Articles
Articles
Longer pieces: a guidance read closely, an enforcement pattern worth naming, or a question I think quality units should be able to answer before somebody asks them in a conference room.
- ai
- aseptic
- capa
- compounding
- cybersecurity
- drug-development
- enforcement
- eu-gmp
- fda
- gmp
- guidance
- inspections
- investigations
- medical-devices
- qmsr
- quality-systems
- readiness
- scale-up
-
What stands between you and a sterility failure
A working list of the controls that keep sterile product sterile, sorted by class, with how each one fails and what FDA has said about it in warning letters issued during 2026.
-
The investigation they should have run
A reconstruction built on one public warning letter: recurring sterility failures blamed on operators, an airflow problem found by the regulator instead, and a step by step account of who you would actually need to talk to, in what order, to reach the real cause.
-
Human error is not a root cause
Inadequate investigation is one of the most consistently cited findings in FDA inspections, and the problem is rarely that firms do not investigate. A walk through the mechanism stage by stage, the decisions each stage forces, and the defects that get records sent back.
-
Do not manage the clock. Manage the risk.
FDA’s one-day inspectional assessments are not a shorter inspection. They are a faster decision about whether your site deserves a longer one.
-
Harmonisation is a business strategy, not an afterthought
Growing companies treat FDA, EU and local requirements as three different languages. They are closer than they look, and building to the highest standard first is cheaper than retrofitting.
-
The GLP-1 story is not about telehealth. It is about distribution.
Four layers, one patient experience, and an architecture FDA has seen before. Once it looks like a system distributing drugs, the FDCA stops caring how the contracts are structured.
-
Clarity enables compliance. Ambiguity drives overcompliance.
FDA’s cybersecurity guidance for medical devices is strong on principle and fuzzy on threshold. That gap does not reduce burden. It moves the interpretation downstream, and companies overbuild to defend themselves.
-
AI is coming to the GMP quality system. Understanding has to come first.
FDA has a planned guidance on AI and machine learning in pharmaceutical manufacturing. Having evaluated process analytical technology from the inside, I think the hard part is not the model. It is the signature underneath it.
-
The regulatory door just opened for new approach methodologies
FDA’s draft guidance on non-animal methods runs on two principles: context of use and weight of evidence. Together they tell you exactly where a predictive model is allowed to sit.
-
A thinking tool, not a decision tool
Virtual cell models trained on causal perturbation data are a real scientific step up. Looked at from a regulatory perspective, they still sit upstream of anything FDA would accept as evidence.
-
If the signal is X, where does the risk live?
A generic, teaching-only framework for thinking about signal-based QMSR inspections: where an investigator is likely to enter the quality system, and the lifecycle path they may follow from there.
-
QMSR aligns the rules. The bigger change is how inspections will feel.
Everyone is discussing the ISO 13485 alignment. The more consequential shift is from QSIT’s four subsystems to following a product risk across the device lifecycle.
Shorter notes live in the blog. Everything is also available by RSS.