Research Roadmap
How ASH is building toward 10,000 articles of open clinical knowledge infrastructure.
The ASH mission is to build an open clinical knowledge & safety operating system for hirudotherapy. We commit to 10,000 articles of verified, structured, and openly auditable content. This page is the public roadmap: where we are, where we're going, and where the current bottleneck is.
All counts on this page are live — they read directly from the structured registry files at render time.
Tier A FDA-cleared conditions: 3 · Tier B off-label: 10 · Tier C investigational: 186
Current foundation
The structured registries that anchor every page on hirudotherapysociety.org.
199
Conditions
FDA-cleared, off-label & investigational — Tier A/B/C
201
Compounds
Salivary pharmacology: hirudin, eglins, destabilase, more
169
RCTs
GRADE-rated trials, PMIDs verified against PubMed
68
Biographies
Investigators across antiquity through modern molecular biology
155
Jurisdictions
Global regulatory survey — honest 'unregulated' markers
10,030
Research articles
PMID-verified library — target 10,000 (long-term roadmap)
10,000 target reached ✓
Roadmap phases
Phase 1 — Foundation
2025 Q2 – 2026 Q1CompleteCore architecture, regulatory framing (Architecture 4.0 ClaimBoundaryBadge), 5 structured registries seeded, ~449 imported research articles, 36 i18n namespaces × 4 locales.
- Architecture 4.0: ClaimBoundaryBadge across all 160+ content pages
- 5 registries seeded: condition, compound, RCT, biography, jurisdiction
- FDA framing corrected (K040187, NRN, Unclassified Pre-Amendment)
- Editorial integrity: 3 fabricated PMIDs caught and corrected (May 2026)
- PageMeta + safety scaffolding (SafetyWrapper, EvidenceGradeBadge)
Phase 2 — Depth (mostly complete)
2026 Q2 – Q3CompletePer-page content depth: every condition page gets PatientSummaryBox + WhatThisDoesNotMeanBox + KeyTrials. Registry expansion targets met or exceeded across the board.
- Condition registry 131 → 199 (target 200+, 99% met)
- Compound registry 81 → 155 (target 130+, exceeded 19%)
- RCT registry 42 → 109+ (target 100+, exceeded)
- Biography registry 39 → 68 (target 80+, 85% met)
- Jurisdiction registry 68 → 107 (target 120+, 89% met)
- Condition PatientSummaryBox backfill 199/199 (100% complete) 🎯
- PMID Integrity Sprint: RCT + condition + compound + biography registries 100% PMID-verified against PubMed (Quality Cycles 06-09)
- i18n quality: 0 missing keys across RU/ES/DE for 36+ namespaces
Phase 3 — Scale (current)
2026 Q4 – 2027 Q2In progressSystematic PubMed import past 10% milestone (10.83% crossed). Three public JSON API endpoints live. Automated freshness monitoring, public correction queue, MeSH-aligned tagging in flight.
- research-articles-imported.ts: 1,433+ verified entries (14.33% of 10k goal — 14% milestone CROSSED 2026-05-27)
- 8 public JSON APIs (site-stats, changelog, research-articles, conditions, compounds, biographies, jurisdictions, rcts) — CDN-cached, CORS-enabled, no-auth, hypermedia discovery
- Structured errata workflow live: /submit-correction with SLA table (24h-30d by category)
- MeSH-aligned tagging on every entry (in progress)
- Automated PMID re-verification cron (weekly) — next
- Public correction queue: /corrections (audit log) + /submit-correction (submission) both live
- Cochrane systematic review tracker — /publication-tracker live
- ICMJE-aligned editorial transparency — /transparency-and-trust live
Phase 4 — Knowledge Operating System
2027 Q3 – 2028Planned10,000-article target. Cross-registry knowledge graph, API access for researchers, structured submission for new entries, downstream MedlinePlus-style integration.
- 10,000 articles in research-articles-imported.ts
- Public knowledge graph API (REST + GraphQL)
- Researcher submission portal with peer-review queue
- MedlinePlus / NLM linkage candidacy
- Open data download (JSONL, BibTeX, RIS)
- Aligned with FDA real-world-data initiatives
Live: Publication Tracker
The roadmap above sets ambition. The Publication Tracker shows the live state: every PMID-verified article in our library, grouped by year and category, with explicit progress toward the 10k goal.
Open the Publication Tracker →Current bottleneck: editorial verification depth
Every PMID added must be verified against PubMed (caught 3 fabricated PMIDs in May 2026). Every condition entry requires PatientSummaryBox + WhatThisDoesNotMeanBox + KeyTrials. This is intentionally slow — the alternative is scale at the cost of accuracy, which fails the user.
Throughput today: ~20 verified condition entries per editorial session (via parallel sub-agent verification). To reach 10,000 articles: ~500 sessions, spread across phase 3-4. Volunteers via the Contact & Transparency page can accelerate this materially.
How to contribute
Clinicians
Submit case series, share institutional protocols, propose new condition entries from clinical practice. Practitioners with hospital-based hirudotherapy programs are especially valuable.
Submit a clinical contribution →Researchers
Suggest RCTs, share unpublished results (with IRB clearance), help build the 10,000-article import workflow. Library/MeSH expertise welcomed.
Researcher portal →Translators
Medical-grade translation pass on RU / ES / DE / KO / VI / ZH-CN locales. Each namespace needs editorial review by a native-speaking clinician.
Translation volunteer →