RetentionHealth
Detects patient drop-off risk without relying on personal data.
The Problem
Most systems are built around individuals. They track identities and store personal data, but they do not define how risk actually emerges. In practice, patient drop-off is not an individual event; it is a pattern. Signals appear early—engagement drops, concerns increase—but those signals are fragmented across systems. When everything is tied to identity, detection becomes slow, noisy, and expensive from a compliance standpoint.
The System
We did not build a tracking system. We defined a signal system: a cohort-based model that detects risk without relying on identity or personal data. Instead of asking who is at risk, the system asks which patterns indicate risk. Detection is based on structured behavioral signals, deterministic evaluation, and cohort-level modeling. No identity tracking. No personal data storage. Risk is computed, not inferred.
How It Works
INPUT → LOGIC → EXECUTION → OUTPUT
INPUT: Behavioral events are captured as raw system activity. LOGIC: Meaningful signals are extracted, grouped into cohorts, and evaluated through deterministic risk scoring. EXECUTION: Cohort-level risk states are computed and continuously monitored. OUTPUT: Clinics get early, actionable visibility into drop-off risk before disengagement becomes irreversible, without handling personal patient data.
What It Governs
How drop-off risk is detected through cohort-level signal patterns without identity dependence.
System Definition Coverage
Inputs
Anonymous behavioral events, cohort assignment data, and clinic-defined monitoring windows.
Constraints
No-PHI handling, identity-independent processing, and deterministic scoring boundaries.
Decision Logic
Signal extraction and weighted cohort scoring classify risk states from pattern behavior.
State & Flow
Event capture -> signal extraction -> cohort aggregation -> risk scoring -> dashboard update.
Outputs
Cohort-level risk visibility, early warning indicators, and intervention-ready system insights.
Validation
Aggregation checks, score reproducibility checks, and compliance-safe data handling verification.
Result
Better retention without increased regulatory burden.
Past Builds
Explore more projects from our studio.
