Patient Activate
Rare Disease Patient Finding & Activation Platform
Overview
Patient Activate connected precision patient finding, compliant engagement, and downstream progression measurement—converting ~500 diagnosed patients and surfacing ~300 treating physicians across 5 rare programs in under a year, with consent-based handoff to pharma brand partners for on-therapy transition
Goal
Linked patient finding, compliant engagement, and downstream progression measurement in one operating system—giving pharma teams a way to surface diagnosed patients, identify treating physicians, and measure consent-based patient handoff to on-therapy transition rather than registration volume alone.
Team
Led strategy and a cross-functional delivery team of 7—analytics, engineering, omnichannel, and CRM. Primary delivery client-facing lead across 5 pharmaceutical partners from scoping through <8-week launches.
Role
Primary client-facing lead across strategy, analytics, engineering, omnichannel, and CRM from scoping through launch.
Problem
5–7 year diagnostic delays, <5,000-patient prevalence, and 4–8+ week PRC cycles made rare-disease activation structurally hard
Patients often see 6+ physicians before diagnosis, while low specialist access, sparse signal density, and slow regulatory cycles make standard awareness and targeting models too blunt to drive measurable progression.
5–7 year diagnosis gaps created the need
- Patients often waited 5–7 years and saw 6+ physicians before correct diagnosis*
- Standard campaigns created awareness, not progression—which shaped our decision to build a staged qualification model rather than optimize for registration volume
*Timmers L, et al. Time to diagnosis and determinants of diagnostic delays of people living with a rare disease. European Journal of Human Genetics. 2024.
<5,000-patient prevalence made targeting harder
- Rare populations were often under 5,000
- Diagnosed-only claims targeting missed symptomatic and misdiagnosed patients
- Claims lacked behavioral and intent signals
4–8+ week PRC cycles made learning slower
- Regulatory review routinely took 4–8+ weeks
- Multi-round approval limited iteration speed
- Claims substantiation increased production burden
- This forced architectural separation of configuration from creative review
*PRC = Promotional Review Committee: the cross-functional legal, regulatory, and medical review process required to approve healthcare promotional materials before launch.
Low-signal conditions made optimization unstable
- Rare disease programs generated too little event volume for standard platform learning
- Over-optimization risked stalling delivery
- We chose delivery stability over algorithmic optimization—building configurable logic a non-technical team could adjust without engineering intervention
Solution
Enabled ~8-week launches with 100%-CMS-configurable logic by separating configuration from PHI
An early integration tied booking confirmation directly to CRM entry—until ~50% API failure rates in Zoho Bookings forced a redesign. Decoupling qualification from booking, with Zoho as protected system of record, made intake reliable at scale and let a non-technical team launch new programs without engineering intervention.
CMS-controlled program configuration
Segment-specific qualification logic
Dynamic form builder
Program-specific booking integration mapping
Consent Engine
Analytics event capture
System Architecture
Configurable Runtime Layer
All behavior controlled via CMS inputs
Anonymous Optimization and Attribution Layer
Protected PHI System
Restricted system access
Sensitive data flowed directly to Zoho, while configuration and optimization stayed outside the protected system.
Scale
Validated repeatable activation across 5 programs, including ~5% penetration in a rare neurological condition in <1 year
Programs remained active for 12+ months and generated qualified patient and HCP yield across populations as small as 1,200—proving the platform could sustain performance where standard activation models don't operate—let alone perform.
| Program | Prevalence | Months | Diagnosed Patients | HCP Leads |
|---|---|---|---|---|
| Rare neurological condition | ~3,000 | 12 | ~150 | ~250 |
| Rare kidney disorder | ~1,200 | 5 | ~140 | ~10 |
| Rare blood disorder | ~4,000 | 12 | ~70 | ~15 |
| Rare brain tumor | ~2,500 | 4 | ~50 | ~10 |
| Rare metabolic bone disorder | ~16,500 | 12 | ~30 | ~15 |
← Swipe to see more →
Intelligence
The Disease Intel Canvas cut cost per identified patient 41% while increasing diagnosed patients ~3x and HCP handoffs ~7x
Upstream of Patient Activate, the Disease Intel Canvas translated structured disease and behavioral signals into reusable patient-finding and planning inputs—compressing the per-launch rebuild that would otherwise have required ~$139K of external support per program.*
*Based on external agency benchmarks for rare disease launch strategy, targeting development, and creative production
Agentic Research & Strategy Generation
Transformed structured disease inputs, search data, and prior-program learnings into reusable audience hypotheses, keyword themes, and channel strategy.
Generative Asset Production
Turned approved strategy into compliant, channel-ready copy, visual prompts, tagged content, and modular creative inputs for launch.
Programmatic Learning Loop
Fed campaign results back into keyword scoring, audience refinement, prioritization, and optimization recommendations to improve future launches.
Disease Intelligence Workflow
View the full operating model behind strategy generation, asset production, and optimization inputs.
Measurement
Optimization shifted toward qualified progression, with best-program cost per diagnosed patient reaching ~$800–$1K
By instrumenting anonymous diagnosis-stage visibility across BigQuery and CRM, the system enabled ~60–85% registration-to-diagnosis conversion in the strongest programs and reduced cost per identified patient from ~$1K to ~$780 over 4 months.
Traditional measurement optimized the wrong signals
- Click-through rates did not reflect clinical relevance
- Registrations did not prove diagnosis progression
- Media drifted toward the cheapest top-funnel event
- So we rebuilt attribution to weight events by clinical relevance—not volume—which is what shifted spend toward diagnosed patients instead of clicks.
We built a staged progression model
- Activation → Qualification → Guide Contact → Dx Confirmation → Client Connection
- Measurement aligned spend to patient progression, not traffic alone
Tooling connected media to downstream outcomes
- Event normalization in BigQuery
- CRM-level diagnosis + HCP status integration
- Source-level attribution across funnel stages
- Conversion weighting by clinical relevance
Our first pipeline combined 4 data sources in a single query—elegant, but too fragile to run reliably. We rebuilt as staged transforms. Less clever, actually reliable.
Optimization improved toward qualified progression
- ~$800–$1K best-program cost per diagnosed patient
- ~60–85% registration-to-qualified conversion in strongest programs
- ~$1K → ~$780 Rare brain tumor cost per diagnosed patient over 4 months
Artifacts
The platform scaled through shared system logic and disease-specific assets
Each program translated common platform capabilities into tailored content, UX flows, and operational outputs shaped by disease burden, sentiment, and diagnostic context.
Acquisition Creative
Qualification/Content Surface
Booking Flow
Program Management Dashboard