Skip to main content Skip to main content

For Researchers

Real-world evidence, at population scale.

A patient-consented, longitudinal, multi-axis dataset normalized to FHIR R4 and OMOP CDM v5.4. Cohort discovery in minutes. IRB-protocoled extracts in days. Federated analytics for the questions a public dataset could never answer.

8 axes
Per-patient longitudinal
Physiological, neurocognitive, emotional, spiritual, relational, environmental, technological, purposeful — measured continuously, not at one visit.
FHIR R4 + OMOP
Native interoperability
Every record is exposed under FHIR R4 (USCDI v4) and dual-mapped to OMOP CDM v5.4. Vocabulary: SNOMED CT, ICD-10-CM, RxNorm, LOINC, ATC.
Patient-consented
Express, granular, revocable
No third-party data brokers. Every row was contributed by the human it describes, who can revoke at any time. We pay them in HCC for the use.

What's in

Every record we hold,
and where it came from.

Encounters
Every clinical encounter across our four clinic types — primary care, vet, optometry, pharmacy, dental — captured by AI Scribe at the point of care. Audio-transcribed, structured, ICD/CPT/SNOMED-coded, attending-attested.
Labs & vitals
LOINC-coded labs from any of the patient's connected systems, plus continuous vitals from connected wearables and the Guardian Orb. Time-aligned to encounters and to axis scores.
Medications
RxNorm + ATC. Active prescriptions, refill history, adherence (PDC) computed weekly, switch events, and pharmacy dispense records.
Imaging
DICOM headers + structured reads, including modality, body site, indication, impression, and assigned reviewer. Pixel data accessible inside enclave only, not export.
Master Equation telemetry
Per-axis scores at daily resolution — eight numbers per patient per day, plus the contributing signals. The signature dataset no other system has.
Patient-reported
Validated instruments (PHQ-9, GAD-7, AUDIT-C, PROMIS, SF-36) plus daily Orb check-ins. Free-text journals are excluded by default — opt-in only.
Genomics
For consenting subset: 23andMe-style array (BeadChip) and clinical exome — sized to the cohort, returned as VCF on the enclave, never exported.
SDoH & environment
Geo-aware exposure: air quality, neighborhood SES (ADI), commute, climate, food access. Coarsened to 5-digit ZCTA for de-identified releases.
Behavioral
Activity, sleep, screen-time, social-platform engagement from connected devices and the CH-Social bridge — opt-in, axis-tagged.
Outcomes
Hard endpoints — hospitalizations, ER visits, mortality (NDI-linked at the enclave), procedure outcomes, axis trajectories. The thing your published dataset doesn't have.

Why this dataset

Three things nobody else has.

There are dozens of HIPAA-compliant clinical data warehouses. There is exactly one of these.

01

The Master Equation

Eight axes of human health, scored daily, for every consenting patient. You can ask questions about emotional resilience and spiritual coherence as covariates of physiological outcomes — and have actual data to answer them. No public dataset has this.

02

Express patient consent

Every row was contributed by the patient who lived it, with informed consent at point of care. Revocable. Auditable. Defensible to any IRB. No third-party broker chain. No coverage entity-of-record uncertainty.

03

Native FHIR + OMOP

You don't write a six-month ETL. The data lands in your enclave already FHIR R4 + OMOP CDM v5.4. Vocabularies normalized. Nulls explicit. We publish the diff every release.

Pathways

Three ways to study with us.

Pick the right altitude. We'll route you to the right operations and pricing on apply.

Pathway A · Enclave query

De-identified, federated, your code in our environment.

For most observational research. You write SQL or R/Python in our analytic enclave; only aggregate, k-anonymized results leave. No row-level export. No PHI leaves the boundary. Fastest start.

  • Cohort discovery within minutes of approval
  • OMOP CDM v5.4 + custom CH schema (axes, equation telemetry)
  • Pre-installed: ATLAS, HADES, R, Python, Spark, dbt
  • Outputs reviewed for re-identification risk before release
IRB: required Time-to-first-result: 2–3 weeks
See sample queries →
Pathway B · Limited dataset

De-identified extract, exported to your enclave.

When you need full row-level data — for novel models, registries, or long-running studies. Requires DUA + executed Limited Dataset Agreement under HIPAA §164.514(e). Dates shifted; ZIP coarsened to 3-digit.

  • Limited dataset per §164.514(e), DUA-bound
  • FHIR Bulk Data export (NDJSON) or OMOP CDM dump
  • Date-shifted, geo-coarsened, identifier-stripped
  • Patient HCC payments roll up at the cohort level
IRB + DUA: required Time-to-first-result: 4–6 weeks
See IRB workflow →
Pathway C · Prospective study

Recruit consenting patients into your protocol.

For trials, longitudinal observational studies, and protocols that need new data collection. We surface your study to eligible patients in their wallet. They consent in-app. The protocol-data goes to you; HCC compensation goes to them.

  • In-app recruitment, e-consent, e-signature with audit trail
  • Patients are paid in HCC per protocol-defined milestones
  • Protocol amendments redrive consent automatically
  • 21 CFR Part 11 + ICH GCP-compliant
IRB + protocol: required Time-to-first-enrollment: 6–10 weeks
Talk to us →
Pathway D · CH-as-control-arm

Synthetic control from CH for your interventional trial.

Pre-screened, propensity-matched, real-world control arms drawn from CH. Useful for rare-disease trials, post-market commitments, and external-control filings to FDA. We'll work with your stat plan.

  • Matched on baseline + axes, not just demographics
  • Documentation suitable for FDA RWE submission
  • Causal inference packages pre-installed in enclave
  • Co-author or work-product, your choice
IRB + sponsor: required Time-to-first-result: by engagement
Talk to us →

Cohorts ready today

Eleven cohorts you can query in week one.

Sample sizes shown are current-platform totals across all clinic types as of the most recent enclave refresh. Each links to a worked sample query you can adapt.

Cohort n (consented) Axes covered Median follow-up Sample query
Adult primary care, longitudinal7,210All 814 moPHQ-9 trajectory by NM bracket →
Type-2 diabetes, A1c tracked1,840PO · NM · ER · ES · PV11 moA1c × Purposeful Vitality →
Hypertension, BP stream2,360PO · ER · ES · TA13 moBP variance × ER score →
Major depression on SSRI680NM · ER · SC · RS · PV9 moResponse × Spiritual Coherence →
GLP-1 agonist initiators410PO · ER · NM · PV7 moGLP-1 weight × axis lift →
Pediatric ADHD320NM · ER · RS · PV10 moStim response × RS network →
Postpartum (within 12 mo)240PO · ER · NM · RS · SC8 moPPD onset × RS support →
Adults with smartwatch HRV3,140PO · ER · TA13 moHRV × ER trajectory →
Companion-animal owners1,210RS · ER · PV15 moPet-loss × ER drop →
Wildfire-exposure ZCTAs580PO · ES · ER · NM12 moPM2.5 × respiratory + ER →
Faith community-attached1,930SC · RS · PV · ER14 moSC × all-cause health →

Sample query · enclave

In our enclave, this returns in under three seconds.

"Among adults with type-2 diabetes, does Purposeful Vitality predict A1c reduction independent of BMI?"

OMOP CDM joined to the CH axis-telemetry table. Output is k-anonymized aggregate (k≥11). The query, the result, and the derivation are all logged to the chain — your IRB will love this.

Twelve more like it on the queries page →

-- adult T2DM, paired baseline + 6mo A1c
WITH cohort AS (
  SELECT person_id, condition_start_date AS dx
  FROM omop.condition_occurrence
  WHERE concept_id = 201826  -- T2DM
    AND person_age >= 18
), baseline AS (
  SELECT c.person_id, AVG(a1c) AS a1c_0
  FROM cohort c JOIN omop.measurement m
    ON m.person_id = c.person_id
   AND m.measurement_date BETWEEN c.dx - 30 AND c.dx + 30
   AND m.concept_id = 3004410  -- A1c
  GROUP BY c.person_id
), pv AS (
  -- CH extension: daily axis scores
  SELECT person_id, AVG(pv_score) AS pv_avg
  FROM ch.axis_daily
  WHERE axis = 'PV'
    AND date BETWEEN dx AND dx + 180
  GROUP BY person_id
)
SELECT pv_quartile, AVG(a1c_delta), k_anon(person_id)
FROM joined GROUP BY pv_quartile;

Trust posture

Your IRB will recognize the names.

All access is gated through the same compliance frameworks our clinical operations sit under. Documents are downloadable from the compliance hub; gated detail through the regulator portal.

HIPAA
Privacy + Security Rules
SOC 2 Type II
Annual independent audit
HITRUST CSF r2
Certified, verifiable
21 CFR Part 11
e-records / e-signatures
ICH GCP
For Pathway C trials
GDPR / UK-DPA
For non-US researchers

A founder's note

"You will see real outcomes in real people. We don't sample bias; we don't sell to brokers; and we will never give you a result we wouldn't want a patient to read. If your hypothesis holds up here, it holds up."

— Maria R. Lahti, MD · Chief Medical Officer · Conceptual Health

Apply for an institutional account.

You'll be paired with a named research liaison inside one business day. Sandbox access is granted in 48 hours; production data follows IRB approval and DUA execution.