Change control in pharmaceutical manufacturing

Change control is the formal system for evaluating, approving, implementing and reviewing changes that may affect product quality, patient safety, validated state, regulatory commitments or the pharmaceutical quality system.

Its purpose is not to prevent change. It is to make beneficial change without losing control. A strong record explains what will change, why, what else may be affected, which evidence is required before use and how the organisation will know the change worked.

What belongs in change control?

• Facilities, rooms, utilities and environmental controls
• Manufacturing, packaging and laboratory equipment
• Processes, parameters, batch sizes, hold times and cleaning methods
• Raw materials, components, suppliers and specifications
• Analytical methods, sampling plans and reference standards
• Computerised systems, configurations, interfaces and reports
• Documents, records, labels and artwork
• Organisational responsibilities or outsourced activities affecting quality
• Regulatory dossiers, commitments and product licences

Like-for-like maintenance may follow an approved maintenance process, while a replacement with different design, material, software, capacity or operating principle may require change control. Define the boundary in procedure and escalate uncertainty.

1. Write a decision-ready proposal

Avoid descriptions such as “upgrade machine” or “revise SOP.” State the current state, proposed state, reason, expected benefit, affected location and intended implementation date. Identify temporary arrangements and whether the change is permanent, emergency or time-limited.

2. Classify without using classification as a shortcut

Minor, major or critical labels can set review depth and approval authority, but the label must follow assessment—not replace it. Consider potential effect on critical quality attributes, process parameters, contamination control, data integrity, validation, stability, supply continuity and regulatory commitments.

Emergency change controls should permit urgent action while preserving authorisation, risk controls, traceability, retrospective review and a route either to permanent approval or reversal.

3. Map impacted systems cross-functionally

The change owner coordinates; subject-matter experts assess their domains; Quality provides independent challenge and approval. “No impact” should carry a rationale.

4. Perform risk assessment that informs the plan

Risk tools should help decide controls and evidence. Start with hazards and failure modes, not a desired low score. Consider severity, probability and detectability using available product and process knowledge. Document assumptions and uncertainty.

Risk assessment should determine:

• Required studies, trials, qualification and validation
• Whether concurrent or prospective evidence is justified
• Additional sampling, monitoring or release controls
• Training and qualification requirements
• Regulatory pathway and implementation restrictions
• Rollback or contingency needs
• Post-implementation review criteria and duration

5. Build an integrated action plan

Actions should have owners, due dates, dependencies and acceptance criteria. Sequence matters: approve specifications before purchase; complete installation qualification before operational testing; approve instructions and qualify users before routine operation; obtain regulatory clearance before implementation where required.

Define the implementation gate—the exact evidence that must be approved before the new state may be used. Examples include approved protocol and report, calibrated instruments, released materials, updated master documents, completed training and authorised system access.

6. Scale qualification and validation to risk

Not every change requires full revalidation, but every decision requires scientific justification. Review the original qualification and validation basis, changed functions, worst-case conditions, interfaces and accumulated changes.

Practical examples

Example 1: replacing a filling-machine sensor

A discontinued sensor is replaced with a newer model. The connection fits, but response time and output range differ. Impact assessment identifies PLC scaling, alarm set-points, calibration, spare-part records and maintenance training. The change is implemented only after documented configuration and challenge testing.

Example 2: new API supplier

The material meets the same written specification, but supplier process, impurity profile, particle-size distribution and transport route may differ. Supplier qualification, regulatory status, method suitability, process impact, stability and initial-batch controls are evaluated before approval.

Example 3: laboratory software upgrade

The vendor calls the release “minor.” The site still evaluates operating-system compatibility, user roles, audit trails, calculations, interfaces, report templates, data migration, backup and restore. Vendor testing supports but does not replace site assurance for intended use.

Example 4: temporary increase in hold time

Production pressure alone does not justify implementation. Existing development and validation data, microbiological or chemical risk, container conditions, sampling and regulatory commitments are assessed. Temporary status, expiry, batch restrictions and reversion are controlled.

7. Control implementation

• Confirm all prerequisite actions and approvals are complete.
• Identify the first affected batch, record, system version or effective date.
• Prevent simultaneous use of obsolete and current instructions.
• Communicate the change across shifts and affected partners.
• Apply enhanced monitoring or additional release controls where planned.
• Document deviations from the approved plan through the quality system.
• Keep a tested rollback or contingency route where failure is plausible.

8. Perform a post-implementation review

Closure should confirm more than task completion. Compare actual results with predefined objectives and risk controls. Review initial batches or runs, deviations, alarms, yield, quality attributes, complaints, stability, audit trails or user feedback as applicable.

For designing outcome-based review criteria, see the effectiveness-check guide.

Common mistakes

• Implementing before approval because the purchase order is urgent
• Treating “like for like” as self-evident
• Assessing only the initiating department
• Using a risk score without describing hazards and controls
• Leaving regulatory assessment until after implementation
• Closing when actions are complete but results are unknown
• Extending temporary changes repeatedly
• Failing to consider cumulative effect of multiple small changes
• Omitting data migration, access and audit-trail effects
• Not identifying the first batch or system version affected

Educational notice: Apply the requirements of your approved quality system, marketing authorisations and relevant regulatory authorities.