Double Cup Coffee
Health-Plan Impact · SDOH & Star Ratings · Illustrative Model

The Social-Need Intervention Report

Health plans are now measured on whether they act on their members' social needs, not just screen for them. Screening is the easy half. This report models where the harder half breaks down, and how a neighborhood coffee network gives a plan a concrete, trackable place to intervene.

HEDIS SNS-E
MA & Medicaid / D-SNP
Equity-Weighted Stars
How to read this report. Regulatory facts (measure structure, documentation rates, program rules) are cited and real. Every member-level figure, chart, and projected lift is synthetic and illustrative, built to show the analysis a pilot would run and the plausible shape of the result, not actual outcomes. Illustrative figures are tagged.
The Wedge
Under HEDIS SNS-E, a plan must deliver an intervention within 30 days of a positive social-need screen, and equity incentives are tilting Star Ratings toward its dual-eligible, low-income, and disabled members. Yet SDOH is documented on only 0.5–2.4% of encounters, and referral loops routinely go open-ended. Plans are under pressure to spend on social needs without a proven, measurable place to send people. Double Cup is built to be that place.
Real
SNS-E Accountability
Screen + Intervene ≤ 30 days
Six HEDIS rates across food, housing & transport: screening and a matched intervention.
Real
SDOH Documentation Today
0.5–2.4%
Share of encounters carrying any SDOH Z-code. The need is there; the record isn't.
Illustrative
Screen → Intervention Gap
~80%
Modeled share of positive social-need screens with no documented intervention in 30 days.
Illustrative
Reachable via a Connector Shop
2 in 3
Modeled positive-screen members living within 1 mile of a candidate Double Cup shop.
Panel 01 · The Intervention Gap

Screening Is the Easy Half

Every step from screening to a documented intervention loses people. The measure, and the money, lives at the bottom of the funnel, exactly where open-loop referrals break down. The lighter bars model what a ready intervention destination recovers.

Modeled · 5,000-member cohort
From Positive Screen to Documented Intervention
Two modeled paths for the same cohort. On the left, a typical open-loop referral, where most positive screens never reach a documented intervention. On the right, the same cohort when a Double Cup Connector shop is the destination and participation is logged back to the plan. The figure beneath each funnel is the share of positive screens that get an intervention within 30 days.

Modeled and illustrative. Hover any stage for its counts.

Panel 02 · What It Moves

The Measures a Connection Intervention Touches

Not every quality measure is in reach of a coffee shop. These are the ones where social connection is a plausible, literature-backed lever. The projected point gains are modeled and directional; the pilot exists to size them for real.

MODELED
HEDIS / Stars · modeled projection
Where Double Cup Plausibly Helps
Projected percentage-point gains in each measure's rate. Directional and illustrative; a pilot would size the real deltas.
The Whitespace
Isolation is a named health risk with no standard place to refer.

SNS-E screens food, housing, and transportation. Social isolation, which the U.S. Surgeon General called an epidemic, isn't its own screened domain yet, and its Z-codes (living alone, lack of social support) are documented on around 1% of encounters.

So plans have a recognized need, growing pressure to address it, and no off-the-shelf intervention built for it. That gap is the opportunity: Double Cup is a low-cost, community-based destination purpose-built for connection, that logs what it does.

Panel 03 · Closing the Loop

An Intervention That Reports Back

A referral only counts if the plan can see it happened. Double Cup is designed to complete the loop, turning a positive screen into a logged, documentable intervention the measure can credit.

01 · Screen
Positive social-need screen
Isolation or living-alone risk surfaces in an assessment or care call.
02 · Refer
Sponsored Double Cup benefit
Plan funds Club access; member is invited to a nearby Connector shop.
03 · Intervene
Participation happens
A check-in, a shared cup, a Connector trained to notice, invite, and refer onward.
04 · Document
Logged back to the plan
A timestamped, privacy-protected intervention record: the evidence SNS-E and Z-code capture require.

This is the difference between a resource list and an intervention: the plan can watch the 30-day window close. Documentation is aggregate and consented. The Club recognizes generosity and participation, never a diagnosis.

Panel 04 · How the Data Is Captured

Where Every Number Would Actually Come From

The metrics above only mean something if the capture is real. Three tiers, ordered by how much cooperation each needs: what Double Cup logs itself, what it attributes through a referral exchange, and what only the plan's claims can confirm.

Condensed view
Tier 1
First-party telemetry
Double Cup owns it · live at pilot start

Every Connector check-in and Double Cup redemption is a timestamped, consented participation event. That event is the intervention record: proof a social-need signal was acted on, exportable to the plan's care system.

Captured as ICD-10 Z-codes, billable service codes (G0136, CHI, PIN), and Gravity Project FHIR resources.
SDOH concern → ICD-10 Z-code

The Connector's structured note attaches the specific driver. A Z-code needs a documented risk or unmet need, not merely the circumstance, which the note is designed to supply.

Z60.2 living alone Z60.4 social exclusion Z59.811 housing instability Z59.00–02 homelessness Z59.41 food insecurity
Intervention → billable service code

So the plan's care team can document and bill the outreach that routed the member here. The PIN codes explicitly cover behavioral-health conditions.

G0136 · SDOH risk assessment G0019 / G0022 · Community Health Integration G0023 / G0024 · Principal Illness Navigation 96160 / 96161 · screening instrument
Carried on FHIR (Gravity Project)

Standards-based, so it reads as interoperable data to a plan rather than a bespoke feed.

Condition (+ Z-code) Procedure (intervention) Observation (screening) Goal
Tier 2
Attributed referrals
Shared via exchange · needs referral wiring

Two directions: where a member came from, and where the Connector sends them onward. Both are captured with shared identifiers, so nothing rests on memory or self-report.

Inbound source codes, an outbound FHIR referral closed over a CIE, and a consented member ↔ participant crosswalk.
Inbound → referral source

Each channel issues its own signed short-code or link, recorded on enrollment and every redemption. Every participant carries who sent them.

source_id campaign_id care-manager activation clinic QR nonprofit / employer / self
Outbound → BH clinic & other destinations

The Connector's "refer onward" opens a referral the receiving org acknowledges and closes, over a community information exchange. This confirms the handoff; the kept clinical visit is confirmed in Tier 3.

ServiceRequest (referral) Task (received → completed) Unite Us / findhelp / CIE
Identity crosswalk (the key)

A consented match of plan member ID to Double Cup participant ID. This single link is what makes Tier 3 possible; without it, nothing connects to outcomes.

member_id ↔ participant_id tokenized / deterministic match
Tier 3
Claims linkage & outcomes
Plan-side · needs data-sharing agreement + consent

The only tier that can see a BH clinic visit or a utilization change. Claims never flow into Double Cup. Double Cup sends exposure only, and the plan matches it against its own claims in a governed environment.

Exposure-only export matched to claims value sets (BH visits, FUM/FUA, AMM, ED/readmit) under a BAA.
What Double Cup exports

Nothing clinical: only who was exposed and when.

participant_id intervention timestamps
What the plan matches it to (claims value sets)
BH visit · CPT 90791/90792, 90832–90838 E/M w/ MH dx · 99202–99215 FUM / FUA · ED 99281–99285 → 7/30-day follow-up AMM · antidepressant fills (NDC) ED & readmit · revenue codes / DRG
Attribution method & governance

Index date = a member's first Double Cup intervention. Exposed members are compared to a propensity-matched cohort the plan pulls, or to the stepped-wedge rollout of the pilot itself. Reporting stays aggregate and de-identified under a BAA / limited data set.

Read top to bottom, this is also a build order. Tiers 1 and 2 are process metrics a pilot can stand up on day one. Tier 3 is the outcome question, and standing up its consent flow and data-sharing agreement is the pilot's real job, which is why every outcome figure in this report is still labeled modeled.

Panel 05 · Who & Where

The Referable Population, and Where to Put the Shop

The same data that flags who to screen also shows where the referable, equity-weighted population concentrates, and therefore where a Connector shop earns its keep. Both views below are illustrative, modeled on a Minnesota-shaped cohort.

Illustrative · isolation signal distribution
Sizing the Screen-and-Refer Population
Modeled isolation-risk score across 5,000 members. Higher tiers are the members a plan should be screening and, on a positive result, referring within 30 days.
Low
Moderate
Elevated
High: screen & refer
Illustrative · geographic concentration
Where a Connector Shop Matters Most
Dot size = members with elevated isolation in each modeled cluster. Bronze = dense referable population but no nearby Connector shop, the siting priority.
Panel 06 · The Intelligence Layer

Where Claude Adds Insight, and Where It Deliberately Doesn't

A Claude-powered API earns its place in the parts of this system that turn on language and judgment, and stays out of the parts that must be auditable or private. That discipline is the point.

Claude (Anthropic API)

Unstructured text into structure, real-time Connector guidance, referral reasoning, and plain-language synthesis. Language and judgment, always with a human in the loop.

vs

Classical models

Risk scoring, dormancy prediction, and outcome attribution. Calibrated, auditable, and owned separately, so a number a plan audits is never "the AI decided."

Flagship · Connector-facing

The Connector Companion

A Claude assistant inside the Connector's app that turns the "notice, invite, refer" training into live support during a real interaction. It works from the public local-resource directory and only what the Connector chooses to share in the moment, never the member's record.

Notice. Surfaces gentle, recovery-friendly openers and cues in the moment.
Invite. Suggests the right next step: an event, the Club, or a warm second cup.
Refer. Ranks nearby BH clinics and services from the resource graph, each with a one-line rationale.
Coach. Rephrases into trauma-informed wording and flags anything that belongs with a human or a crisis line.
Draft, don't decide. Every suggestion is a draft the Connector accepts, edits, or ignores.
Grounded. Answers cite the directory, so it can't invent a resource.
Capture assist

Note → structured codes

Converts a Connector's free-text note into the Z-codes, service codes, and FHIR resources from Panel 04, with a human confirming. Zero-retention: nothing stored or trained on.

Matching

Best onward referral

Reasons over a messy, non-uniform resource graph to rank the best clinic or service for an expressed need, with rationale, feeding the closed-loop referral in Tier 2.

Synthesis

Plain-language briefings

Writes the coffee-shop, sponsor, and health-plan narrative on top of the aggregate dashboards, refreshed on demand. A human signs off before anything reaches a payer.

Kept clear of PHI by design. The intelligence layer works from the public resource directory, de-identified aggregates, and what a Connector shares in the moment. It never touches the identity crosswalk or the plan's claims; Tier 3 stays in a governed environment with no model access, and where any member text is processed it runs under a zero-retention BAA with a human confirming the output. The auditable risk and attribution models stay separate, so "the AI decided" never enters a payer conversation.