IQ/OQ/PQ validation protocol

When a process needs validation

Some processes make a result that can be checked completely afterward: a hole diameter can be measured on every part. Others cannot. A seal, a weld, a sterilization cycle or a molded part's internal stress can only be tested destructively, or the test would miss what matters. For those, the process itself is proven capable before it is used, so that the output can be trusted without testing it all. That proof is process validation. The first question in the protocol is therefore whether the output can be fully verified, and if it can, whether verification alone is enough.

The three stages

• Installation qualification (IQ) gives documented evidence that the equipment is installed as specified: utilities, environment, calibration of the instruments, software version, drawings, manuals, spare parts and maintenance in place.
• Operational qualification (OQ) shows that the process works across its whole operating window. It is run at the limits of the parameters, the worst case, and it challenges the alarms and interlocks. OQ is where the validated window comes from.
• Performance qualification (PQ) shows that the process consistently produces conforming product under normal production conditions: real operators, real materials, normal settings, over several runs or lots so that the normal sources of variation, such as shifts, material lots and start-ups, are included.

Some industries use "process qualification" or "process performance qualification" for the last stage; the idea is the same.

Writing the protocol

The protocol is approved before anything is run. It says what will be tested, how, how many, and what result counts as a pass. An acceptance criterion written after the result is known is not a criterion. Give the statistical rationale for sample sizes. For example, zero failures in 30 samples gives 90% confidence that the failure rate is below about 7.4%; zero in 90 gives 95% confidence that it is below about 3.3%.

Three consecutive PQ runs is a common convention, not a rule. The number should follow from the risk of the process and from what the statistics need.

Deviations

When a test fails, or the protocol cannot be followed as written, the failure is recorded as a deviation: what happened, its effect on the rest of the validation, the cause, and the resolution. A failed test stays failed in the record. The retest is a new test case, run after the cause is fixed. Rerunning until a test passes, and keeping only the pass, is one of the classic findings in a validation audit.

The report and after

The summary report states whether the evidence shows the process is validated, the operating window it is validated for, and any open items. It is approved after the last test and deviation are closed. Validation then has to be maintained: monitor the process, and revalidate when a change or a trend takes it outside what was proven. List those triggers in the report.

Related tools

A first article inspection verifies one part against the drawing; validation proves the process. The control plan holds the monitoring that keeps a validated process in its window, and the hazard analysis shows which processes matter most.

On the exam

CQE III.D asks you to interpret verification and validation results, including IQ, OQ and PQ. CMDA IV.H covers process validation by IQ/OQ/PQ along with cleanliness, test method and rework validation, and III.D.6 the assessment of process validation within production and process controls. CQPA IV.C covers validation and qualification methods for approving new products and processes, CSQP III.C.3 internal process validation in a qualification plan, and CCT III.F validation of calibration processes and software. Expect to say which stage a test belongs to, why OQ uses worst-case settings, and what to do with a failed test.

Other tools and templates