinical judgement. AI prepares; a person decides.

Case

PittHope — Pitt-Hopkins syndrome

PittHope is a non-profit initiative supporting families living with Pitt-Hopkins syndrome, an ultra-rare genetic condition in which most children do not speak. Their experience reaches research only through the person who cares for them. Balexus built and maintains the platform.

83
families on the platform
19
countries, 5 languages
16
modules in production
1 in 3
queries arrive on WhatsApp

Platform figures as of 20 September 2026. The data belongs to the families and is governed by the non-profit; Balexus is the technology provider.

The most useful thing this project taught us: in Latin America one in three queries does not arrive through an app, it arrives on WhatsApp. If capture does not live where the family already lives, there is no data.
For your programme

What we can build for you

Registry and natural-history study

A platform with protocol, visit schedule, standardised instruments and research-ready export.

Consent and data governance

Consent architecture, traceability, de-identification and the technical framework for third-party data access.

Interoperability and extracts

A data model aligned to registry standards, FHIR export and documented extracts for research teams.

Long-term follow-up

Sustained capture for years after treatment, in whatever country the family lives in, for therapies that must prove durability.

Let's talk

Let's talk about your programme

If you are designing the evidence strategy for a therapy in a rare condition, or you need a patient community to generate data a regulator will accept, we can build that layer.

Or write to us directly at [email protected]