FDA–PQRI: Process Drift

Detection, Measurement, and

Control in Pharma Manufacturing
PQRI-FDA Workshop Summary on Process Drift
Margaret M. Szymczak, Richard L. Friedman, Rajendra Uppoor, and Avraham Yacobi

F
FDA’s Office of Compliance and Office of DA’s Office of Compliance and Office of Phar-
Pharmaceutical Science, both part of the Center maceutical Science, both part of the Center for
for Drug Evaluation and Research (CDER), and Drug Evaluation and Research (CDER), and the
the Product Quality Research Institute (PQRI) Product Quality Research Institute (PQRI) co-
co-chaired a workshop for industry on process chaired a workshop for industry on process drift in
drift in December 2010, in North Bethesda, December 2010, in North Bethesda, MD. The work-
MD. This paper summarizes the discussions and
shop was well attended by participants representing
feedback obtained during the meeting.
industry, academia, and CDER. The program received
high marks for its content and quality. The presen-
Margaret M. Szymczak is an independent tations were deemed relevant and compelling, with
pharmaceutical consultant and chairperson of
the Product Quality Research Institute (PQRI) many examples illustrating a life-cycle approach to
Manufacturing Technical Committee. Richard L. active monitoring and improving manufacturing
Friedman and Rajendra Uppoor work in FDA’s Center performance. Both APIs and myriad pharmaceuti-
for Drug Evaluation and Research. Avraham Yacobi
cal dosage forms were discussed. There was general
is an independent pharmaceutical consultant and a
member of PQRI’s Board of Directors. agreement that this workshop deserves additional
exposure. It was proposed that PQRI offer additional
Note: The views presented in this summary article programs in the US and abroad, and publish a sum-
do not necessarily reflect those of FDA or of the
organizations in which the authors are employed.
mary of the workshop. This article is intended to meet
the latter objective.

70 Pharmaceutical Technology October 2011 P h a r mTe c h . c o m

FDA–PQRI: Process Drift
Defining process drift •D
 iscussion of the impact of process drift on prod-
uct performance, safety, and efficacy.
Process drift in the manufacture of an API or a drug
product was defined by workshop attendees as:
Preventing process drift
An unintended, unexplained or unexpected trend of measured
process parameter(s) and/or resulting product attribute(s) away Prevention of process drift and continual improve-
from its intended target value in a time-ordered analysis over the
ment was a central theme of the workshop. In addi-
lifetime of a process or product.
tion to the quality benefits, workshop participants
In many instances, process drift is the consequence discussed important business efficiencies gained
of variation in a variety of process inputs, includ- from improving quality, such as reduction in inven-
ing raw materials, manufacturing personnel, and tories, unexpected shutdowns, operational costs, and
machine (man-machine) interactions or process- capital expenditures. Trend monitoring is an impor-
ing conditions. In some cases, the testing of materi- tant preventive program of a drug manufacturer to
als or measurement of process parameters may also identify impending issues and allow time to investi-
experience drift. As an example, process drift is oc- gate and resolve the problems before there is an im-
curring when a unit operation of a manufacturing pact on the patient. Statistical Process Control (SPC)
process moves toward the edge of its acceptable pro- and other time-ordered analyses can be used, ideally
cess ranges. Consequences can ultimately include in real-time or shortly after completion of the process
significant variability in drug release, content uni- stage, to identify undesirable variation and enable
formity, assay, or other product attributes. Unaccept- appropriate action. Such early-warning systems en-
able variation during processing may lead to problems able detection of intra-batch and inter-batch process
with consistency of in-process output, finished- drift in a pharmaceutical operation and are an inte-
product quality at release, or increased risk of failing gral part of a 21st-Century quality system. Suitable
specification before shelf-life expiry. It is important to quality standards for APIs, excipients, container-clo-
understand, monitor, and control process drift so that sure systems, and drug products must be designed,
action can be taken before it impacts the patient. implemented, monitored and periodically evaluated
Understanding the major sources of variation is es- by persons involved in pharmaceutical manufactur-
sential to the design and control of robust processes ing and quality control/quality assurance, including
in the manufacture of pharmaceuticals. Over time, a management. Senior management’s role was strongly
component or process may drift from its target due to emphasized as central to an organization’s success
intrinsic or extrinsic factors. These factors may initially at achieving proactive quality assurance rather than
be unknown or considered to be of lesser significance. reactive quality control.
Failure to adequately control processes and prevent de- Modern technolog y and process control ap-
fects can pose risk to patients/consumers, affect prod- proaches were discussed. Detection, diagnosis, and
uct availability, and yield undesirable regulatory and enhanced upstream control of process variables
business outcomes. can be modeled and process trajectory more tightly
The objectives and scope of the FDA–PQRI work- measured and controlled to ensure consistent mate-
shop included: rial output. With this enhanced quality information
for each unit operation, intelligent control systems
• Discussion of a life-cycle approach to monitoring can be employed to provide dynamic feedback for
manufacturing performance that assures prompt larger manufacturing process improvement efforts.
detection and correction of meaningful variation This improved knowledge will enable the impact of
• E xploration of technological and management changes to the product or process to be predicted bet-
system approaches to better identify, measure, and ter so that the undesirable impact on product quality
control ­process variation and mitigate undesirable and stability can be prevented. Ultimately, enhanced
product variability process information will facilitate a state of process

72 Pharmaceutical Technology October 2011 P h a r mTe c h . c o m

FDA–PQRI: Process Drift
control (including continuous process verification) Workshop participants recognized “process drift” as
throughout the life cycle and yield drug quality ben- an “unintended, unexplained, or unexpected trend of
efits to both industry and consumers. measured process parameter(s) and/or resulting prod-
uct attribute(s) away from its intended target value in
When robust systems are not implemented and ca-
a time-ordered analysis over the lifetime of a process
pable tools are not used to prevent process drift, result-
or product.”
ing manufacturing problems may include: low product
yield, batch delays, ingredient and packaging variabil- When a substantial number of errors that occurred
ity, batch failures, product quality-related clinical fail- in pharmaceutical manufacturing were analyzed
ures, investigations, recalls, product seizures, injunc- over a specified period, their distribution was 40%
tions, and consent decrees. due to recurring human errors, 30% due to recur-
ring process errors, and 30% nonrecurring errors, as
Detection and resolution of process drift was
stated by one of the speakers during the workshop.
discussed for APIs, key excipients, and a variety of
Human errors are due to inadequate training, poorly
dosage forms (solid orals, transdermals, topicals,
understood standard operating procedures and pro-
injectables, metered-dose inhalers and dry powder
cessing parameters, lack of skill sets and procedural
inhalers). The impact of alternate source qualifica-
control, and inadequate resources. Other sources are
tion for drug product ingredients and packaging
inadequate change management and risk manage-
components was also included. The effect of process
ment. Unsuitable equipment, deficient preventive
drift on specifications and shelf-life was addressed.
maintenance, and inadequate equipment calibration
In addition, participants discussed proper supplier
also lead to movement away from the normal process
management, product quality monitoring and con-
variability. Inadequate understanding of the inher-
trol systems (e.g., PAT), corrective and preventive
ent variability in and inappropriate characterization
action (CAPA) programs, and quality-by-design
of excipients, APIs, and other components can lead
(QbD) approaches. Each approach was considered
to unacceptable process variation and can be one of
an integral piece to establishing and maintaining a
the most complex areas to track during the life cycle
stable manufacturing system.
of a pharmaceutical product.

Breakout sessions Scientific rationale is needed to select variables on

which to focus most when monitoring process drift. For
Breakout sessions among participants focused on: raw materials, the certificate-of-analysis (CoA) testing
• The definitions and terms for describing process alone is insufficient to determine excipient variability.
variation and process drift In comparison with commonly used univariate analy-
ses, use of multivariate analytical methods for assess-
• The most frequent causes of variation in pharma-
ing properties of raw materials, in-process materials,
ceutical manufacturing
and finished products by modeling relevant material
• Current strategies for monitoring and detecting attributes and process variables can support increased
process variability quality assurance. However, appropriate education,
training, and manufacturing process understanding
• Evolutions in industry management approaches
is necessary to properly use these methods. Analytical
and the regulatory environment that could promote
method drift should also be monitored.
more proactive process improvement throughout
the product life cycle Evolutions in industry management approaches
could promote a more proactive process improve-
• Impact of process drift on product bioavailability,
ment throughout the life cycle of a product. Justifi-
safety, and efficacy.
cation to senior management is necessary to design
The following section summarizes key outcomes of the and implement quality assurance concepts. Work-
breakout sessions. shop participants agreed that upper management

74 Pharmaceutical Technology October 2011 P h a r mTe c h . c o m

FDA–PQRI: Process Drift
must support the efforts to assure drug quality and Overall, the participants felt that further work is
understand the inherent costs of poor quality. Esti- needed to develop better methods that recognize and
mation of the cost of failure can provide an indica- enhance understanding of the effects of process drift
tion of the ramifications and support the resources on product quality and product performance.
needed. Better metrics are needed to quantify the
risk–benefit ratio. Conclusion
Participants also indicated that the regulatory land- Feedback obtained from FDA–PQRI workshop attend-
scape needs to be more predictable. Some also felt that ees through a post-event survey has been positive. The
the time and cost to provide information to the regu- topic of enhancing life-cycle management of process
lators and receive approval for process improvements drift should be elaborated at future workshops, includ-
can be prohibitive. ing discussion of advancements and innovations in man-
Process drift has the potential to affect product per- ufacturing technology that allow for dynamic process
formance for all types of dosage forms. More mean- control.
ingful analytical tools for understanding and detect-
ing process drift are needed. Identification and use of Additional reading
improved input material measurements and process
1. PQRI–FDA Workshop on Process Drift program, www.
information are necessary to improve the likelihood of pqri.org/pdfs/processdrift_finalprogram.pdf.
alerting a manufacturer to process variability that may 2. PQRI–FDA Workshop on Process Drift summary, Gold
impact product quality, including drug bioavailability. Sheet, www.pqri.org/workshops/ProcDrift/imagespdfs/
For example, to date, the best in vitro marker available DrugMakers_Goldsheet_article.pdf. PT
to industry for drug bioavailability is dissolution. Even
when in vitro–in vivo correlation is not conclusively
Reprinted and posted online with permission from
established, significant emerging changes in product
Pharmaceutical Technology Oct. 2011. Pharmaceutical
quality and manufacturing consistency are detected
Technology is a copyright Advanstar Communications
by testing a compressed tablet for tablet disintegration
Inc. publication.
and drug dissolution.

76 Pharmaceutical Technology October 2011 P h a r mTe c h . c o m