Instrument Qualification: DQ, IQ, OQ, PQ in Plain Terms
Share
An instrument qualification file is usually presented as a certificate. The certificate says the instrument is qualified. It rarely says which question was asked, and the four stages exist precisely because there are four different questions.
The distinction is not academic. An HPLC can be installed correctly, function correctly, pass calibration, and still produce wrong numbers on a bile-acid assay, because the fifth question, whether the whole method works on this sample type, was never asked. That question has a name and a stage. It is the one buyers should ask about.
Four stages, four promises
| Stage | Question | Evidence it leaves behind |
|---|---|---|
| Design qualification | Is this the right instrument for the work we intend? | User requirement specification and the vendor's response to it |
| Installation qualification | Was it installed as specified, in the right environment? | Installation record, utility checks, serial numbers, software version |
| Operation qualification | Do its functions meet specification? | Tests of flow, wavelength accuracy, temperature, injection precision |
| Performance qualification | Does the method work on the intended sample? | Results on a known sample, against an agreed acceptance criterion |
Design qualification is the stage most often absent in practice and most often blamed later. It is the only stage that can prevent buying the wrong instrument. If a laboratory buys a detector that cannot reach 205 nm because nobody wrote down that the assay needs 205 nm, the remaining three stages will be executed faithfully on hardware that cannot do the job, and every one of them will pass.
Installation is not a delivery note
Installation qualification records the state of the instrument at the moment it starts its life: serial numbers of critical components, firmware and software versions, utility conditions, and the environment it sits in. That record has one practical use. When something changes later, it is the only way to know what changed.
The environment element is underrated. A detector whose baseline drifts because the laboratory's air conditioning cycles will fail a system suitability check on a warm afternoon and pass it in the morning, and the investigation will look at the column for three days before anyone looks at the thermometer. Recording ambient conditions during qualification gives the later investigation a baseline.
Operation qualification is about the machine
Operation qualification tests functions against the vendor's specification: flow rate accuracy, gradient performance, wavelength accuracy, column oven stability, autosampler precision, detector noise and drift. These are generic tests. They are supposed to be generic, because they describe the instrument rather than the analysis.
The trap is treating a passing OQ as a licence to report results. It is not. OQ says the pump delivers what it says it delivers. It says nothing about whether the method separates what it needs to separate, which is a property of column, mobile phase, temperature and sample together.
Performance qualification is the one that matters
Performance qualification runs the actual method on the actual sample type and compares the outcome with an acceptance criterion agreed in advance. For an assay, that might be a known reference material within one per cent of its assigned value. For an impurity method, it might be a spiked sample at the specification limit with recovery inside a stated window.
This is the stage that answers the customer's question, and it is the stage most often compressed into "the instrument was qualified by the vendor". A vendor can qualify an instrument. A vendor cannot qualify it for a method the vendor has never seen.
The acceptance criterion has to be written before the experiment. A criterion chosen after seeing the numbers is not a criterion, it is a description, and inspectors read it that way.
Calibration sits beside qualification, not above it
Calibration is a narrow, periodic check against a reference. It is necessary and it is not sufficient. A balance can be calibrated and still be unsuitable for weighing twenty milligrams, because calibration confirms the reading at the points tested and says nothing about the repeatability at the low end. That particular trap is large enough to deserve its own treatment, and it is covered in the note on minimum sample weight.
The same logic applies to pH meters, Karl Fischer titrators and UV spectrophotometers. Each has a calibration routine, each routine covers a subset of what the method needs, and the gap between the two is where results go wrong without any instrument actually failing.
Software and the audit trail
Chromatography data systems hold the integration parameters, the calculations and the audit trail. Qualifying the instrument and leaving the software unassessed is like calibrating a balance and never checking whether the printer prints the right number.
Three things are worth confirming in any laboratory whose data you intend to rely on: that access is role-based and individual, that the audit trail cannot be switched off, and that integration changes carry a reason. These are also the three things that make an electronic record admissible as a record rather than as a screenshot, which is the subject of the data integrity note.
The interval question
Qualification is not permanent. Major repair, relocation, a software upgrade, or a change in the method it supports each create a reason to revisit. The interval between requalification is a documented decision, and the decision should name what would trigger an out-of-cycle requalification.
A file whose newest entry is four years old is not evidence about the instrument in use today. It is evidence about an instrument that existed four years ago, which may or may not be the same one after two repairs and a firmware update.
What a buyer should ask for
Four items, none of which requires the full folder. The date of the most recent performance qualification. The acceptance criterion that was used. Whether the software has been assessed for access control and audit trail. And what has changed since, in repair or version terms.
Those four questions separate a laboratory that maintains its instruments from one that maintains a binder. Both will tell you the instrument is qualified. Only one can tell you what that means.
Where this leaves a supply-chain partner
We do not describe an instrument as qualified without naming which stage the evidence comes from and when it was last performed. Where a customer's method differs from ours, the qualification of our instrument does not transfer, and the honest answer is that a method transfer is needed rather than a copy of our certificate.
That answer is slower in the first email and considerably faster in the third month.
What triggers requalification
Qualification is a statement about a configuration at a moment. When the configuration changes, the statement stops applying, and the only question is whether anybody noticed. The triggers are worth writing into a list, because the list is what turns a policy into a practice.
| Trigger | What it invalidates | Ordinary response |
|---|---|---|
| Major repair, such as a pump or detector replacement | Operation qualification, partly performance qualification | Repeat OQ on the affected function, then a PQ check |
| Relocation, including within the same building | Installation and operation qualification | Repeat IQ and OQ, then PQ |
| Firmware or software upgrade | Data system qualification, audit trail verification | Repeat the software assessment and the audit trail test |
| New method on the same instrument | Performance qualification for that method | Method-specific PQ before reporting results |
| Preventive maintenance with parts replaced | Depends on the part; lamps and seals usually need a check | System suitability evidence, documented |
The last row is where most sites improvise. A lamp change on a UV detector shifts the energy profile and can move a baseline, and the first evidence of it is usually a system suitability failure on a batch that was already injected. Documenting the check that followed the maintenance closes that gap for the cost of one injection.
Qualification during a method transfer
When a method moves between laboratories, the receiving instrument must be qualified for that method before the comparison runs. Doing it the other way round produces a comparison between a qualified laboratory and an unqualified one, and the difference between the two results cannot be attributed.
This is why a transfer protocol normally opens with a qualification check rather than with the comparative runs. The order is not bureaucratic; it protects the conclusion. A third-party laboratory's confirmation of a result is a useful cross-check and a poor substitute for this step, because the outside laboratory's qualification says nothing about the instrument in the originating site.
The documentary trail matters more here than at any other point, because a transfer is exactly when a customer's auditor will ask which instrument produced which number. The general expectations for equipment and records sit in the ICH quality guidelines, the European interpretation in EudraLex volume 4, and the pharmacopoeial method definitions at the EDQM. Three sources, one requirement: the record has to identify the instrument, and the instrument has to have been shown to work on that method.
A short question closes this section for any supplier: what was the date of the last performance qualification on the instrument that ran our last batch, and what method was it qualified against? An answer with two dates is an answer. A single date is a certificate.
Frequently asked questions
Is annual calibration the same as qualification?
No. Calibration checks one measurement against a reference at one point in time, usually with a certificate. Qualification is the documented evidence that the instrument as installed, as operated and as used on real samples performs within stated limits. An instrument can pass calibration every year and still be unqualified for the method it is being used for.
Which stage do buyers actually need to see?
Performance qualification, because it uses the method and the sample type in question. Installation and operation qualification tell you the instrument exists and functions. Performance qualification tells you it produces acceptable results on material like yours. Ask for the PQ summary and the date, not the whole folder.
Can a supplier qualify an instrument once and never revisit it?
Qualification is not permanent. Major repair, relocation, software upgrade or a change in the method it supports all create a reason to reconsider. A folder whose most recent entry is four years old is a historical document, not evidence about the instrument in use today.
What is the difference between OQ and PQ?
Operation qualification tests the instrument's functions against vendor specification, usually with reference materials or internal standards. Performance qualification tests the complete method on the intended sample type. OQ answers whether the machine works. PQ answers whether the analysis works.
Does software need qualification too?
Yes, and it is the part most often skipped. Chromatography data systems hold the audit trail, the integration parameters and the calculation. A system that has never been assessed for access control and audit trail completeness cannot support the records it produces, whatever the instrument in front of it is doing.
What happens if an instrument fails PQ?
Results produced since the last successful qualification come into question, and the scope of the question is the interval, not the individual run. That is why the interval between qualifications is a decision worth documenting rather than a gap that fills itself.