P e e r R e v i e w e d : P r o c e s s Va l i d at i o n

FDA 2011 Process Validation Guidance:

Process Validation Revisited
Paula Katz and Cliff Campbell

This article contains the views and opinions of the authors and does not necessarily represent those of FDA or the
United States.

This article is based on a technical training semi-

ance). The 2011 Guidance revises and replaces FDA’s
nar presented to United States Food and Drug
Guidance for industry entitled Guideline on General
Administration policy advisors, management, and
Principles of Process Validation, issued in May 1987
field staff in Silver Spring, Maryland, in May 2012. It
(the 1987 Guideline).
summarizes the regulatory drivers that led to the pub-
The 2011 Guidance defines process validation as
lication of FDA’s 2011 Process Validation Guidance
“the collection and evaluation of data, from the pro-
for industry. In particular, the article emphasizes that
cess design stage through commercial production
process validation is a meaningful scientific endeavor
which establishes scientific evidence that a process
that strives to ensure process control and product
is capable of consistently delivering quality product.”
quality rather than a discrete and isolated activity.
The 2011 Guidance promotes a “lifecycle” approach
The article proceeds to describe practical steps that
to process validation that includes scientifically
product manufacturers can take when applying the
sound design practices, robust qualification, and
Guidance to both legacy and new product manufac-
process verification. In particular, the 2011 Guidance
turing situations. The article introduces the concept
describes process validation activities in three stages:
of “validation trilogies” as a proposed method of
• In Stage 1, process design, the commercial
aligning the three inter-related stages of process
process is defined based on knowledge gained
validation (process design, process qualification, and
through development and scale-up activities.
continued process verification) with the key concept
• In Stage 2, process qualification, the process
of product lifecycle. The article concludes with an
design is evaluated and assessed to determine if
exhortation to industry to capitalize on this updated
the process is capable of reproducible commer-
view of process validation as a means to carry out
cial manufacturing.
sound, science-based process improvements and an
• In Stage 3, continued process verification, ongo-
integrated approach to product quality.
ing assurance is gained during routine produc-
tion that the process remains in a state of control.

AGENCY CONTEXT In addition to discussing activities typical of each

stage of process validation, the 2011 Guidance pro-
Background vides recommendations regarding appropriate docu-
In January 2011, FDA announced the availability of mentation and analytical methods to be used during
a final guidance for industry entitled Process Valida- process validation. Figure 1 illustrates how the three
tion: General Principles and Practices (the 2011 Guid- stages of process validation relate to one another and

18   Journal of GXP Compliance

 Paula Katz and Cliff Campbell

Figure 1:
Stages of process validation. Used with permission of Grace E. McNally.

Stage 2
Evaluate/Confirm Process Qualification
Stage 1
(PQ)
Process
Design

Chang Design of
Process
es Facilities &
Performance
Qualification
Qualifiction
of Equipment
(PPQ)
and Utilities

Ch
an
ge
s
Stage 3
te
Continued ib u
Distribute Distr
Process
Verification

to the actions taken during and after each stage. product and process. This view of process validation
underscores the importance of detecting, under-
Regulatory Drivers and Expectations standing, and controlling sources of variability over
Nearly a quarter of a century elapsed between time in order to consistently produce safe, effective
the time FDA first issued the 1987 Guideline and drugs that meet all quality attributes. In turn, the
the publication of the 2011 Guidance. The 2011 emphasis on understanding and controlling process
­Guidance is entirely consistent with the basic prin- variability leads to a clarification that FDA expects
ciples of process validation articulated in the 1987 manufacturers to employ objective measures and ap-
Guideline–and indeed, with principles imbedded propriate statistical tools and analysis.
in the current good manufacturing practice (cGMP) Again, none of these concepts are new to process
regulations in 21 Code of Federal Regulations (CFR) validation. Rather, the 2011 Guidance reinforces
Parts 210 and 211 as published and described in the central themes of the cGMP regulations that drive
1976 preamble to those regulations. Nonetheless, successful process validation and the production of
more than 25 years worth of experience and regu- quality products over time. The 2011 Guidance un-
latory oversight, along with the cGMPs for the 21st derscores the link between process validation and
Century Initiative (1), prompted FDA to revisit the existing regulations such as the following:
principles and concepts in an effort to update and • § 211.100(a) requires “written procedures for
clarify FDA’s thinking on process validation. production and process control designed to
Among other motivating factors, FDA sought to assure that the drug products have the identity,
emphasize process design and maintenance of pro- strength, quality, and purity they purport or
cess control during commercialization. By aligning are represented to possess...” Manufacturers are
process validation activities with a lifecycle approach, required to design a process, including opera-
the 2011 Guidance communicates that process vali- tions and controls that yield a product meeting
dation is an ongoing program rather than a discrete these attributes.
and isolated activity. Under the 2011 Guidance, pro- • § 211.110(a), sampling and testing of in-process
cess validation is presented as a series of activities materials and drug products, requires that
that manufacturers carry out over the lifecycle of the manufacturers establish control procedures “to

Autumn 2012 Volume 16 Number 4 19

Peer Reviewed: Process Validation

monitor the output and to validate the perfor- common misconceptions about process validation.
mance of those manufacturing processes that One of the most widely-discussed has been the re-
may be responsible for causing variability in buff of process validation’s (perceived) three-batch
the c­ haracteristics of in-process material and requirement. Prior to the issuance of the 2011 Guid-
the drug product.” From a process validation ance, “...it was widely accepted throughout industry,
perspective, even well-designed processes and, indeed, implied or stated in some FDA guid-
must include in-process control procedures to ance documents, that process validation was a static,
assure final product quality. Furthermore, § three-batch demonstration event. (2)”. With the ad-
211.110(b) requires that in-process specifica- vent of the 2011 Guidance and its emphasis on de-
tions be “derived from previous acceptable pro- sign, lifecycle, and control of variability, the “rule of
cess average and process variability estimates three” has been effectively rejected. Although some
where possible and determined by the applica- may harp at the idea that there is no longer a magic
tion of suitable statistical procedures where number, FDA’s position remains that there never was
appropriate.” Manufacturers must continu- a three-run requirement in the first place. Despite
ally analyze process performance and control the pervasive practice of three-batch validation, note
batch-to-batch variability using appropriate that even the 1987 Guideline used the following lan-
statistical techniques. guage to describe the validation protocol:
• Sampling methodology becomes a key factor “It is important that the manufacturer prepare a
in carrying out process validation insofar as it written validation protocol which specifies the pro-
concerns monitoring and evaluating variability, cedures (and tests) to be conducted and the data to
especially in process qualification (Stage 2) and be collected. The purpose for which data are collect-
continued process verification (Stage 3). cGMP ed must be clear, and data must reflect facts and be
regulations specify that samples must: collected carefully and accurately. The protocol should
• Represent the batch under analysis (§ specify a sufficient number of replicate process runs to
211.160(b)(3)). demonstrate reproducibility and provide an accurate
• Meet specifications and statistical quality measure of variability among successive runs.”
control criteria as condition of approval and
release (§ 211.165(d)) The 2011 Guidance is deliberately less prescrip-
• The batch must meet its predetermined speci- tive than the 1987 Guideline. Under the 2011 Guid-
fications (§ 211.165(a)). ance, the process performance qualification proto-
Finally, § 211.180(e) requires that information col need not specify the number of batches to be
and data about product quality and manufacturing performed. Instead, the 2011 Guidance describes
experience be evaluated at least annually to deter- how manufacturers should develop a protocol that
mine the need for changes in specifications or manu- builds upon process design knowledge to identify
facturing or control procedures. Regular review and criteria and process performance indicators that
analysis of product quality and process performance allow for science and risk-basked decision-making
data to monitor trends is, by definition, an essential about the manufacturing process. Does the pro-
feature of continued process verification. cess consistently produce quality products? Is it in
a state of control? The 2011 Guidance emphasizes
MISCONCEPTIONS ABOUT PROCESS VALIDATION documenting and evaluating evidence that answers
these questions rather than satisfying a three-batch
Validation Protocols and the “Rule of Three” checklist. The notion that manufacturers must
In addition to revisiting important long-standing make deliberate, rational decisions about whether
principles of process validation by linking them to their specific processes are validated and their prod-
regulatory requirements, the 2011 Guidance dispels ucts ready for commercial release is hardly new. As

20   Journal of GXP Compliance

 Paula Katz and Cliff Campbell

Figure 2:
Process validation learning progression (2). Used with the permission of Grace E. McNally.

Learning progression

Good planning, expected path

Continued
Comprehensive process Sound, thorough Verification,
design, scientific process qualification Process learning and
process understanding confirms design improvement

Poor design, planning, process and understanding

Unexplained variation,
Product and process problems,
PQ checklist
Poor, minimal Process not in control.
exercise w/little
design Major learning!
understanding
Potentially substandard
product on market

the 2011 Guidance states, “[f]ocusing exclusively output. Because the 2011 Guidance is based on
on qualification efforts without also understanding the premise that process validation must be tied to
the manufacturing process and associated varia- product and process lifecycle rather than a static
tions may not lead to adequate assurance of qual- event, different attributes and parameters may
ity.” Accordingly, the 2011 Guidance recommends have different roles in the process. These may pose
an approach to process validation that is tailored to greater or lesser risks to product quality over time,
and based upon up-front learning and knowledge and manufacturers are expected to reevaluate the
about the product and process rather than simply level of risk assigned to attributes and parameters
getting to the goal of three acceptable batches (see as new information becomes available and respond
Figure 2, below). accordingly. Again, this expectation implies that
process validation is an on-going practice rather
“Criticality”–Revisiting Attributes and Parameters than a single event. Viewing process validation in
Another important change in the language adopted this light facilitates process improvements that can
by the 2011 Guidance is the notion that criticality in turn improve product quality.
is a continuum rather than a binary (“yes or no”)
state. The 2011 Guidance does not designate spe- Use of Statistics in Process Validation
cific attributes and parameters as “critical.” Instead, An additional item of note in the 2011 Guidance is
the 2011 Guidance stresses the need to exercise its emphasis on the use of statistics. As with other
control over attributes and parameters commen- elements of the 2011 Guidance, FDA’s choice to un-
surate with their risks to process variability and derscore the importance of using statistical analysis

Autumn 2012 Volume 16 Number 4 21

Peer Reviewed: Process Validation

in process validation is hardly a new idea. Indeed, deviations observed during process validation. This
the cGMPs discussed above indicate that the use cGMP regulation has also been cited in situations
of statistical tools and analyses is a required part of where firms released product despite revalidation ef-
compliance with the cGMPs for drug manufactur- forts that failed to demonstrate process robustness.
ing. The 2011 Guidance reminds manufacturers of In some cases, firms have responded to inspectional
this requirement and reaffirms the role that statis- observations with claims that they had “controls”
tics can and should play in all three stages of pro- in place to control for process variability, but ensu-
cess validation. Again, FDA is not prescriptive about ing warning letters cited 21 CFR § 211.100(a) on
this issue. The 2011 Guidance references a number the grounds that such controls for variability had
of acceptable industry standards but clarifies that not been deliberately and prospectively assessed in
manufacturers must make deliberate decisions about process validation studies. In such cases, firms rely-
which statistical tools and analyses are appropriate ing on the recommendations in the 2011 Guidance
for their products and processes. Choosing suitable might have fared better if they returned to the stage 1
statistical tools depends on factors such as the size of “drawing board” upon discovering significant prob-
the data set, and the selection of variables, attributes, lems during process qualification, rather than reach-
and parameters being used to make inferences about ing premature conclusions about process capability
process performance (process capability and process and performance and distributing product to the
stability) and product quality. market (see Figure 1).
In the context of process validation, another com-
ENFORCEMENT STRATEGY AND STATUS monly-referenced regulation is 21 CFR § 211.110,
As illustrated above, process validation has always often in connection with missteps observed dur-
been and continues to be an enforceable cGMP re- ing stages 2 and 3. For example, many firms utilize
quirement. Moreover, the 2011 Guidance, like all standard operating procedures (SOPs) in routine
FDA guidance documents, represents the current commercial production that permit batch release
thinking on the topic and does not create or con- outside of established in-process specifications.
fer any legal rights or obligations. Nothing about the Consider the continued process verification (stage
2011 Guidance changes FDA’s enforcement policy 3) implications of a SOP allowing for drug product
with respect to process validation in a strict sense. batches to be released despite some level of failures
Rather, the expectation is that the 2011 Guidance of in-process testing. There is nothing inherently
provides greater clarity regarding FDA’s expectations wrong with such SOPs, but in the context of pro-
and the types of activities that firms should conduct cess validation and depending on the facts of the
during each of the stages of process validation. case, such SOPs could lead to a violative situation.
A review of inspectional observations and warn- Under 21 CFR § 211.110(a), manufacturers must
ing letter citations since the publication of the 2011 establish control procedures that monitor the out-
Guidance indicates that FDA has continued to cite put and validate the performance of manufacturing
firms for process validation deviations across a processes that may cause variability in the charac-
wide range of product and facility types. Among teristics of in-process material and the drug prod-
the most-frequently cited regulations are 21 CFR § uct. Even if the SOP discussed above conditioned
211.100(a) and § 211.110(a)-(b). The former is typi- batch release on the proviso that no more than a
cally invoked when product quality issues and fail- certain number of units failed to meet specification,
ures can be linked to stage 1 errors (poor process then a pre-determined level of in-process specifica-
design). For example, warning letters have cited 21 tion failures would also need to trigger follow-up
CFR § 211.100(a) for “inadequate” process validation investigation(s) to determine the root cause of pro-
efforts that relied on incomplete validation reports, cess failures as part of the firm’s process control
such as reports that failed to include and evaluate all and monitoring program for cGMP compliance.

22   Journal of GXP Compliance

 Paula Katz and Cliff Campbell

From a process validation standpoint, the inability equipped to address FDA’s questions about process
to quickly detect unreliable batch operations and validation activities, but, more importantly, better
correct deviations has a clear impact on product able to utilize their own data and process under-
quality. standing to improve quality over the lifecycles of
Finally, the link between process validation their products.
and in-process specifications is also apparent in
citations and observations related to 21 CFR § INDUSTRY IMPLEMENTATION
211.110(b). For example, § 211.110(b) might be This section proposes a practical view on how man-
cited when firms blindly refer to or rely on sta- ufacturers might carry out some of the 2011 Guid-
tistical methods and tools, (e.g., using process ance’s recommendations.
capability index (Cpk) values without previously
demonstrating statistical control, understanding Getting Started
the distribution of underlying data, etc.), to sug- Unit operations (or process steps) constitute the
gest that a process is in control in spite of ob- central spine of both the manufacturing and vali-
served specification failures or variability. This dation process, and are emphasized accordingly
type of post-hoc rationalization is tantamount to within the Guidance. This is good news for manu-
“testing into compliance,” and it is not adequate facturers, and provides the first step in a systematic
under the regulations. Using statistics alone is response. Using a limited number of unit opera-
insufficient; such tools must be applied appropri- tions as building blocks, complex manufacturing
ately in order to provide valuable and meaningful processes can be designed, qualified, and verified
inferences about the state of control for a given across each of the three stages of process validation.
process and the quality attributes of the products These building blocks (c. 20-25 per process) rep-
within and between batches. The 2011 Guidance resent the foundation layer of the platform-driven
affirms the regulatory requirement that firms strategy articulated here.
make deliberate decisions about use of statistics Figure 3 is a simple but effective example of how this
in light of their own products and processes, and can be implemented in practice. There is nothing fun-
that controls and variability should be assessed damentally new here, except for the fact that the Guid-
through completion of successful process valida- ance introduces some alternative terminology, and that
tion studies. the proposed framework is depicted pictorially.
As industry becomes more familiar with the 2011 The approach can be summarized as follows:
Guidance and with FDA’s recommendations for • The purpose of a unit operation is to deliver or
executing and demonstrating process validation, protect some aspect(s) of the target product pro-
firms should be better able to show that they under- file (also known as attributes at risk).
stand how process inputs and parameters impact • The identification and management of signifi-
the safety, efficacy, and quality of drug products. cant variables constitutes the control strategy
Successful process validation is a matter of carrying for the unit operation.
out comprehensive design work, executing qualifi- • Significant variables can entail equipment mon-
cation efforts that employ meaningful performance itoring (EM), material analysis (MA), or quality
criteria and extend beyond rote checklist exercises, control (QC) testing.
and implementing process monitoring programs • Process analytical technology (PAT) is treated
that offer useful information about whether or as a means to an end, rather than a variable per
not the process remains in control (see Figure 2). se, within the framework. PAT does not replace
Manufacturers that can document these important required in-process testing and finished prod-
steps and the knowledge gained from them in a uct release testing, although it can provide real-
systematic way will find themselves not only better time data for use in such cGMP tests.

Autumn 2012 Volume 16 Number 4 23

Peer Reviewed: Process Validation

Figure 3:
Process validation framework.

• The compilation of control strategy commit- Stage 1–Process Design

ments can be prospective or retrospective Stage 1 involves the itemization of significant vari-
(new versus legacy products), and be based on ables and their rationales for each of the process’s
a combination of manufacturing experience, unit operations, followed by the definition of oper-
technical literature, quality by design, risk ating limits and related monitoring requirements/
analysis, etc. techniques for each variable. As previously indicat-
ed, associating variables with unit operations can be
Diagrams such as Figure 3 above can be readily a largely generic activity, whereas the definition of
customized based on specific requirements of partic- operating limits and methods of monitoring is more
ular products. Identity/strength/quality/etc. are more context specific. As can be seen from Figure 4, stage
accurately defined as “super-attributes,” in most cases 1 has its own internal lifecycle, with the names of
referencing a number of sub-attributes, also known the significant variables normally being known in
as critical quality attributes (chemical, physical, mi- advance of the associated operating limits. What the
crobiological). These items can be explicitly named diagram depicts, and what the Guidance invokes, is
and further quantified within the cells of the matrix. the beginning of a structured and interconnected
Support processes are also amenable to a similar level chain of validation evidence.
of analysis. Equipment monitoring equates to what The extent to which rationales should be provid-
are traditionally known as critical process parameters ed is a vexed issue. Cogency and consistency work
within the industry. Material analysis (for in-process best, the objective being to demonstrate and justify
materials) equates to what are traditionally known the linkage between significant variables and quality
as in-process controls. Note that incoming materials attributes, (i.e., this variable protects or jeopardizes
must also be in a state of control; the default indus- this attribute, the rationale being underpinned by
try response to this item is based on a combination risk assessment). The linkages can be derived empiri-
of supplier audits, quality agreements with suppliers, cally, and do not necessarily imply full-blown quality
certificates of analysis, raw material testing, and re- by design. Note also that from the point of view of a
lated monitoring programs. standardized response, pharmaceutical manufacture

24   Journal of GXP Compliance

 Paula Katz and Cliff Campbell

Figure 4:
Stage 1–Process Design.

is comprised of a relatively small number of vari- uitous validation batches, particularly during the
able types (10s rather than 100s), many of which are transition phase. In such situations, carrying out
shared across unit operations and processes–not to the recommendations envisioned by the 2011 Guid-
mention organizations. ance is based on the proviso that the validation re-
port makes a commitment to an ongoing monitor-
Stage 2–Process Qualification ing and review program. On that basis, processes
Stage 2 is a seamless extension of stage 1, and in- can be provisionally declared to be “in control” rela-
volves the definition and execution of a testing strat- tive to the level of evidence available when the dec-
egy on behalf of the variables itemized and quan- laration was made. This is a key aspect of the 2011
tified in stage 1. As with stage 1, stage 2 also has Guidance; validation is an unequivocal function of
its own internal lifecycle, approximating to the ac- time and the inferences permissible based on the
ceptance criteria, protocol preparation, protocol ex- available data.
ecution, and report aspects of traditional validation. In regard to satisfying the “in control” expecta-
What the 2011 Guidance is emphasizing here, to the tion, what this means in practice is that control
dismay of diehards, is that a testing strategy is only charts can be initiated for all significant variables
as good as the corresponding sampling and analysis during stage 2, and this continues until a sufficient
commitment, and that assertions to the effect that number of batches have been manufactured to en-
“this process is validated” only carry weight if all of able a declaration to be made to the effect that “this
its significant variables are in a state of control. This process is currently capable–for these variables.”
is conveyed schematically in Figure 5. This is easier said than done when dealing with low
Many organizations are of the view that GMP volume products and statistically insignificant data-
compliance will continue to require three ubiq- sets. In such cases, statistical inferences should not

Autumn 2012 Volume 16 Number 4 25

Peer Reviewed: Process Validation

Figure 5:
Stage 2–Process Qualification.

be contrived, but the spirit of the 2011 Guidance nique, allowing manufacturers deliberate scope in
can still be satisfied and defendable conclusions this area.
drawn, to the effect that “these variables are within
their operational limits–and therefore in a state of Stage 3–Continued Process Verification
control, at this time.” Stage 3, while conceptually straightforward, is prov-
Fitness for purpose of facilities and equipment ing to be problematical for a number of manufactur-
is an obvious prerequisite of process qualification. ers. This is due in part to the perception that the re-
The 2011 Guidance actually incorporates this as a quirement for process monitoring is totally new. The
stage 2 activity, but it has been excluded from Fig- 2011 Guidance merely formalizes what was always
ure 5 above for the sake of simplicity. The key point an implicit expectation. Significant variables quanti-
here is that systems and components must (as was fied at stage 1 and qualified at stage 2 should be sub-
always the case) be suitable for their intended use ject of continued process verification (CPV) during
and perform properly. Taking metrology as an ex- routine manufacturing at stage 3. The expectation is
ample, the “intended use” stipulation merely means summarized schematically in Figure 6.
that instruments have been specified, calibrated, A logical strategy in regard to stage 3 implemen-
and maintained relative to their process duty, tation takes the following course. As part of the
namely the measurement/control of significant “handshake” between stage 2 and stage 3, continue
variables with defined tolerances. This also squares monitoring all significant variables until sufficient
the circle in regard to instrument “criticality,” such data has been acquired to enable process capabil-
instruments being those that measure or control ity to be declared, the default monitoring frequency
significant variables. Note that the 2011 Guidance here being “every batch.” Take remedial action for
is non-committal in regard to qualification tech- any rogue variables in parallel with the monitoring

26   Journal of GXP Compliance

 Paula Katz and Cliff Campbell

Figure 6:
Stage 3–Continued Process Verification.

effort. Assess the data on completion, and confirm ment free, but rather that the documented evidence
that all variables are “in control”. For each unit oper- of compliance with process validation regulatory
ation, select one or two leading variables, (i.e., those requirements is migrating to a data-centric repre-
that are predictive of process performance or process sentation, whether this be captured in hard copy or
distress). Focus the ongoing stage 3 monitoring pro- electronic format.
gram on these variables, along with any “intensive When implementing stage 3, manufacturers
care” variables that may also be in play. Continue to should consider the semantic difference between
capture process performance for the remaining vari- the terms “continued” and “continuous”. The 2011
ables via the batch record or its attachments. Once Guidance deliberately speaks to continued process
sufficient data have been acquired, assign alert/ac- verification, which some organizations have misin-
tion limits for leading variables. Using manual or au- terpreted to mean continuous, with mandatory en-
tomated data acquisition procedures, monitor these ablement via PAT. The expectation is decidedly not
variables for stability (i.e., absence of drift) as well that in-process or release testing required under the
as capability (i.e., within operating limits) as close cGMP regulations be replaced by PAT approaches.
to real time as is practical. Recalculate alert/action Rather, the expectation is for ongoing, (i.e., inter and
levels once sufficient data become available, not sim- intra-batch, monitoring, and review).
ply on a quarterly or annual basis. Synchronize the Process owners are encouraged to compile inter-
above efforts with the facility’s alarm management, batch data registers for their significant variables,
event logging, and dashboard systems, these items these forming the basis of CPV control charting
being considered as facilitators of stage 3. and process monitoring programs. Process owners
This is another key aspect of the 2011 Guidance; should also reflect on the term “significant” when
the review process is dynamic and data-driven designing their CPV programs. With “significant”
rather than static and document-based. That is not comes “significance,” the implication being that
to say that validation has suddenly become docu- material attributes, process parameters, and in-

Autumn 2012 Volume 16 Number 4 27

Peer Reviewed: Process Validation

Figure 7;
Process validation trilogy.

process controls are no longer monitored in isola- PROCESS VALIDATION TRILOGIES–PER SIGNIFI-
tion, but visibly correlated against the associated CANT VARIABLE
product attributes that they are intended to deliver From a manufacturer’s point of view, implementing a
or protect. There is ample opportunity for imagina- new guidance typically involves fine-tuning or mod-
tive and ergonomic control chart design and revi- ifying existing policies and procedures. Depending
sion here. on the level of validation maturity within the organi-
Manufacturers should not be intimidated by the zation, such an approach may not always be sufficient
degree of statistical know-how that compliance or appropriate. The traditional approach to compli-
with the relevant process validation cGMP regula- ance is based on an established lifecycle, with the
tions and recommended stage 3 activities may seem phases of validation occupying pole position within
to imply. As a benchmark, early warning track and the model. Because these phases are disconnected
trigger systems have been in place within the clini- in time, manufacturing systems and processes, and
cal setting for many years, with relatively little by their significant variables, reappear in a diversity of
way of statistical sophistication (e.g., “contact doc- plans, protocols, and reports, often inconsistently
tor for early intervention if patient triggers one red and incompletely across their lifecycles.
or two amber scores at any one time”). For all their From a knowledge management perspective, such
simplicity, such systems really do pack a punch, fragmentation is counterintuitive and not conducive
capturing multiple variables, including risk catego- to process understanding or economy of compliance.
ries, alert levels, and response mechanisms within The emergence of the 2011 Guidance provides in-
a single chart. CPV 101 can follow a similarly fru- dustry with an opportunity to reassess the suitabil-
gal course, with specialist support from in-house ity of existing methods to satisfy the requirements
or contracted statistical resource being provided as of a risk-based approach. For example, the simple
required. manoeuvre of flipping the X and Y “axes” of the cur-

28   Journal of GXP Compliance

 Paula Katz and Cliff Campbell

rent model (such that X = significant variable and Y part is encouraged to avail of the opportunities for
= validation phase) results in all of the expectations streamlined compliance that the new Guidance in-
of the 2011 Guidance being immediately facilitated vites. Manufacturers with a thorough understanding
(see Figure 7). of their processes and platforms, supported by an
As a result of this inversion, significant variables interconnected and systematic approach to the in-
become the dominant item within the lifecycle, formation lifecycle, have nothing to fear from the ar-
with the phases of validation providing the requi- rival of the Guidance. Defining and cross-correlating
site books of evidence (also known as trilogies) on a significant variables in the first place, and monitor-
subordinate basis. The fact that the three chapters of ing/improving their performance across time in the
a variable’s biography are written sequentially rather second, is a highly logical and valued-adding activity
than simultaneously–or in reverse order for legacy that should be shared by process owners and quality
products–is a technicality. And again, there is noth- units in order to reduce negative quality outcomes
ing radically new here, the proposal being in total and improve process performance.
accord with a V-model approach.
The challenge/opportunity for industry is to cul- REFERENCES
tivate a mindset that is prepared to revisit those as- 1. FDA, Pharmaceutical cGMPs for the 21st Century - A Risk-
pects of validation practice that are proving to be Based Approach, 2004, available at: http://www.fda.gov/
inadequate or unfit for current purpose. Manufac- Drugs/DevelopmentApprovalProcess/Manufacturing/Ques-
turers have little trouble in assembling libraries of tionsandAnswersonCurrentGoodManufacturingPracticesc-
documentation for the many thousands of technical GMPforDrugs/ucm137175.htm
items within their care. It should not be too arduous 2. FDA, “SUPAC-IR Questions and Answers about SUPAC-IR
to compile a more compact and informed narrative Guidance,” 1997, available at http://www.fda.gov/Drugs/
that provides accurate line-of-sight for significant GuidanceComplianceRegulatoryInformation/Guidances/
variables across their lifecycles. Whether such an ucm124826.htm. (See, e.g., “SUPAC-IR Questions and
initiative constitutes incremental or step change is Answers about SUPAC-IR Guidance,” describing the “first
for readers to consider. three production batches” as “validation batches;” Guide
to Inspections of Oral Solid Dosage Forms Pre/Post-Approval
CONCLUSIONS Issues for Development and Validation, 1994, available at
FDA’s 2011 Guidance on Process Validation sets available at: http://www.fda.gov/ICECI/Inspections/Inspec-
out a framework that is entirely consistent with tionGuides/ucm074928.htm, indicating that “at least three
longstanding principles and existing regulatory re- batches are needed to demonstrate consistency.”). GXP
quirements. By aligning process validation activi-
ties and expectations with the lifecycle concept, the ABOUT THE AUTHORS
2011 Guidance offers a perspective that underscores Paula Katz, JD is a Senior Policy Advisor in the Office of Manu-
facturing and Product Quality at CDER’s Office of Compliance.
the importance of risk and science-based decision
She earned her J.D. at the University of Virginia School of Law
making from the outset of product design, through and B.A. at the University of Virginia. She can be contacted at
process qualification, and into continued process [email protected].
verification. The 2011 Guidance illustrates that us- Cliff Campbell, B.E., is an independent consultant focused on
ing objective measures to detect, understand, and simplification and economization of the validation lifecycle.
He earned his B.E. at University College Cork, specializing in
control sources of variation can ultimately improve
Control Engineering. He can be contacted at cliffcampbellcon-
product quality and safety over time. Industry for its [email protected].

Article Reprint posted with permission from Autumn 2012, issue of Journal of GXP.
Copyright 2012, an Advanstar Communications Inc. publication. All rights reserved. www.ivtnetwork.com

Autumn 2012 Volume 16 Number 4 29