Continued Process Verification
Continued process verification (CPV) is the third stage of the modern process-validation lifecycle: the ongoing collection and analysis of process and product data during routine commercial production to provide continual assurance that the process remains in a state of control. It replaces the outdated notion that validation ends after three successful batches, embedding statistical monitoring, trending and periodic review into the product lifecycle so that drift and special-cause variation are detected and acted on before they become failures.
What an explainer is not
A topic explainer is SPEQ’s synthesis of what a practice involves, cited to the standards that govern it. It does not reproduce their text, and it does not determine which of them apply to your product or process.
[ POSITION IN THE FRAMEWORK ]
7 DIMENSIONS · 23 LINKSContinued process verification is validation's Stage 3 across the GMP and quality-system disciplines: statistical monitoring that catches drift before an OOS and routes signals into CAPA, evidenced in the manufacturing systems below.
06 · QUALITY MATURITY — CONTINUED PROCESS VERIFICATION, REACTIVE TO ADAPTIVE
Validation is 'three batches and done'; routine production is judged only by pass/fail specification, so drift is invisible until an OOS.
A CPV plan and some trending exist, but limits come from the specification, not the process's own history, and charts are rarely reviewed.
CQAs/CPPs are monitored with statistically justified limits; signals route into deviation and CAPA, and data feeds the product quality review.
Control charts distinguish special- from common-cause variation; slow drifts that never breach specification are detected and investigated.
CPV drives continual improvement — reducing variability, refining the control strategy, and informing requalification across the product lifecycle.
SPEQ’s shared five-stage progression, labelled synthesis — not the FDA QMM rating scale. Where does your organization sit? Score your quality system →
07 · REGULATORY & EVIDENCE
GOVERNING STANDARDS · 3
Derived from the 3 standards SPEQ maps to this subject, across 3 regulatory bodies: FDA, EMA, ICH.
RECORDS & OBJECTIVE EVIDENCE
- A CPV plan identifying CQAs/CPPs, data sources, and review responsibility
- Statistically justified alert/action limits derived from process history
- Control charts and process-capability trending across batches
- Investigations and CAPA arising from CPV signals
- Product quality reviews incorporating CPV trend data
COMMON INSPECTION FINDINGS
- 'Three batches and done' with no ongoing verification after qualification
- Limits set from the specification rather than the process's own performance
- Control charts generated but never reviewed or acted upon
- Adverse trends within specification not investigated
- Confirmed process shifts not routed into change control or requalification
The third stage of the lifecycle
The FDA's process-validation guidance defines validation as a three-stage lifecycle: Stage 1 process design, Stage 2 process qualification, and Stage 3 continued process verification. The paradigm shift is that validation is not a one-time event that concludes with qualification batches; it is a lifecycle in which the process is continually verified during commercial production. CPV is the stage that makes "validated" a maintained state rather than a historical claim.
This aligns with the ICH Q10 pharmaceutical quality system and its emphasis on continual improvement and process performance monitoring, and it is reinforced in EU GMP through the ongoing process verification expectation of Annex 15. The three converge on the same idea: a validated process must be monitored across its life, with the monitoring itself designed on the process understanding developed in Stage 1.
What CPV actually requires
A CPV program identifies the critical quality attributes and critical process parameters worth monitoring, defines how their data will be collected and analyzed, sets statistically justified alert and action limits, and assigns responsibility for review and response. The monitoring should be broad enough to detect variability from all relevant sources — raw materials, equipment, environment, operators — and the plan should specify sampling and testing frequency proportionate to risk and to the maturity of process understanding.
The distinguishing feature is the use of statistical methods rather than pass/fail specification checks alone. A batch can be within specification while a parameter trends steadily toward the edge; CPV's job is to catch that trend. SPEQ synthesis: the discriminating question for a CPV program is whether it would detect a slow drift that never breaches specification — a program that only reacts to out-of-specification results is doing release testing, not continued process verification.
Statistical process control and trending
CPV relies on statistical process control: control charts that distinguish common-cause (inherent, stable) variation from special-cause (assignable, actionable) variation, and process-capability indices that quantify how well the process fits within its specification limits over time. Control limits are derived from the process's own historical performance, not from the specification, so a signal is raised when the process behaves unusually for itself — before it produces an out-of-specification result.
Trending these indicators over time and across batches reveals gradual shifts, cycling, or increasing variability that single-batch review would miss. When a signal appears, it triggers investigation under the quality system, and the outcome may be a corrective action, a process improvement, or a revalidation. The data also feeds the product quality review, closing the loop between routine monitoring and periodic holistic assessment.
Acting on the data and closing the loop
CPV is only worth the data collection if it drives action. Signals must route into deviation and CAPA processes, adverse trends must be investigated for root cause, and confirmed process shifts must lead to change control and, where warranted, requalification. The program should also feed continual improvement — CPV data often reveals opportunities to tighten control or reduce variability that were not visible from qualification batches alone.
The connection to the broader quality system is what makes CPV credible to an inspector: the monitoring plan, the statistical analysis, the trending, the investigations, and the resulting changes should form a visible, auditable chain. A CPV program that collects charts nobody reviews, or raises signals nobody investigates, is a documentation exercise, not process verification. The intent is continual assurance that the process remains in control — and demonstrable evidence that the organization acts when it is not.
FREQUENTLY ASKED
Is process validation finished after three successful batches?
No. The modern lifecycle model treats validation as ongoing: Stage 3 continued process verification collects and analyzes process and product data throughout commercial production to provide continual assurance the process stays in control. The "three batches and done" idea is outdated — qualification batches are Stage 2, not the end of validation.
How is CPV different from routine release testing?
Release testing checks whether each batch meets specification (pass/fail). CPV uses statistical methods and trending to detect variability and drift across batches — including shifts that stay within specification but signal the process is moving out of its normal state. CPV is designed to catch problems before they become out-of-specification results.
What statistical tools does CPV use?
Primarily statistical process control — control charts that separate common-cause from special-cause variation using limits derived from the process's own history, plus process-capability indices tracked over time. Signals trigger investigation under the quality system, and the data feeds the product quality review and continual improvement.
How does CPV connect to the quality system?
CPV signals route into deviation, investigation and CAPA processes; confirmed shifts drive change control and potential requalification; and the trended data feeds the periodic product quality review. This visible, auditable chain from monitoring to action is what distinguishes real continued process verification from collecting charts nobody reviews.