Natural Language Analytics Agent for Real-World Clinical Evidence
Built for the AI and Big Data Expo — Intelligent Enterprise Solutions Hackathon · May 11–19, 2026
ClinicalMind lets clinicians, researchers, and healthcare analysts ask clinical questions in plain English — and get instant SQL, results, visualizations, and publication-ready clinical interpretations.
No SQL knowledge required. No data science background needed. Just ask.
"Do CRC patients with depression have higher 30-day readmission rates?"
→ Generates OMOP SQL → Executes → Chart + Table + Clinical Interpretation
→ Exports submission-ready Table 1, Table 2, Figure 1 (JAMA/Lancet style)
- 🔍 Natural Language to SQL — Ask in plain English, get validated DuckDB queries
- 🏥 Multi-Source Simulation — Hospital EHR, Pharmacy Claims, Insurance Claims, Population Registry
- 📊 Instant Visualization — Bar charts, results tables, clinical interpretation
- 📄 Publication-Ready Outputs — CONSORT Figure 1, JAMA-style Table 1 & Table 2 with p-values, HRs, 95% CIs
- 🌐 Zero Infrastructure — Fully offline-capable single HTML file
- 📚 Journal-Grounded Questions — Research questions derived from NEJM, JAMA, Lancet, Lancet Psychiatry, Acta Oncologica 2023–2024
| Source | Organization | Standard | Patients | Focus |
|---|---|---|---|---|
| 🏥 Hospital EHR | St. Mercy General, Chicago | OMOP CDM v5.4 | 347 | Readmissions, LOS, mortality |
| 💊 Pharmacy | MedRx Network, Nationwide | NCPDP D.0 | 1,240 | Adherence, polypharmacy, refill gaps |
| 🛡 Insurance | BlueCross Shield National | X12 837/835 | 8,420 | Costs, claim denials, ER utilization |
| 🌐 Registry | National Cancer Registry | SEER/NAACCR | 48,750 | Staging, survival, geographic trends |
ClinicalMind was built on real dissertation research:
- Domain: Colorectal cancer (CRC) + psychiatric comorbidities (depression, anxiety)
- Dataset: OMOP CDM v5.4 (TriNetX / OHDSI-compatible)
- Methods: Cox proportional hazards model, Mantel–Haenszel adjustment, CONSORT reporting
- Key Finding: CRC patients with comorbid MDD show 25% excess mortality (HR 0.48, 95% CI 0.05–4.61) — consistent with Acta Oncologica 2024 meta-analysis
| Layer | Technology |
|---|---|
| AI Agent | Claude claude-sonnet-4-20250514 (Anthropic) |
| Query Engine | DuckDB (in-memory, OMOP CDM v5.4) |
| Backend | FastAPI + Python + Uvicorn |
| Frontend | Vanilla HTML/CSS/JS + Chart.js |
| Standards | OMOP CDM, NCPDP D.0, X12 837/835, SEER/NAACCR |
| Deploy | GitHub Pages + Custom Domain |
clinicalmind/
├── index.html # Full demo app (offline, single file)
├── backend/
│ ├── app/
│ │ ├── main.py # FastAPI app with CORS + static serving
│ │ ├── api/routes.py # /api/query, /api/schema, /api/health
│ │ ├── agent/nl_to_sql.py # Claude NL→SQL agent with retry loop
│ │ └── db/connection.py # DuckDB + mock OMOP dataset
│ └── requirements.txt
├── railway.toml # Railway deployment config
└── README.md
# Clone
git clone https://github.com/YudhaESap/clinicalmind.git
cd clinicalmind
# Backend
cd backend
pip install -r requirements.txt
cp .env.example .env
# Add ANTHROPIC_API_KEY to .env
uvicorn app.main:app --reload
# → http://localhost:8000The backend serves index.html at / — no separate frontend server needed.
Just open index.html in any browser. The offline demo includes:
- 4 pre-loaded healthcare data sources
- 20 journal-grounded clinical questions
- Instant results with charts and interpretations
- Publication-ready Table 1, Table 2, Figure 1
Questions are derived from real research in top clinical journals:
Lancet Psychiatry 2023 — Disparities in cancer care for patients with mental illness
NEJM 2023 — Rising incidence of early-onset colorectal cancer
JAMA 2023 — Mental health parity gaps in oncology insurance coverage
Acta Oncologica 2024 — 25% increased CRC mortality in severe mental illness
JCM 2024 (Roderburg et al.) — Temporal trends of psychiatric disorders in CRC
J Clin Oncol 2023 — Antidepressant non-adherence and chemotherapy discontinuation
Health Affairs 2024 — Economic burden of untreated psychiatric comorbidity in cancer
Gut 2023 — Global trends in young-adult CRC incidence
| Method | Endpoint | Description |
|---|---|---|
POST |
/api/query |
NL question → SQL → results + interpretation |
GET |
/api/schema |
OMOP CDM table schema |
POST |
/api/raw-query |
Execute raw DuckDB SQL |
GET |
/api/health |
Health check |
Yudha
GitHub: @YudhaESap
© 2026 YudhaESap. All rights reserved. Unauthorized use, reproduction, or distribution is prohibited.