TENSOR Med is an AI-supported clinical decision-support platform for rheumatology that generates living, fully cited treatment reports for every patient. The system synthesizes guidelines, trial evidence, and the patient's complete clinical context—including comorbidities, pregnancy status, and lab values—into recommendations that refresh automatically at each visit. Every recommendation includes traceable citations, confidence grades, and a recorded reasoning chain that clinicians can accept, modify, or decline with documented rationale.
Funding
Funding not disclosed
Founders
Product
Problem
Rheumatology clinicians must reconcile complex, frequently updated treatment guidelines with a growing body of trial evidence and each patient's unique clinical context—including comorbidities, reproductive status, and prior treatment responses. This cognitive burden increases the risk of guideline drift, overlooked drug interactions, and recommendations that do not fully account for individual patient constraints.
Solution
TENSOR Med provides an AI-supported clinical decision-support platform that generates a living, fully cited treatment report for every patient across axSpA, RA, PsA, gout, SLE, and GCA/PMR. The system employs four parallel AI analysts—examining treatment response, guideline position, trial evidence, and drug safety—that propose next steps, respond to each other, and are reconciled by a moderator into a single recommendation. Every recommendation carries traceable citations to specific guideline passages or trial data, a confidence grade, and a visible reasoning chain. The report refreshes automatically about thirty seconds after new data arrives, and a separate verification pass corrects any cited values back to the patient record before display. Clinicians can accept, modify, or decline recommendations, with each decision recorded as a constraint that shapes future reports for that patient.
Target Audience
Primary users are rheumatologists, rheumatology fellows, and advanced practice clinicians managing patients with inflammatory arthritis, gout, SLE, or vasculitis in outpatient or hospital settings.
Features
- Four-analyst AI reasoning system (treatment response, guideline position, trial evidence, drug safety) with moderator reconciliation and explicit disagreement reporting
- Living treatment report that auto-regenerates ~30 seconds after new data arrives, reflecting the entire current record
- Encoded citation library covering ASAS/EULAR (axSpA), ACR/EULAR (RA), GRAPPA (PsA), EULAR (SLE), and ACR (gout) guidelines plus pivotal trials and label-aligned drug monitoring profiles
- Clickable citations that open the exact source passage or jump to the specific patient-record value cited
- Post-generation verification pass that cross-checks every cited score against the record and flags unreconcilable values rather than silently changing them
- Document ingestion supporting PDFs, photos (OCR-first redaction), structured hospital exports, and spreadsheets, with structured preview before anything commits to the record
- Multi-language parsing for English, Turkish (including non-Unicode older hospital PDFs), and Spanish, with AI reasoning in all three languages
- Pregnancy and breastfeeding awareness, including teratogen washout calculations and conception-planning support
- Multi-condition reasoning for overlapping diagnoses (e.g., axSpA with IBD, RA/PsA overlap) with side-by-side disease-specific scores
- Duplicate-patient detection before record creation, plus optional care-team access restriction with logged break-glass emergency access