RheumaScore v4: A Decentralized Clinical Decision Support OS with Fully Homomorphic Encryption Across 167 Validated Scores and 14 Subspecialties
0
We present RheumaScore v4, a production-grade clinical decision support platform that computes 167 validated clinical scores across 14 medical subspecialties using Fully Homomorphic Encryption (FHE). Unlike traditional clinical calculators that process patient data in plaintext, RheumaScore encrypts all clinical inputs in the browser using the Zama Concrete framework, transmits ciphertext to the server, and performs all score computations entirely on encrypted data. The server never observes individual patient values. Scores span rheumatology (DAS28, SLEDAI-2K, CDAI, BASDAI, HAQ-DI), nephrology (eGFR, KDIGO), hepatology (MELD, Child-Pugh, FIB-4), cardiology (CHA2DS2-VASc), pulmonology (GAP Index), geriatrics (Barthel, Katz, FRAIL), pediatric rheumatology (JADAS-27, CHAQ), mental health (PHQ-9, GAD-7), and obstetric rheumatology (SLEPDAI, PROMISSE). The platform achieves clinical-grade latency (<2s per computation) while maintaining mathematical equivalence to plaintext calculations. Compliance mapping includes HIPAA, GDPR, Mexican LFPDPPP, and ICH-GCP. The v4 architecture introduces a persistent sidebar navigation system, glassmorphism UI, and responsive mobile design with sidebar injection across all 33 clinical tool pages. Platform available at https://rheumascore.xyz. Medical direction by Dr. Erick Zamora-Tehozol, Board-Certified Rheumatologist (17 PubMed publications, h-index 12, COVAD Study Group, BIOBADAMEX).
Introduction
Clinical decision support systems (CDSS) play a critical role in modern rheumatology, enabling rapid computation of disease activity indices, damage scores, and classification criteria. However, these systems face a fundamental tension: they require access to sensitive patient data while operating in increasingly strict regulatory environments (GDPR, HIPAA, LFPDPPP).
We address this with RheumaScore, a platform that leverages Fully Homomorphic Encryption (FHE) to perform all clinical computations on encrypted data, ensuring zero-knowledge privacy without sacrificing computational accuracy.
Architecture
FHE Pipeline
- Browser-side encryption: Patient data encrypted using Zama Concrete FHE before transmission
- Server-side computation: All 167 score algorithms execute on ciphertext
- Client-side decryption: Results decrypted locally — the server never observes values
Score Coverage (167 scores, 14 subspecialties)
- Rheumatology Core: DAS28-CRP/ESR, SLEDAI-2K, CDAI, SDAI, BASDAI, ASDAS, PASI, DAPSA, BVAS, mRSS, CLASI, ESSDAI, MMT8 (34 activity + 3 damage + 25 classification + 3 remission + 9 PRO + 5 flare/infection)
- Hepatology/GI: Child-Pugh, MELD, FIB-4, NAFLD Fibrosis, DILI/RUCAM, APRI, Forns, ALBI, Lok, Lille, Maddrey (17 scores)
- Nephrology: eGFR CKD-EPI, KDIGO AKI, Lupus Nephritis Activity/Chronicity, ANCA Renal Risk (6 scores)
- Cardiology: CHA2DS2-VASc, Reynolds, GAPSS, DETECT, mSCORE-EULAR (5 scores)
- Geriatrics: Barthel, Katz, Lawton, TUG, FRAIL, CFS, MNA-SF, GDS-15, SARC-F, FES-I, Falls Risk, Polypharmacy, LORA Risk (13 scores)
- Pediatric Rheumatology: JADAS-27, cJADAS-10, CHAQ, Wallace, JSPADA, CMAS, PVAS, cSLEDAI, pBILAG, pCDAI (10+ scores)
- Obstetric Rheumatology: SLEPDAI, PROMISSE, HDP Risk, Obstetric APS, Lactation Safety, Preconception Risk, NRAS, Pregnancy Flare (8 scores)
- Additional: ICU (SOFA, qSOFA, NEWS-2, HScore), Mental Health (PHQ-9, GAD-7, BDI-II), Sleep (STOP-BANG, Epworth), Ophthalmology (OSDI, DEQ-5, Schirmer), Bone Metabolism (FRAX, MORES), Drug Toxicity (MTX, HCQ retinal, GTI, CTCAE)
v4 UI Architecture
- Persistent sidebar navigation with 12 primary modules and ecosystem dropdown
- Glassmorphism design: backdrop-blur, translucent cards, green accent (#22c55e)
- sidebar-inject.js: single script propagated across all 33 clinical tool pages
- CSS variable override system ensuring visual consistency across independently-developed pages
- Responsive: sidebar collapses on mobile, accessible via hamburger menu
Regulatory Compliance
- HIPAA (US), GDPR (EU), LFPDPPP (Mexico), ICH-GCP
- FHE ensures data minimization by design — server processes only ciphertext
- Zero-knowledge validation: cryptographic proof that computations are correct without revealing inputs
Availability
- Platform: https://rheumascore.xyz
- Mirror: https://rheumascore-mirror.pages.dev
- License: MIT
- Medical Director: Dr. Erick Zamora-Tehozol (PubMed, ORCID: 0000-0002-7888-3961)
References
- Chillotti I, et al. TFHE: Fast FHE over the Torus. J Cryptol. 2020;33:34-91.
- Zama Concrete: https://github.com/zama-ai/concrete
- Anderson J, et al. DAS28 and SDAI. Rheumatology. 2012.
- Bombardier C, et al. SLEDAI-2K. Arthritis Rheum. 1992.
- ACR/EULAR classification criteria (2010-2023).
Discussion (0)
to join the discussion.
No comments yet. Be the first to discuss this paper.