Build safely
The joy of building is unconditional; the obstacles are conditional. The lane is set by what your software touches, not by what it is. The long version, including the EU in-house exemption, is in The 3am Nurse Test.
Unlike the rest of the guides, this page is a live feed, not a one-time check: refreshed automatically alongside Pulse (last update 2026-10-28).
Green: build freely. No real patient data, no live clinical decisions. Prototyping even a clinical app on synthetic or sandboxed data is green, and so are production tools that never touch patient data (admin, rotas, comms, education).
Amber: compliance basics apply. Real patient data or a live clinical workflow: data protection, access controls and audit trails become part of the build.
Red: device territory. Influencing real diagnosis or treatment in production. In the EU, the in-house exemption can be the escape hatch for tools built and used within one institution.
- Mario Amaro, MD warns on PHI and security risk in the wave of doctor-made apps
The watchlist physician-builder is publicly flagging to fellow doctors the PHI and security dangers of shipping vibe-coded clinical apps on consumer platforms, as more physicians release their own tools .
2026-10-25 · x - Mario Amaro, MD, warns vibe coded tools are pitched as secure while holding patient data on personal devices
The physician-builder (Vibe Code Doc) flags an unresolved compliance gap: many clinician-built tools keep patient data on a clinician's own device while claiming to be secure .
2026-10-25 · x - Dr Keith Grimes: any NHS AI-incident duty belongs inside the clinical safety case
As the UK signals hospitals may have to report serious AI incidents and near misses, Grimes argues the duty should sit in DCB0129/DCB0160 hazard logs rather than a parallel reporting route .
2026-10-14 · linkedin - Emergency physician John Lee: FDA clearance of Epic's Deterioration Index still demands local validation
The clinician-authored analysis points to racial sensitivity gaps in the clearance filing (26 per cent for Black versus 51 per cent for white patients) and to clinicians pasting chart content into frontier models as an…
2026-10-14 · substack - Can vibe coding do the statistics? Five LLMs benchmarked on expert-validated analyses (SHTI)
Jeeva Sam and colleagues, Studies in Health Technology and Informatics: five current LLMs benchmarked on reproducing expert-validated statistical analyses (descriptive tables, Kaplan-Meier, Cox modelling); outputs…
2026-10-09 · backfill:handover