Compliance-first medical AI: lessons from Nomogram for SMILE
What it takes to ship AI inside an ophthalmic clinic — OD/OS dual-eye analysis, role-based audit logs, PDF reporting and how Nomogram keeps surgeon data isolated end-to-end.
Key points
- SMILE refractive-surgery pre-op evaluation
- OD / OS bilateral data analysis
- Auto-generated PDF clinical reports
- Role-based permissions and audit logs
Compliance is the product
In a clinic, an AI feature that can’t be audited can’t be used. Nomogram was designed compliance-first: every inference is attributable, every access is logged, and surgeon data never leaves its isolation boundary.
OD / OS dual-eye analysis
SMILE refractive planning treats each eye independently. Nomogram keeps OD (right) and OS (left) data separated end-to-end, produces structured PDF reports for the clinical record, and maintains a role-based audit trail throughout.
What we learned
Shipping AI into healthcare is less about raw model accuracy and more about the guardrails around it — data isolation, permissions, reporting and reproducibility. Those constraints shaped the whole system.