In‑Process Verification (IPV) – Proving Quality During Manufacture

This topic is part of the SG Systems Global regulatory & operations glossary.

Updated November 2025 • IPC, PAT, CPV • Quality, Manufacturing, Validation, Compliance

In‑Process Verification (IPV) is the structured set of checks, tests and confirmations performed while the batch is being made to demonstrate that critical steps have been executed correctly and that the process remains in control. Instead of relying only on finished‑product testing, IPV uses in‑process controls (IPC), sampling, PAT tools and system verifications at key points to build real‑time evidence that the batch will meet its specifications.

“If you wait until the end to find a problem, you’ve already wasted the batch. IPV is about catching issues while you can still act.”

1) Where In‑Process Verification Fits in the Lifecycle

IPV sits between process validation on paper and final batch release. During design and validation you define critical quality attributes (CQAs) and critical process parameters (CPPs). IPV is how you actively check those attributes and parameters in real time for every batch—not just in a validation campaign or retrospective data analysis.

In practice, IPV is executed through the Batch Manufacturing Record (BMR) or eBR, equipment controls, inline sensors and defined sampling plans. The data then feeds CPV trending and ongoing QbD‑style improvements, closing the loop between design, execution and performance monitoring.

2) Regulatory Anchors and Terminology

While regulators use slightly different labels—“in‑process testing”, “in‑process controls”, “on‑line/off‑line testing”, “process monitoring”—the core idea is the same: demonstrable control of the process while it runs. 21 CFR 211 and EU GMP both explicitly reference in‑process testing and controls as part of GMP; modern guidance emphasises that these can, in some cases, reduce or replace finished‑product testing if appropriately justified.

In that sense, IPV is not a new regulatory invention but a more deliberate, integrated way of designing and documenting in‑process checks as part of the overall control strategy. When inspectors ask, “How do you know the process stayed in control?”, IPV is the concrete, data‑backed answer rather than vague reassurances about “experienced operators” or “historically good performance”.

3) IPV vs In‑Process Controls (IPC) and CPV

In‑process controls (IPC) are the actual tests or measurements performed during manufacture—pH checks, blend uniformity tests, torque readings, weight checks, visual inspections. IPV is the structured verification framework around those controls: where they sit in the process, how often they occur, what limits apply, what actions are triggered and how they are documented.

CPV then uses the accumulated in‑process data (including IPV results) to demonstrate long‑term consistency and detect slow drifts or shifts. Think of it this way: IPC is the raw act of measuring, IPV is the designed pattern of checks that make a batch “verifiable”, and CPV is the statistical analysis over time that proves and maintains the state of control.

4) What IPV Actually Covers in a Batch

A robust IPV plan typically spans multiple layers. At the unit‑operation level it covers critical parameters like temperature, pressure, agitation speed, flow rates and hold times. At the material level it confirms correct component identity, potency adjustments, charge weights and yields. At the product level it may include intermediate assays, blend uniformity, in‑process moisture and tablet weight/hardness checks.

On top of that, IPV encompasses verification of line clearance, label set‑up, label/barcode verification, container closure checks and other inspection steps that are crucial for finished‑goods compliance even if they do not directly measure CQAs. All of this must be defined, justified and recorded, not improvised batch‑by‑batch.

5) Role of PAT, SPC and Online Sensing

Modern IPV is heavily augmented by Process Analytical Technology (PAT) and statistical process control (SPC). Inline NIR probes, online particle‑size measurements, continuous weight monitoring and real‑time spectroscopic assays provide continuous evidence that the process remains on target, rather than sporadic snapshots.

SPC tools (control charts, Cp/Cpk calculations, alert/action limits) then distinguish normal variation from signals that require intervention. In a well‑designed IPV strategy, PAT and SPC outputs are not just monitored on screen; they are integrated into the decision logic—pausing, adjusting or rejecting stages when patterns indicate loss of control, and documenting those actions in the eBR or control system.

6) Documentation in BMR/eBR and Control Plans

IPV must be visible in your documentation, not just in people’s heads. At minimum, IPV elements should appear in the master BMR / eBR, in the process control plan (PCP) and in validation and QbD documentation describing the product control strategy.

For automated environments, IPV may be implemented as enforced checks in the MES, interlocks in equipment PLC/SCADA, or mandatory data fields in the eBR that prevent progression until limits are met or deviations raised. For hybrid or paper processes, it still needs to be explicit: what is checked, by whom, how often, against which limits, and with what documented response to failures.

7) IPV and Finished‑Product Testing

Regulators still expect appropriate finished‑product testing, but they are clear that testing alone does not assure quality. IPV provides assurance that the process has consistently produced material meeting its CQAs, so finished‑product tests become confirmation rather than the first time anyone really looks at the batch.

In some regulated contexts, where process understanding and PAT/IPV are strong, firms may justify reduced sampling plans or, for certain attributes, replacement of end‑product testing with in‑process results. That is not a loophole; it demands a stronger demonstrated process understanding and a robust, well‑documented IPV/CPV regime, not a weaker one.

8) Triggering Deviations, OOS/OOT and Corrective Actions

IPV is only credible if failures lead to action. A parameter excursion, failed in‑process assay or out‑of‑trend pattern is not “just data”; it is a trigger for OOS/OOT workflows, deviations, root‑cause analysis and, where needed, CAPA.

That means IPV design must include explicit links from each check to the required response: when to pause the line, when to scrap material, when to re‑work, when to expand sampling, and when to open a formal investigation. In an electronic environment, this is often implemented as enforced decision branches and automatic creation of QMS records rather than manual “please remember” steps.

9) Data Integrity, Systems and Validation

Because IPV evidence is part of batch release justification, it must meet the same data‑integrity standards as lab results or final COAs. That includes secure, time‑stamped acquisition; traceability back to instruments and operators; protection against unauthorised changes; and retention in line with record‑retention requirements.

Systems used to capture IPV data—MES, PAT platforms, process historians, LIMS—fall under 21 CFR Part 11, Annex 11 and CSV expectations. Inspections increasingly probe how IPV‑critical data flows across these systems and how you ensure its integrity and availability during batch review and regulatory submissions.

10) Metrics and Use in CPV and PQR/APR

Over time, IPV generates a rich dataset: the distribution of in‑process results, frequency of near‑misses, drift toward alert limits, intervention rates and impact on yield or quality events. These are prime inputs for CPV programmes and PQR/APR.

Typical IPV‑related metrics include percentage of batches with one or more IPV failures, time from IPV failure to intervention, correlation between IPV patterns and deviations or complaints, and the proportion of finished‑product failures that were predictable from earlier IPV signals. If those metrics never feed back into risk assessments or control‑strategy updates, IPV is being under‑used.

11) Sampling Plans and Verification Frequency

IPV design is tightly linked to GMP sampling plans. For each CQA or supporting attribute you need to define how many samples, from where, how often, and with what statistical confidence. For continuous or large‑volume processes, relying on a single grab sample is rarely defensible.

In practice, firms combine statistically designed sampling for critical attributes with high‑frequency or continuous monitoring for key parameters. Verification frequency may also vary by risk tier—for example, higher sampling during process start‑up, after maintenance or changeover, tapering down once the process has demonstrated stable behaviour within defined boundaries in CPV data.

12) Role of Data Platforms and Analytics

As IPV matures, the challenge becomes less “do we measure?” and more “can we interpret the flood of in‑process data?”. Aggregating IPV data into a validated GxP data lake and analytics platform allows multivariate analysis across lines, sites, materials and product families.

Advanced analytics can highlight subtle interactions (e.g. certain raw‑material lots and equipment combinations) that drive IPV trends long before formal OOS events. As with PAT, these tools should guide human judgement and QRM, not auto‑pilot critical decisions without explainability or appropriate validation and governance.

13) People, Training and Culture

IPV only works if operators and supervisors treat it as a non‑negotiable part of the process, not as optional extra tests. That means training that explains why IPV checks exist, how they protect patients and the business, and what actions are expected when checks fail or drift toward limits.

From a quality‑culture perspective, IPV can be a litmus test: do people log and escalate marginal results, or do they quietly re‑test until they get something “acceptable”? Clear expectations, aligned incentives and visible management support are needed to make “escalate early” the safest option for individuals, not the riskiest politically.

14) Implementation Steps and Maturity Path

Most organisations already do some level of in‑process checking; the real step is formalising it into a coherent IPV strategy. A pragmatic path starts with inventorying existing checks, mapping them to CQAs/CPPs and classifying them by risk and data quality (manual vs automated, paper vs electronic).

From there, you can redesign the control plan: close gaps for unmonitored risks, remove low‑value checks that never find anything, digitise high‑risk IPV steps into MES/eBR, and connect them to deviations and CPV. Over time, adding PAT, SPC and analytics moves you from basic IPV (“we measure”) to mature IPV (“we understand and act quickly on what we measure”).

15) FAQ

Q1. How is In‑Process Verification different from standard in‑process controls?
IPV is the designed, documented framework of checks that proves process control for each batch. In‑process controls are the individual tests; IPV is the coherent strategy tying them to CQAs, CPPs and release decisions.

Q2. Can IPV replace finished‑product testing?
Sometimes for specific attributes, but only with strong process understanding, PAT/IPV data, robust CPV and a documented regulatory justification. In most cases it reduces reliance on, rather than eliminates, end‑product tests.

Q3. How does IPV relate to Continued Process Verification (CPV)?
IPV produces the in‑process data that CPV analyses over time. If IPV is weak or poorly recorded, CPV becomes a statistical exercise on incomplete or low‑value data and cannot convincingly demonstrate a state of control.

Q4. What systems do we need to support IPV?
At minimum, reliable instruments and clear documentation. For higher maturity, an integrated MES/eBR, PAT tools, process historian, LIMS and a validated analytics platform make IPV scalable and auditable across sites.

Q5. What is the first practical step to strengthen IPV?
Map existing in‑process checks against CQAs/CPPs, identify gaps and low‑value tests, and update the control plan and BMR/eBR so each critical attribute has clear IPV coverage, limits and defined actions.

Related Reading
• Process Control & Validation: IPC | Process Validation | CPV | PAT | SPC
• Records & Systems: BMR | eBR | MES | Process Historian | GxP Data Lake
• Quality & Risk: OOS | OOT | QRM | QbD | PQR/APR
• Compliance & Data Integrity: 21 CFR 211 | 21 CFR Part 11 | Annex 11 | CSV | GxP