hh-health-ai/healthcare-equity/channel-check-prep

Channel-Check Prep

Convert footprint and KOL-site data into a ranked, MNPI-safe human-diligence target list for expert calls and field checks.

Workflow

1. Start from the thesis question — what claim needs human confirmation (adoption pace, usability friction, competitive displacement, reimbursement hassle, utilization reality) — and from footprint outputs (provider-footprint, site-adoption-tracker, kol-site-map) if available; run them first when absent.
2. Build the target list from the NPI Registry connector data: segment by role (evangelist KOL · high-volume community adopter · recent adopter · non-adopter/skeptic · former user) and by geography/site type. Deliberately include non-adopters and churned sites — confirmation-only lists are the classic channel-check failure.
3. Rank targets by information value: who can uniquely answer the thesis question × recency of direct experience × independence (trial investigators are informed but conflicted — label them). 8–12 targets with one line each: who, where, why them, the single question they can answer.
4. Frame the questions per usability/adoption drivers in ${CLAUDE_PLUGIN_ROOT}/references/usability-adoption-checklist.md (workflow fit, training burden, staffing, economics at the site level) — these are the answerable, non-MNPI questions that predict adoption.
5. MNPI guardrail (non-negotiable): questions must stay within personal experience and general practice patterns — never solicit enrollment numbers in blinded trials, unreleased results, or confidential contract terms. Flag any question that skirts the line and rephrase.
6. Hand off execution to healthcare-equity meetings-experts (EN-01 call prep, EN-02 debrief, EN-03 network plan); this skill's product is the who and why, that skill's is the how.
7. End with the EVIDENCE BRIEF block only when the list itself embeds findings (e.g., footprint concentration discovered en route); otherwise deliver the ranked list with sources and dates.