Care Journey & Patient Value

The idea in one line

Take the financial lens a consumer marketer uses to run a business — lifetime value, acquisition funnel, retention/churn, share-of-wallet, expansion revenue — and apply it to a Medicare patient relationship, reconstructed empirically from claims. Everyone else's "patient journey" is a workshop whiteboard. Ours is how thousands of real beneficiaries actually moved through care, with real dollars on every step.

How to read this prototype. Two screens, one engine. ① Care Journey Map is the time view — how patients move. ② Revenue Stack is the money view — what gets monetized, and by whom. Both are projections of a single per-patient claims event stream (see Data Model). Dotted-underlined terms and the info dots carry definitions throughout.

The translation that makes it click

Consumer-marketing concept	Medical equivalent	Data readiness
Customer journey / funnel	Observed care journey — the service sequence a patient walks over time	Built · #433
Customer Lifetime Value (CLV)	Patient LTV — revenue across the full care relationship	Built · #432
Cohort retention / churn	Active / stopped-but-living / deceased + length-of-service	Built · #432
Share of wallet	Share of care — what one player captures vs. what leaks to others	Net-new · hard
Expansion / cross-sell revenue	Revenue stack — devices, drugs, virtual visits, referrals beyond the core service	Preliminary · #430/#431
Vertical integration / make-vs-buy	Control vs. partner across the care value chain	Framing layer

Lost to referral leakage

10–30%

of potential revenue (industry est.)

Complex CCM = the spine

3–5×

the monthly revenue/patient of any other service

Scope

Market-general

works for any care-mgt company; ACCESS is one slice

Positioning fork (needs David). Lean market-general: present as a standalone "care economy intelligence" lens, lead the ACCESS demo with the ACCESS slice. The engine reads claims either way — it's a positioning call, not a data one.

① The Observed Care Journey Map

How do real patients actually move through this company's care — where do they enter, branch, stall, and drop off — and where does revenue accumulate along the path? Built from the claims-derived transition matrix in GHI #433.

Entity Encode flow by

Care journey flow

Phases left-to-right in clinical order. Switch the encoding toggle to Revenue and the same flow re-weights toward Complex CCM — the money view of the identical patient paths.

Block height ∝ patients reached Onboarding Complex CCM (spine) Ongoing care-mgt Transition — width ∝ patients; fades source→destination phase color 0.0 mo avg time to next phase Drop-off / stalled

The insight here Solace doesn't run a clinical pathway — it runs a revenue assembly line, and Complex CCM is the engine. Reconstructed from claims, ~40% of patients move through one near-deterministic funnel — E&M → SDOH → CCCM — and when you flip to the revenue view, that same funnel is where the dollars pool (CCCM bills ~3–5× any other service per patient/month). Two things fall out that you can act on:

The drop-off between care-plan and monthly CCM is recoverable money — patients who got onboarded but never started the billable monthly service. Mortality you can't fix; stalled onboarding you can.
The shape is a fingerprint. Point it at Pairtu and the assembly line collapses to a single E&M → PIN step — a prospect sees in one glance how a competitor's model differs from theirs.

Care phase key

What each acronym in the flow means. These come from the care-management code families in dim_care_mgt_hcpcs (#433).

How to read it. Each column is a care phase. Ribbon thickness = patients (or dollars) flowing that path, and each ribbon fades from its source-phase color to its destination-phase color (the same colors as the phase-key dots). The small number on each ribbon is the typical time (avg months) before patients advance to the next phase — hover a ribbon for its patient count and share. The red taper at each node is drop-off — patients who reach that phase but don't advance. Flip the toggle to Revenue and watch the flow re-weight toward Complex CCM: same patients, but now you see where the money is made.

What the flow exposes (Solace)

A product pipeline, not a clinical path. ~40% of patients flow through E&M → SDOH Assessment → CCCM → … — an operational onboarding funnel built on G0136 → 99487/99489 layering.
A fixable leak: the patients who get an SDOH/care-plan step but never start monthly Complex CCM. Stalled onboarding is recoverable revenue; mortality is not. The map separates the two.
Pace matters. SDOH → CCCM takes ~1.6 months on average; CCCM → CHI ~3.3 months. Slow transitions are where patients fall out.
Comparison is the kill shot. Switch to Pairtu: a single E&M → PIN path, 95% less stacking. A prospect instantly sees how a competitor's model differs from theirs — GHI #432's literal stated purpose.

② The Revenue Stack & Value Chain

Across everything a patient generates in billable care, how much does this player capture today — and what are the adjacent rungs they could own, build, partner, or cede? Each bar splits captured (solid) from whitespace (hatched).

Altitude

Revenue stack — captured vs. whitespace

Posture: OWN already captured · BUILD whitespace worth vertically integrating · PARTNER whitespace better captured via a collaborator · CEDE not worth chasing.

Captured today Whitespace

The insight here Solace dominates the rung it's on — and that rung is a small slice of what its own patients are worth. Care management is ~95% captured and firmly owned. But the same panel generates multiples of that revenue in adjacent rungs — RPM/RTM, devices, drugs/GLP-1, downstream referrals — that flow to other players or go unclaimed. The growth isn't in the rung you already win; it's in the posture you take on the rest:

The biggest dollars say partner, not build. The two largest whitespace pools — drugs/GLP-1 and specialist referrals — are exactly the ones hardest to own (cross-entity, regulated, different capability). Chase them through a collaborator, not a build.
The buildable adjacency is the near-term play. Telehealth, RPM/RTM and devices sit right next to what Solace already does — modest each, but a credible vertical-integration path that #430/#431 are already scoping.

The strategic story. 2026 is a wave of healthcare vertical integration (payers + PBMs + specialty pharmacy + care services consolidating). "Which links of the chain do I control vs. collaborate on?" is exactly the decision our customers are making right now — this screen hands them the quantified map. The biggest unclaimed rungs here — drugs/GLP-1 and devices — are precisely what vinci #430/#431 are scoping.

Data Model & Gap Analysis

Both screens are projections of one object — a per-patient claims event stream with revenue attached. Most of it already exists in Christian's vinci notebooks. This is what's built, what's preliminary, and what's net-new.

The engine: one per-patient event stream

Object	Grain	Key fields	Status
`bene_journey_events` the engine	one row per bene × service	`bene_id`, `entity`, `service_date`, `service_yearmo`, `hcpcs_cd`, `care_mgt_cat`, `family`, `phase`, `allowed_amount`, `seq_index`, `days_since_prior`, `served_by`	Net-new join components exist separately

Dimension tables (reusable building blocks — exist today)

Table	Maps	Source
`dim_care_mgt_hcpcs`	HCPCS → `care_mgt_cat` / phase / family (CCM, CCCM, CHI, PIN, BHI, SDOH…)	`care_mgt_hcpcs_lookup_table.ipynb`
`dim_access_track_diag`	diagnosis → eCKM / CKM track patterns	used in `eckm_ccm_rpm_stack.ipynb`
`dim_zip_geo_mapping`	ZIP → geo / market	shared dim
`carrier_claimlines_clean_2025`	cleaned claim-line source (allowed amount, NPI, bene, date)	pipeline output

Projections & gap analysis

Built in a vinci notebookPreliminary started, not solidNet-new doesn't exist yet

Artifact (table / output)	Powers	Status	Source / what's missing
`journey_transition_matrix` from→to, benes, %_of_from, median_gap_mo	Journey Map ribbons + timing	Built	`pin_landscape_by_entity_time_period.ipynb` · #433 (Steps 4c–4f)
`journey_sequence_patterns` first-appearance sequence, benes, %	Journey Map entry/top-paths	Built	`SEQUENCE_ANALYSIS.md` · #433
`bene_revenue_segment` segment (deceased/active/stopped), months_of_service, first/last_service, avg_monthly_revenue, total_allowed	LTV, churn, drop-off labels	Built	#432 (Solace revenue calc)
`revenue_by_hcpcs` / `by_care_mgt_cat`	Revenue-mode flow weighting; CCCM spine	Built	#432
`rpm_rtm_stack` billing_npi × bene × yearmo, is_stack, annual rev	Revenue Stack — RPM/RTM rung	Built	`eckm_ccm_rpm_stack.ipynb`
`device_rung`	Revenue Stack — devices	Preliminary	#430 — example data only; needs $ per bene by claim type (DME/prof/SNF/HH)
`drug_glp1_rung`	Revenue Stack — drugs/GLP-1	Preliminary	#431 — prescriber counts only; needs Part D spend attribution per bene
`bene_journey_events` (unified)	Both screens (the engine)	Net-new	Join #432 revenue onto #433 sequence into one persisted stream
`revenue_stack_rung` rung, captured_amt, whitespace_amt, posture	Revenue Stack — captured vs. whitespace	Net-new	Needs total-billable-care denominator per bene (all claims, not just entity's)
`share_of_care` / referral leakage	Share-of-care / leakage (deferred screen)	Net-new	Cross-entity claims + entity resolution — the hard one. Not v1.
`served_by` dimension	Shared with the competition lens	In progress	Being built in competition work — reuse, don't fork
LTV projection (forward)	Cohort-economics screen (deferred)	Net-new	Foundation in #432; survival/projection logic TBD

Bottom line. The two screens in this concept stand almost entirely on Built work — journey sequencing (#433) and per-bene revenue/segmentation (#432). The net-new lift is (1) joining them into one event stream and (2) the total-billable-care denominator that turns "captured" into "captured vs. whitespace." The device/drug rungs and share-of-care are explicitly downstream.

Glossary

Term	Definition
Patient LTV	Lifetime value — total revenue a patient generates across the full care relationship. The medical version of customer lifetime value.
Care journey	The observed, time-ordered sequence of services a patient receives, reconstructed from claims (not a workshop diagram).
Care phase / family	A group of HCPCS codes playing the same role: E&M Office, SDOH Assessment (G0136), Complex CCM/CCCM (99487/99489), Illness Navigation/PIN (G0023/24), Community Health Integration/CHI (G0019/22), Behavioral Health Integration/BHI.
Segment	Mutually-exclusive bene status (#432): deceased (date of death) → active (billed in trailing window) → stopped-but-living (= churn; capture length-of-service).
Transition matrix	For each phase, the share of patients moving to each next phase + median time gap. The Markov edges behind the Journey Map.
Revenue stack	The layered set of revenue rungs around a patient: care-mgt, E&M, virtual visits, devices, drugs, labs, referrals, plus ACCESS/RPM/RTM payment programs.
Captured vs. whitespace	Captured = what the target entity bills today. Whitespace = the rest of the patient's billable care (addressable adjacency).
Posture (own/build/partner/cede)	The strategic call per rung: already have it / vertically integrate / collaborate / walk away.
Share of care	Medical share-of-wallet: the fraction of a patient's total billable care one player captures vs. loses downstream.
Referral leakage / keepage	Revenue lost when patients get care outside the network. Industry estimates: 10–30% of system revenue.

Concept doc: design/screens/care-journey/care_journey_value_concept.md · Data engine: vinci GHIs #430–433 · Worked example: Solace.