A change control has to clear review before it runs

This blog has already covered what a validation master plan decides that no single protocol can, including what triggers a revalidation once a process, material, or piece of equipment changes. It stopped short of the document that actually starts that chain: the change control record itself. On the device side, this blog has also covered why a corrective action isn't finished the moment its fix ships — a reactive record, opened after something has already gone wrong. A change control runs the opposite direction: it has to clear review before anything changes at all.

211.100(a) treats the procedure's own text as something to approve

21 CFR 211.100(a) requires written procedures for production and process control, designed to assure that a drug product has the identity, strength, quality, and purity it's represented to have. The same subsection extends that approval duty to the procedure's own later amendments: a change to a written production or process control procedure has to be drafted by the appropriate organizational unit and reviewed and approved by the quality control unit, the same as the original procedure was, before the change takes effect. The regulation never uses the phrase ‘change control’ — that's industry's own name for the mechanism 211.100(a) requires without naming it.

A change control is the planned half; 211.100(b) governs the other

211.100(b) requires written procedures to actually be followed, and any departure from them to be recorded and justified — the deviation side of the same rule, covering what happens when a batch doesn't match the procedure on a given day. A change control runs in the opposite direction: it's the mechanism for updating the procedure itself, on purpose, before the next batch runs under it. Treating a deliberate change as a deviation to explain after the fact, instead of a change to evaluate and approve before it happens, skips the review 211.100(a) actually requires and leaves the record describing an event as unplanned that was never unplanned at all.

ICH Q10 scales the evaluation to risk, not to habit

ICH Q10's pharmaceutical quality system builds a change management system around that same approval duty, but widens it past any single change at a time: proposed changes are evaluated using quality risk management, with the depth of that evaluation scaled to the risk and the knowledge behind the specific change rather than applied the same way to every request. The same system is meant to track a change's effect across a product's full lifecycle and against a firm's own regulatory filings, not just the production floor — a change that looks purely procedural on the line can still require updating a marketing authorization the production record alone would never flag.

The device side used to say this almost word for word

Until the QMSR replaced it in February 2026, FDA's device-side quality system regulation ran the same rule as 211.100(a) under a different number: 21 CFR 820.70(b) required a manufacturer to verify, or validate where appropriate, a change to a specification, method, process, or procedure before putting it into effect, and to approve that change before implementation rather than after. The numbered section is gone, but this blog has already covered what the QMSR actually changed when it folded the device rule into ISO 13485, and the same before-not-after sequencing survives inside the standard the device side now runs on. A change approved only after the batch has already run has broken the same rule on both sides of the aisle, just under two different document numbers.

Where this meets the file

A change control record built around that sequence — the proposed change, its risk evaluation, the quality unit's approval before implementation, and the regulatory-filing or revalidation check the evaluation triggers — is previewed in the launch catalog alongside the validation master plan it often has to update. If your site structures this approval differently, the shelf takes that correction directly.