Analytical Procedure Lifecycle: ICH Q14 and Q2(R2)

For most of the last thirty years, an analytical method was treated as a fixed artifact. You wrote a procedure, validated it against the eight performance characteristics in ICH Q2(R1), filed the validation report, and locked it down. Any change after that was a deviation to be feared, because the method had no defined room to move. Q14 and the revised Q2(R2), adopted by the ICH in November 2023, change the frame. A method is no longer a frozen recipe. It is a system with a defined purpose, an understood relationship between its inputs and its results, and a managed life that runs from development through routine use to eventual retirement.

This article walks that lifecycle. It starts where a beginner needs to start, what these two guidelines actually are and how they fit together, and builds toward the parts that matter to people who own methods in a regulated environment: how to define an analytical target profile, what the revised validation parameters require, and how to keep a method in control across change control, transfer, and continued verification. The content applies across small molecules, biologics, vaccines, cell and gene products, and the analytical work behind medical devices and combination products. The framework does not care what you are measuring, only that you can show the measurement is fit for the decision it supports.

Two guidelines, one method

ICH Q2 and ICH Q14 are deliberately a pair. Q2(R2), titled Validation of Analytical Procedures, tells you how to demonstrate that a procedure is fit for its intended use, the performance characteristics, the data you collect, the acceptance criteria you set. ICH Q14, titled Analytical Procedure Development, tells you how to get to a procedure worth validating in the first place, and how to manage it over time. Q14 was new; Q2 already existed and was revised in parallel so the two would speak the same language. Both reached Step 4 of the ICH process in November 2023, and regulators have since been adopting them into their regional frameworks, for example the EMA and the US FDA publishing the Step 4 texts for implementation.

The relationship mirrors the one between product-side guidelines that practitioners already know. ICH Q8(R2) describes pharmaceutical development by quality-by-design principles; Q9(R1) covers quality risk management; Q10 covers the pharmaceutical quality system. Q14 is, in effect, the analytical analogue of Q8, it imports the same vocabulary of design, criticality, and control strategy and applies it to measurement systems instead of to the product itself. If you have worked with a quality target product profile and critical quality attributes, the analytical target profile and analytical procedure attributes in Q14 will feel familiar.

Two important points about scope. First, Q14 explicitly allows two approaches: a minimal (sometimes called traditional) approach and an enhanced approach. The enhanced approach is the quality-by-design style, define the goal, study the variables, understand the relationships, build a control strategy. The minimal approach is the conventional one, and it remains fully acceptable. Q14 does not force every laboratory into a full design-of-experiments campaign for every assay. It offers the enhanced toolset and lets the method’s complexity and risk dictate how far you go. A simple loss-on-drying determination does not need the same development rigor as a chiral impurity method for a complex molecule.

Second, the enhanced approach can buy regulatory flexibility. When you have demonstrated genuine understanding of a method, for example, by establishing a proven acceptable range for operating parameters, Q14 opens the door to managing certain changes within that range under your pharmaceutical quality system rather than through a prior-approval regulatory submission. That is the analytical equivalent of the design-space concept on the product side. The flexibility is earned through documented understanding, not granted automatically.

How to choose between minimal and enhanced

The choice is a risk decision you should be able to defend in writing, not a style preference. Use a short justification, captured in the development report or method strategy document, that weighs:

• Criticality of the result. A method that releases product against a tight specification, or that supports a stability shelf-life claim, justifies more development understanding than an in-process check used only for directional trending.
• Method complexity and known fragility. Gradient chromatography with a hard-to-resolve critical pair, chiral separations, trace-level genotoxic impurity methods, and chemometric models all reward enhanced work because small parameter shifts move the result.
• Expected change frequency. If you anticipate column re-supply, instrument platform migration, or scale-driven sample changes over the product life, building proven acceptable ranges up front pays back later as avoided revalidation and lighter regulatory reporting.
• Lifecycle horizon. A commercial product with a 10-year-plus life is worth more development investment than a method supporting a single early-phase study.

A practical pattern many organizations adopt: minimal approach for compendial and simple physical tests, enhanced approach for the assay and impurity methods that gate release and stability of the lead product. You are allowed to mix. Q14 does not require a single approach across the whole method portfolio.

The analytical target profile

Everything in the enhanced approach hangs off one document: the analytical target profile, or ATP. The ATP is a prospective statement of what the method must deliver. It is not a description of a specific technique. It is a performance specification that any candidate technique either meets or fails to meet.

A useful way to think about it: the ATP defines the requirement; the analytical procedure is a solution to that requirement. You could satisfy the same ATP with HPLC, with a UPLC method, or potentially with a different separation chemistry entirely, as long as each delivers the required performance. This separation of requirement from solution is what makes the rest of the lifecycle coherent, because it gives you a stable reference point that does not change when the technique does.

A well-formed ATP for an assay or impurity method typically specifies:

• The quantity being measured and the reportable range, expressed against the relevant specification. For a release assay you might state a range covering, say, 70 to 130 percent of label claim; for an impurity method you anchor the range to the reporting, identification, and qualification thresholds in ICH Q3A and Q3B.
• The required accuracy and precision, often stated together as a target measurement uncertainty or as a combined performance criterion. Rather than separate, loosely connected pass/fail limits, the modern framing asks: across the reportable range, how close to the true value, and how reproducibly, must a reported result be for it to be fit for decision-making?
• Specificity expectations, the method must distinguish the analyte from the matrix, from known degradants, and from process-related impurities.
• The decision the result supports, because that is what ultimately sets how tight the performance needs to be.

The discipline of writing an ATP forces a conversation that the old approach often skipped: what does this measurement actually have to be good enough to do? A method used to release a product against a 95.0-105.0 percent specification has different uncertainty demands than one used for in-process trending. The ATP makes that explicit before any development work begins, which is also why inspectors increasingly expect to see the rationale for performance criteria rather than criteria that appear from nowhere.

A worked ATP

Concrete beats abstract. Here is a worked ATP for a finished-product assay of a small-molecule oral tablet. The same structure scales to a biologic potency assay or a residual-host-cell-protein method; only the analyte and the numbers change.

Notice what the ATP does not say. It does not name HPLC, does not fix a column, does not state a wavelength. Those are properties of the solution, chosen during development. If two years from now UPLC replaces HPLC, you evaluate the new procedure against this same ATP. The target does not move.

A short note on setting the uncertainty target. Work backward from the specification width and the decision risk you are willing to carry. If the release specification is 95.0-105.0% (a 10% window) and you want low risk of a false out-of-specification on a truly conforming batch, a total uncertainty of a couple of percent is reasonable; a method contributing 4-5% uncertainty against a 10% window will generate false failures and frustrate the manufacturing floor. This is the conversation the ATP is meant to force, and it is exactly the rationale an inspector wants to see.

Enhanced development: from ATP to control strategy

Once the ATP exists, enhanced development is the work of finding and understanding a procedure that meets it. Q14 frames this through a small set of linked concepts.

Analytical procedure attributes and parameters. The method has attributes, the measurable characteristics of its output, such as resolution between critical pairs, peak symmetry, or signal-to-noise at the reporting threshold. It has parameters, the things you set or that vary, such as mobile-phase composition, column temperature, flow rate, injection volume, gradient slope, or detector wavelength. Development is about learning which parameters drive which attributes, and how strongly.

Risk assessment. Before experimenting, a structured risk assessment, typically the same tools Q9(R1) describes, such as a cause-and-effect (fishbone/Ishikawa) analysis feeding a failure-mode assessment, identifies which parameters are likely to matter. This focuses the experimental effort. There is no value in running a full design of experiments across twenty parameters when prior knowledge and risk ranking tell you three of them dominate.

Robustness as a design input, not an afterthought. In the traditional model, robustness was a box ticked late in validation. In the enhanced model, deliberate variation of the high-risk parameters, usually through a design of experiments, happens during development. The output is an understanding of how the method behaves across ranges, which lets you define a proven acceptable range for each important parameter and, where justified, a multivariate method operable design region (MODR), the analytical analogue of a design space, within which the method is known to meet the ATP.

Control strategy. The development work culminates in a control strategy: the set of system suitability tests, defined operating ranges, and procedural controls that keep the method delivering ATP-conforming results in routine use. System suitability is the daily gatekeeper. For a chromatographic method this usually includes resolution criteria for the critical pair, repeatability of replicate injections of a reference standard (commonly expressed as a relative standard deviation limit), tailing factor limits, and a sensitivity or signal-to-noise check at the reporting level. The point of enhanced development is that these system suitability criteria are derived from the understanding you built, not borrowed from a template.

A worked development sequence

The steps below are the order an enhanced development actually runs. Numbers are illustrative but realistic for a reverse-phase gradient assay.

1. Translate the ATP into method attributes. The +/- 2.0% uncertainty target and the specificity requirement become concrete attributes: resolution between D2 and the main peak greater than 2.0, peak purity passing, replicate-injection relative standard deviation under 1.0%.
2. Scope the parameters and rank risk. Fishbone the procedure, then rank with a failure-mode and effects analysis. Suppose gradient slope, column temperature, and mobile-phase pH score highest; flow rate and injection volume score medium; detector wavelength scores low because the analyte has a broad absorbance plateau.
3. Screen, then optimize. Use a fractional-factorial or definitive-screening design across the high-risk parameters to find the main effects and interactions, then a response-surface design around the promising region to map resolution and run time as functions of those parameters.
4. Define ranges. From the surface, read off a proven acceptable range for each parameter, for example column temperature 33-37 C, mobile-phase pH 2.4-2.8, gradient slope within +/- 5% of nominal, each shown to keep critical-pair resolution above 2.0. Where parameters interact, capture the joint region as a MODR rather than independent ranges.
5. Set system suitability from the data. The daily resolution limit, the relative standard deviation limit, and the tailing-factor limit come from the observed distribution at the edges of the ranges, not from a generic template. If resolution drops to 2.1 at the worst-case corner, a system suitability limit of “not less than 2.0” is justified by data, not habit.
6. Write the development report. It carries the ATP, the risk assessment, the design and raw data, the ranges, the MODR if claimed, and the control strategy. This report is the evidence base an assessor reads, and it is what justifies any later in-PQS change.

Two failure modes are worth naming. The first is performing a design of experiments and then discarding the knowledge, running the study, declaring the method solid, and filing only a single set of fixed conditions. That throws away the regulatory flexibility the work could have earned. The second, and more common, is the opposite: claiming an enhanced approach in a submission while the underlying data only supports point estimates at nominal conditions. Reviewers and inspectors read the data, not the label. If you assert a proven acceptable range, the experimental evidence for that range has to be in the file.

Validation under Q2(R2): the revised parameters

Validation is where the method’s fitness is demonstrated against the ATP. Q2(R2) keeps the familiar performance characteristics but reorganizes and clarifies them, and crucially, it removes the old assumption that every method must be characterized for every parameter regardless of purpose. What you validate is driven by what the method is for.

The core performance characteristics remain:

The most consequential conceptual shift in Q2(R2) is the treatment of accuracy and precision together. Historically the two were validated as separate silos with separate criteria. The revision recognizes that the quantity a user actually cares about is the total error, or measurement uncertainty, of a reported result, which is a combination of systematic bias (accuracy) and random variation (precision). Q2(R2) explicitly accommodates approaches that assess them jointly, and it broadens the acceptable statistical toolkit. This aligns the validation deliverable with the question the ATP asked: is a reported result close enough and reproducible enough to support its decision?

Q2(R2) also formally embraces analytical procedures beyond classical small-molecule chromatography. It accommodates multivariate methods, for example spectroscopic procedures coupled with chemometric models, such as near-infrared with a partial-least-squares calibration, and addresses the additional validation considerations those bring, including the calibration model itself and the criteria for when it must be re-evaluated. For biological and biotechnological products, the revision is more explicit than its predecessor about methods whose response is relative rather than absolute, such as potency bioassays.

What good validation evidence looks like

For each characteristic, “done correctly” means the study design reflects how the method will actually run and the acceptance criterion traces to the ATP. The table below pairs each characteristic with a representative study design, a sample acceptance criterion, and a worked result, for the tablet assay ATP above.

A few practical points for setting validation acceptance criteria:

• Tie criteria to the ATP, not to habit. A specificity requirement of baseline resolution exists because the ATP demanded freedom from a specific interference, not because resolution greater than two is a customary number.
• Make the precision design match the routine reality. Intermediate precision should vary the factors that genuinely vary in routine use, different analysts, days, instruments, and reagent lots. A precision study run by one analyst on one instrument in one afternoon understates the variability the method will actually show, and inspectors look specifically for whether the study reflects real operating conditions.
• Document the rationale for the reduced set. If you do not validate a given characteristic, for instance, you may not need a quantitation limit for a high-level assay, state why it does not apply. The absence should be a justified decision, not a gap.
• Pre-specify the statistics. Decide before you run whether accuracy and precision are assessed jointly or separately, what model fits linearity, and how outliers are handled. Choosing the analysis after seeing the data is a finding waiting to happen.

For the protocol mechanics, the order of execution, and how to write the deliverables, see method validation execution and method validation essentials. The statistical tools behind accuracy, precision, and capability are covered in statistics in quality.

Linking validation back to the ATP

Validation closes the loop the ATP opened. The ATP set the performance requirement; validation demonstrates the chosen procedure meets it. That makes the validation report more than a compliance artifact. It becomes evidence that a specific solution satisfies a specific, pre-stated requirement, the structure an inspector or reviewer can follow without having to reverse-engineer your intent.

It also makes the consequences of change tractable. Because the ATP is stable and technique-independent, a change to the procedure can be evaluated against a fixed target. The question is never “is the method different?”, methods drift and evolve. The question is “does the changed method still meet the ATP?” That reframing is the entire basis of lifecycle change management.

Lifecycle management: change control and the continuum

A method’s life after validation is governed by the pharmaceutical quality system described in ICH Q10, operating within the GMP expectations of regions such as the EU GMP guide and the requirements of 21 CFR Part 211 in the United States. Q14 connects to this rather than replacing it. Three mechanisms carry the weight.

Change control. Every proposed change to a validated method, a new column supplier, a revised sample preparation, a different instrument platform, a modified gradient, enters change control with a risk assessment. The assessment asks two questions. Does the change stay within an already-understood range, such as an established proven acceptable range or MODR? And does the changed method still meet the ATP? The answers determine the path:

The regulatory reporting category, whether a change requires prior approval, can be filed in an annual report, or is handled through a post-approval change management protocol, is jurisdiction-specific and follows the relevant variation or supplement framework. ICH Q12 (Technical and Regulatory Considerations for Pharmaceutical Product Lifecycle Management) provides the mechanism, including established conditions and post-approval change management protocols, by which the understanding generated under Q14 can be converted into reduced reporting obligations. Q14 and Q12 are designed to work together: Q14 generates the knowledge, Q12 provides the regulatory tools to act on it. The mechanics of established conditions and post-approval change management protocols are covered in ICH Q12 lifecycle management, and the general discipline of controlling changes to validated systems in change control for validated systems.

A worked change-control decision

Take a common real case: the original column is discontinued and a replacement of nominally equivalent chemistry must be qualified.

1. Raise the change in the quality system with a description, the reason, and the affected documents (method, specification, validation report).
2. Risk-assess against the ATP. Column chemistry was studied during development; the development data show critical-pair resolution holds across two alternate columns of the same stationary phase. The proposed column is a third of that family.
3. Decide the path. Because column equivalence sits inside the understood region, the handling is verification of system suitability and a small comparative run (a few batches analyzed on old and new columns, results within the method’s known variability), not a full revalidation. Document the assessment.
4. Set the regulatory action. If the established conditions defined under Q12 cover column family rather than a specific part number, this may be a lower reporting category or even an in-PQS change. If the original filing pinned the exact column, you may owe a notification. Confirm against the regional variation framework before you act.
5. Implement, verify, and close. Update the method, run system suitability and the comparative check, confirm the ATP is still met, then close the change with the evidence attached.

Contrast that with a change that swaps the detection principle, say from UV to charged-aerosol detection for a chromophore-poor impurity. That is a different measurement principle. It is treated as a new procedure: new development against the same ATP, full validation, and very likely a prior-approval submission. Same ATP, completely different lifecycle path, because the change sits outside any understood range.

Method transfer. A method rarely stays in the laboratory that developed it. Transfer to a quality control laboratory, a contract laboratory, or a second manufacturing site is itself a lifecycle event. The common transfer models are comparative testing (sending and receiving labs analyze the same samples and compare results against pre-set acceptance criteria), covalidation (the receiving lab participates in part of the validation), revalidation at the receiving site, and, only where justified by a documented risk assessment, transfer waiver based on the receiving lab’s demonstrated equivalent capability.

The ATP and the development understanding make transfer materially easier. If the receiving laboratory knows which parameters are critical and what their acceptable ranges are, the transfer protocol can focus its scrutiny where it matters rather than re-proving the entire method. A solid system suitability test, derived from real understanding, is the single most useful thing you can hand a receiving lab, because it gives them a daily, objective check that the method is performing as intended in their hands. The most frequent transfer failure is not a method that cannot be reproduced, it is a transfer protocol whose acceptance criteria were never tied to the method’s actual performance characteristics, so a difference appears and no one can say whether it matters. A worked transfer acceptance criterion: for an assay, the difference between the means of sending and receiving labs on the same homogeneous samples does not exceed 2.0%, set from the method’s intermediate precision, not picked arbitrarily. The full transfer workflow, models, protocols, and acceptance-criterion setting is in analytical method transfer and the broader technology-transfer discipline in quality in technology transfer.

Continued performance verification. Validation proves fitness at a point in time. Continued verification confirms the method stays fit over its operational life. The day-to-day instrument of this is system suitability, every run is, in effect, a small ongoing check that the method still meets its control criteria. Above that sits trending: tracking system suitability results, control sample or reference standard results, and out-of-specification and out-of-trend investigations that implicate the method rather than the product. A method that is creeping toward its system suitability limits, or whose control-sample results are drifting, is signaling that it needs attention before it produces an unreliable reportable result. This trending links naturally to the product-side continued process verification expectation in process validation guidance, the analytical method is part of the control strategy whose ongoing performance the quality system is obliged to monitor. See continued process verification for the product-side analogue and out-of-trend investigations for handling the signals a drifting method throws off. The instrument side of continued fitness, calibration and qualification, lives in analytical instrument qualification and the calibration and metrology program.

Roles and responsibilities

Lifecycle ownership fails when no one is clearly accountable for the ATP across its life. A workable split:

The single most important handoff is from development to the method owner. If the development understanding (which parameters matter, what their ranges are, why the system suitability limits are what they are) is not transferred along with the procedure, the method owner cannot defend a change later and the regulatory flexibility evaporates. Make the development report a controlled, transferred document, not a file that stays in development’s drawer.

What inspectors and reviewers actually look for

Whether through a submission review or an on-site inspection, the scrutiny of analytical methods tends to converge on the same questions.

• Is there a clear, pre-stated requirement? An ATP, or at minimum a documented statement of intended use and performance requirements that the validation criteria trace back to. Acceptance criteria that appear without rationale are a recurring finding.
• Does the validation data support the claims? If the file claims an enhanced approach with operating ranges, the experimental evidence, including the design-of-experiments data behind any proven acceptable range or MODR, must be present and must actually support the stated ranges. Asserted understanding without underlying data is a classic gap.
• Does the precision study reflect reality? Intermediate precision that varies only trivial factors, or that was run too narrowly, draws questions about whether routine variability was captured.
• Is change control real? Inspectors look at whether method changes went through assessment, whether revalidation decisions were justified against the method’s performance, and whether the regulatory reporting category was handled correctly. A method that quietly evolved through a string of undocumented “minor” tweaks is a serious finding.
• Is the method actually in control now? System suitability data, trending, and the handling of method-related investigations show whether continued verification is a living practice or a paper exercise. A method that frequently fails system suitability, or that passes only after repeated injections, is telling its own story.

Compendial methods, those from pharmacopoeias such as the USP, the European Pharmacopoeia, or the Japanese Pharmacopoeia, deserve a specific note here. A compendial method is considered validated for its stated purpose, but it must still be verified under the conditions of actual use, per the relevant general chapter on verification of compendial procedures. Verification confirms the method works for your specific matrix and laboratory; it is a lighter exercise than full validation but it is not optional, and skipping it is a common observation. The mechanics of that verification are covered in compendial method verification.

Common mistakes and real finding patterns

Patterns that recur across inspections and submission deficiencies, stated generically:

• System suitability that masks a failing method. “Inject until it passes” is the classic. Re-injecting a reference standard repeatedly until the relative standard deviation drops, without an investigation, is data manipulation, and the audit trail shows it. See chromatography data system integrity and audit trail design and review.
• Trial injections and unreported runs. Aborted or unlabeled “test” injections that are never reconciled against the reportable result. Reviewers read the full instrument audit trail, not just the final report.
• Acceptance criteria with no provenance. Limits copied from an old method or a generic template, with no link to an ATP or a performance characteristic. When asked “why this number,” the answer is “it has always been that.”
• A precision study that flatters the method. One analyst, one instrument, one day, then a wide intermediate-precision claim built on it. Routine variability was never captured.
• Enhanced label, minimal data. A submission asserts proven acceptable ranges or a MODR, but the file holds only nominal-condition results. The claim and the evidence do not match.
• Silent method drift. A chain of small undocumented tweaks (a slightly different gradient here, a re-equilibration change there) that together moved the method outside its validated state with no change control.
• Skipped compendial verification. A pharmacopoeial method run on a new matrix without verifying it works in that matrix, then a surprise interference appears in routine use.
• Orphaned transfer criteria. Transfer acceptance limits set without reference to the method’s own precision, so a real but acceptable lab-to-lab difference triggers a failure no one can adjudicate.

Interview questions and how to answer them

If you own or audit methods, these come up. Short, confident answers that show you understand the why:

• “What is the difference between an ATP and a method?” The ATP is the requirement, written as performance and independent of technique. The method is one solution that meets it. You can change the solution without changing the requirement, which is what makes lifecycle change management coherent.
• “What changed between Q2(R1) and Q2(R2)?” Q2(R2) keeps the performance characteristics but lets you scope what you validate to the method’s purpose, formally allows accuracy and precision to be assessed together as total measurement uncertainty, broadens the acceptable statistics, and explicitly covers multivariate and biological methods. It was adopted alongside the new Q14 in November 2023.
• “Minimal versus enhanced, when do you use each?” Risk decision. Minimal for simple, low-risk, or compendial tests. Enhanced for critical, complex, frequently-changing methods where the development understanding earns regulatory flexibility. You can mix across a portfolio.
• “What is a proven acceptable range and how is it different from a MODR?” A proven acceptable range is a studied range for one parameter where the method still meets the ATP. A MODR is the multivariate region across interacting parameters where it does, the analytical analogue of a design space. The MODR captures interactions a set of independent ranges would miss.
• “How does Q14 connect to Q12?” Q14 generates the method understanding; Q12 provides the regulatory tools (established conditions, post-approval change management protocols) to convert that understanding into reduced reporting obligations. Knowledge without the Q12 mechanism stays unused; the mechanism without knowledge has nothing to act on.
• “A column is discontinued. Walk me through it.” Raise a change, risk-assess against the ATP, check whether the replacement sits inside the studied column understanding, choose verification versus revalidation accordingly, set the reporting category against the established conditions and the regional framework, then implement, verify, and close with evidence.
• “How do you set transfer acceptance criteria?” From the method’s own performance, usually its intermediate precision, not an arbitrary number. The criterion should distinguish a difference that matters from normal lab-to-lab variation.
• “How do you know a method is still in control?” System suitability every run, trending of system suitability and control-sample results, and investigation of out-of-trend or method-related out-of-specification events. Drift toward limits is the early warning to act on before a wrong result is reported.
• “What is the difference between validating and verifying a compendial method?” A compendial method is validated for its stated purpose by the pharmacopoeia. You still verify it works in your matrix and your lab under actual conditions. Verification is lighter than validation but mandatory.

Where this leaves the practitioner

The shift Q14 and Q2(R2) describe is from did you validate it? to do you understand it, and are you keeping it in control? The mechanics, the ATP, enhanced development, risk-based validation, lifecycle change control, transfer, continued verification, all serve that change in posture.

For a practitioner, the practical takeaways are concrete. Write the requirement before you choose the technique, and write it as performance, not as a method name. Spend development effort where the risk assessment says it belongs, and keep the knowledge you generate rather than collapsing it to nominal conditions. Set validation criteria that trace to the requirement and a precision design that reflects how the method will really be run. Treat every change, transfer, and trend as a lifecycle event evaluated against the same stable target. And remember that the enhanced approach is an option scaled to risk, not a mandate, a simple, low-risk method handled the minimal way is entirely compliant, and over-engineering it helps no one.

The methods that age well under this framework are not the ones with the thickest validation reports. They are the ones whose owners can answer, at any point in the method’s life, a single question with evidence: is this measurement still good enough for the decision it supports? That question is what the ATP poses, what validation answers, and what continued verification keeps answering for as long as the method is in use. For the broader quality-by-design context behind enhanced development, see quality by design and design of experiments; for how methods feed batch decisions, see batch disposition decisions; and for the lab investigations that flag a failing method, the OOS investigation process.