Read enough user requirement specifications and validation protocols and a pattern appears: the same measured quantity is called a critical quality attribute on one page, a critical process parameter three pages later, and a critical material attribute in the risk assessment appended at the back. The three abbreviations look interchangeable. They are not, and the difference decides what a process analyzer is actually being installed to do.
The distinction matters most at the point where a measurement plan is written. An attribute describes something about a material - the stuff in the vessel, the powder in the feeder, the resin leaving the reactor. A parameter describes something about how the process is being run - a temperature setpoint, a feed rate, an agitator speed. Analyzers usually measure the first category. The distributed control system usually measures the second. When a specification blurs them, the result is a probe justified against a requirement it cannot satisfy.
This piece sets out the working definitions, the documents each one comes from, and the places where the boundaries are genuinely ambiguous rather than merely sloppy. It assumes familiarity with the PAT framework itself and with Quality by Design, the broader regulatory philosophy the vocabulary belongs to.
Where the terms come from
The FDA’s 2004 PAT guidance established the ambition: build quality into the process by understanding and controlling it, rather than testing it into the product afterwards. It did not, however, supply the vocabulary that industry now uses daily.
That vocabulary arrived with ICH Q8(R2) Pharmaceutical Development, whose glossary defines both critical quality attribute and critical process parameter. ICH Q9 supplies the risk-management machinery used to decide which attributes and parameters earn the “critical” label. ICH Q10 places the resulting controls inside a pharmaceutical quality system, ICH Q11 extends the logic to drug substance manufacture, and ICH Q13 carries it into continuous manufacturing, where the distinction between a material attribute and a process parameter becomes operationally sharper.
Critical quality attribute (CQA)
ICH Q8(R2) defines a CQA as a physical, chemical, biological or microbiological property or characteristic that should be within an appropriate limit, range or distribution to ensure the desired product quality.
Three things follow from that wording. A CQA is a property of a material, not of the equipment or the recipe. It is defined by a limit, range or distribution rather than a single number. And its criticality is anchored to product quality - which, in the ICH framing, ultimately traces back to patient or user harm.
CQAs are usually written against the finished product, but the concept applies at intermediate stages too. The moisture content of a granulate leaving a dryer, the residual monomer in a resin before it is drummed, the particle-size distribution of an intermediate: each is an attribute of a material, and each may be critical to the quality of what eventually ships.
Critical process parameter (CPP)
ICH Q8(R2) defines a CPP as a process parameter whose variability has an impact on a critical quality attribute, and which therefore should be monitored or controlled to ensure the process produces the desired quality.
The definition is explicitly relational. Nothing is a CPP in isolation. A parameter earns the designation because a link to a specific CQA has been demonstrated, and the demonstration is what the development record has to contain. A jacket temperature is a process parameter always; it becomes critical only where its variability moves an attribute that matters.
This is also where the ordinary control-engineering meaning of “parameter” and the regulatory meaning diverge. A plant may hold hundreds of parameters under closed-loop control for reasons of throughput, energy or equipment protection. Those are operating parameters. Only the subset with a demonstrated quality linkage belongs in the CPP list, and inflating that list has a cost: every CPP drags documentation, monitoring and deviation handling behind it.
Critical material attribute (CMA)
Here the ground shifts. ICH Q8(R2) does not define “critical material attribute” as a glossary term. It speaks instead of material attributes - of the drug substance, of excipients, of in-process materials - and asks whether their variability affects product quality. The design-space definition in the same document treats material attributes and process parameters as the two families of input variables whose combination has to be shown to deliver quality.
The abbreviation CMA grew out of that treatment and is now common in industry documents, ISPE-adjacent literature and vendor specifications. It is useful shorthand: an input material property that has been shown to affect a CQA. It is not, however, a harmonised ICH term, and inspectors are entitled to ask what a company’s own quality system means by it. Defining it once, in the site glossary, and using it consistently is worth more than assuming it carries the same weight as CQA or CPP.
The practical value of the term is that it separates two kinds of variability that behave very differently. Process parameters are things the operator sets. Material attributes are things the operator inherits - from a supplier, from an upstream unit, from a recovered stream. Dealing with inherited variability is an argument for measurement rather than for tighter setpoints.
How the three connect
The causal chain the framework assumes runs in one direction: material attributes and process parameters are inputs, and quality attributes are outputs. A CMA plus a CPP produces a CQA. Design space, in ICH Q8(R2), is the demonstrated region of that input combination within which quality is assured.
That chain is what makes criticality assessment a risk exercise rather than a measurement exercise. ICH Q9 supplies the tools, and the implementation material developed alongside Q8, Q9 and Q10 makes an important point that is frequently lost: criticality is tied primarily to severity of harm, and adding a control does not make a critical attribute uncritical. Controls reduce risk. They do not delete criticality. A CQA held reliably in range by a well-tuned loop is still a CQA.
What this means for a measurement plan
Once the three categories are separated, the question a PAT project has to answer becomes concrete: which of them is the analyzer for?
Most spectroscopic analyzers measure attributes. An inline probe on a reactor reports composition, conversion or endpoint - properties of the material in front of it. Those readings may serve as a direct CQA measurement, as a surrogate for one, or as an in-process control that keeps a downstream CQA in range. Each of those three roles carries different validation expectations, and they should be distinguished in the specification rather than left implicit.
Parameters, by contrast, are usually already instrumented. Temperature, pressure and flow arrive through the control system, and a PAT project rarely improves on them. Where analyzers add something is on the attribute side, in the gap between a recipe that is being followed correctly and a material that is nonetheless drifting.
The placement decision follows from the same logic. Whether a measurement belongs inline, online, atline or offline depends on how fast the attribute moves and how quickly a response has to follow, not on which term sounds most advanced. And the user requirement specification should name the specific CQA, CPP or CMA the measurement supports, because that naming is what makes the requirement testable at qualification.
Where teams get it wrong
Four confusions recur. An in-process test result is treated as a CQA when it is really an in-process control limit set well inside the specification. A parameter is called critical because it is tightly controlled, rather than because a quality link was demonstrated. A supplier specification is imported wholesale as a set of CMAs without checking which attributes actually propagate to the finished product. And the analytical procedure itself is left out of the risk assessment, despite ICH Q14 extending the same parameter-and-attribute reasoning to the measurement method - the method has parameters that matter too, and a drifting method looks exactly like a drifting process.
The FDA’s process validation guidance closes the loop on the last of these. Continued process verification expects ongoing assurance that the process remains in a state of control, which in practice means the measurements underpinning CPP and CQA monitoring have to stay trustworthy across the product lifecycle, not only at qualification.
None of this is difficult vocabulary. It is simply vocabulary that rewards being used precisely, because the words determine what has to be demonstrated, documented and defended.