Digital health platforms
Add a clearer understanding of each case without replacing the product layer above it.
PI Health
A clearer way to understand changing health data.
PI Health helps pull different health signals about one person into one view so teams can see what changed, what may be causing it, what may deserve attention first, and when the evidence is too weak to say more.
Looking for how this is evaluated in practice?
Featured pathway
A bounded first pilot in women’s health.
The clearest near-term starting point is a small retrospective midlife pilot that uses existing data to support review, not to replace it.
Pilot outline
- 6 to 10 retrospective cases
- Existing longitudinal data only
- No live triage
- No diagnosis or autonomous recommendations
- 4 to 8 weeks
Why health first
Health is where mixed signals matter most.
PI Health is strongest where home data, symptoms, labs, and events do not fit neatly into one score or one threshold, and where within-person change matters more than a generic average.
Intended role
- Built for within-person review, not population scoring
- Helps teams review complex cases
- Looks at change from a person's usual baseline
- Shows what may be causing the current picture
- Shows what may be helping versus hurting
- Shows where several factors may be working together
- Keeps early findings separate from stronger confirmation
- Shows weaker points, less obvious levers, and what matters most
- Only surfaces permitted action when the governed lane clears
- Stays cautious when the evidence is thin
Where it fits
PI Health works best inside other products and workflows.
The best near-term fit is inside digital health, clinical, or monitoring products that already handle workflow and engagement.
Clinician-facing tools
Support review where the data is mixed, scattered, or hard to read quickly.
Research workflows
Study repeated within-person trajectories without flattening the individual story into one average curve.
Partner pilots
Test one clear workflow problem before making broader claims.
Evidence direction
The strongest proof so far is in health.
Evidence from the McPhases dataset
PI was evaluated on the McPhases Women’s Health dataset across real within-person participant histories, with staged reasoning, continuity-aware rollups, and a cohort-level evidence map.
Evidence from a MIMIC-based cohort
PI was also applied to a deidentified inpatient cohort, where it processed 100 users, carried 37 within-person cases through full Stage 3 review, and let the active inpatient story emerge across domains before action.
Read the evidence briefEvidence from a Garmin-derived case
PI was applied to one within-person wearable and metabolic case where fasting glucose stayed the anchor target, recovery strain shaped the broader picture, and governed action remained gated.
Read the evidence briefClearer reasons to watch
PI is a fit for settings where review burden is high and the information is spread across many sources and teams need a clearer reason to pay attention.
Separate evidence briefs are now available for women’s health, a MIMIC-based inpatient cohort, and a wearable metabolic case. Additional cohort briefs can follow as new runs clear review.
Current posture
Ready for the right kind of conversation.
- Internal and partner use
- Pilot and research discussions
- Sub-layer use inside existing health products
- Clinical review and support exploration
What is not claimed
- Autonomous triage
- Autonomous diagnosis
- Unbounded treatment recommendation
- Finished public clinician-grade product
PI’s health story is stronger when it remains disciplined about intended purpose and non-claims.
Health conversations
Start with the women’s health pilot path.
The best first discussion is usually narrow: one workflow, one care pathway, or one bounded pilot question.