Every pharmaceutical manufacturer can recite the acronym — Installation Qualification, Operational Qualification, Performance Qualification. Most can hand you a binder of completed protocols. Fewer can clearly explain what each one actually proves, and even fewer can tell you what's at risk when one of them is done poorly or skipped to meet a project deadline.
Each phase answers a different question. Get any of those questions wrong and the equipment may still run, but the validation package won't hold up under an FDA inspection or a serious internal audit.
IQ — Installation Qualification
IQ proves that the equipment was installed correctly, against design specifications, with everything it's supposed to have and nothing it isn't.
The question IQ answers is: did we install the right thing in the right way? That means verifying utilities (power, compressed air, water for injection, drains), confirming serial numbers and model numbers match the purchase order, checking that all required components are present and intact, recording calibration certificates for every instrument, and capturing the as-installed configuration. Software components get verified too — HMI version, PLC firmware, network configuration.
IQ is mostly a paperwork-and-photographs exercise, but it's the foundation everything else rests on. If the IQ is wrong, the OQ and PQ are testing something that doesn't match what's actually installed.
OQ — Operational Qualification
OQ proves that the equipment operates within its specified ranges, across the full operating envelope, repeatably and predictably.
The question OQ answers is: does this thing actually do what the manual says it does? That means running the equipment through worst-case conditions — minimum and maximum setpoints for temperature, pressure, airflow, speed, whatever the critical process parameters are — and verifying it responds correctly and stays within tolerance. Alarms get tested. Interlocks get tested. Safety functions get tested. Recipe management, if applicable, gets exercised.
OQ is where the most calibration work happens, because every parameter you're verifying needs to be measured with a NIST-traceable reference. An OQ run with uncalibrated reference instruments isn't an OQ — it's a hope.
PQ — Performance Qualification
PQ proves that the equipment can reliably produce in-spec product under actual or simulated production conditions, run after run, by trained operators using the approved SOP.
The question PQ answers is: does this thing make good product the way we'll actually run it? That means using real or surrogate process material, real SOPs, real operators, real cleaning cycles between runs. Typically three consecutive successful production-scale runs are required to claim qualification, with full sampling and analytical verification of the resulting product.
PQ is the most expensive phase because it consumes real material and real production time. It's also the easiest to rush. The temptation to "we'll get to it after launch" has sunk more validation packages than any other single decision.
What it looks like on a fluid bed dryer
| Phase | On a Glatt or ACG fluid bed |
|---|---|
| IQ | Verify model, serial, all instruments present, calibration certs on file, utilities correct, software version logged |
| OQ | Operate at min and max inlet temperature, verify airflow ranges, test high/low alarms, exercise interlocks, verify exhaust filter pressure differential thresholds |
| PQ | Three consecutive production-scale runs of an actual product with full LOD, particle size, and yield data captured |
What happens when one is skipped or done poorly
The phases protect each other. Cutting corners on one shifts risk — and inspection scrutiny — onto the others.
"We'll do IQ and OQ now, and PQ at launch with the first production batch." This is the most common shortcut, and it's the one inspectors flag most often. PQ data from a single batch isn't qualification — it's a release decision being made by a customer.
Other failure modes we see regularly:
- OQ run using a reference instrument that wasn't itself calibrated — the OQ certifies nothing
- IQ photographs of the equipment in a configuration that doesn't match the final installed state
- PQ run by engineers, not production operators — doesn't prove the SOP works in the hands of the people who'll actually use it
- Three "consecutive" PQ runs that are actually three attempts where the failing two were quietly removed from the package
Any one of those gets challenged, and the whole validation file loses credibility. The cost of doing each phase properly the first time is always less than the cost of revalidating after an inspection observation.
Key Takeaways
- IQ proves installed correctly, OQ proves operates within spec, PQ proves makes good product reliably
- Every OQ measurement needs a NIST-traceable reference — uncalibrated instruments certify nothing
- PQ has to be performed under actual production conditions, by actual operators, using the actual SOP
- Skipping or weakening any phase shifts risk onto the others and almost always shows up later as an inspection observation