Content uploaded by James K Drennen, III
Author content
All content in this area was uploaded by James K Drennen, III on Feb 27, 2014
Content may be subject to copyright.
The Financial Returns on Investments in Process
Analytical Technology and Lean Manufacturing:
Benchmarks and Case Study
Robert P. Cogdill &Thomas P. Knight &
Carl A. Anderson &James K. Drennen III
Published online: 18 October 2007
#International Society for Pharmaceutical Engineering 2007
Abstract The combined deployment of process analytical
technology (PAT) and Lean manufacturing offers extraor-
dinary financial opportunities for pharmaceutical manufac-
turers at every scale. While many articles have been
published describing the economic and quality opportuni-
ties presented by improved pharmaceutical manufacturing
performance, greater understanding of the financial benefits
of PAT and Lean at the individual company level is needed
to support accurate valuation of corporate investments in
manufacturing performance upgrades. This paper describes
research using industrial benchmarks and published data for
publicly-traded companies to demonstrate the value poten-
tial posed by combined deployment of PAT and Lean in
pharmaceutical manufacturing operations. A method of
estimating the financial return on investments in PAT and
Lean is described by considering their impact on the
profitability of a hypothetical mid-sized generic pharma-
ceutical manufacturer. The results of the case study show
that based on benchmark data for the generic drug
manufacturers, it is possible to return savings of up to 6%
of revenues by improving process capability and supply
chain management through strategic deployment of PAT
and Lean manufacturing.
Keywords PAT .Lean manufacturing .Process capability .
Supply chain management .Pharmaceutical manufacturing
Abbreviations
C/T Process cycle time
cGMP Current good manufacturing practice
CoQ Cost of quality
Cpk Process capability index
EBITDA Earnings before interest, taxes, depreciation,
and amortization
FG Finished goods (inventories)
KPI Key performance indicator
NVA Non-value-added
OTC Over-the-counter
QbD Quality by design
PAT Process analytical technology
RAROI Risk-adjusted return on investment
%RFT % Right-first-time
RM Raw materials (inventories)
RTR Real-time-release
SKU Stock-keeping-unit
TQM Total quality management
VAR Value-added ratio
WACC Weighted-average cost of capital
WIP Work-in-progress (inventories)
J Pharm Innov (2007) 2:38–50
DOI 10.1007/s12247-007-9007-x
R. P. Cogdill (*)
Duquesne University Center for Pharmaceutical Technology,
410A Mellon Hall, 600 Forbes Ave.,
Pittsburgh, PA 15282, USA
e-mail: cogdillr@duq.edu
URL: www.dcpt.duq.edu
T. P. Knight
Invistics Corporation, 5445 Triangle Pkwy.,
Suite 300, Norcross, GA 30092, USA
URL: www.invistics.com
C. A. Anderson
Strategic Process Control Technologies, LLC,
306 Winter Run Lane, Mars, PA 16046, USA
URL: www.spctechllc.com
J. K. Drennen III
Mylan School of Pharmacy, Duquesne University,
600 Forbes Ave., Pittsburgh, PA 15282, USA
URL: www.pharmacy.duq.edu
Introduction
The combined deployment of process analytical technology
(PAT) and Lean manufacturing offers extraordinary financial
opportunities for pharmaceutical manufacturers at every
scale. In the few years since the kickoff of FDA’s twenty-
first century cGMP and PAT initiatives [1,2], many articles
have been published describing the economic and quality
opportunities presented by improved pharmaceutical manu-
facturing performance. In general, though, financial oppor-
tunities have been described in terms of aggregate impact on
the industry [3–6] or, more recently, benefits to public health
and the overall economy [7]. While such projections are
interesting and worthwhile, greater understanding of the
financial benefits of PAT and Lean at the individual company
level is needed to support accurate valuation of corporate
investments in manufacturing performance upgrades.
This paper describes research using industrial bench-
marks and published data for publicly-traded companies to
demonstrate the value potential posed by combined
deployment of PAT and Lean in pharmaceutical manufac-
turing operations. A method of estimating the financial
return on investments in PAT and Lean is described by
considering their impact on the profitability of a hypothet-
ical mid-sized generic pharmaceutical manufacturer. The
remaining portions of this paper describe PAT and Lean
manufacturing, how they work synergistically to unlock
value from operations, and current benchmarks for operat-
ing performance in the pharmaceutical industry. Most
details regarding the implementation of systems for PAT
and Lean are not discussed since they are often quite
subjective, and are beyond the scope of this work. While
some of the financial and operational assumptions used for
this work are specific to the presented case study, the
estimation techniques presented are generally applicable.
PAT and Efficient Quality Management
The main goal of PAT has often been summarized as simply
being to “achieve greater process understanding”and
thereby mitigate risks to product quality. Perhaps to the
detriment of the rate of implementation, the potential for
financial returns on investment in PAT has been treated as a
secondary issue. This should not be unexpected, though,
since the FDA has been clear that improving quality
assurance and streamlining oversight activities has been
their goal in promoting PAT and reforming cGMPs. While
increasing process understanding is a laudable goal, it
doesn’t necessarily induce much action on the part of
manufacturers if it is unclear as to whether the integration
of process analytics into their operations will yield positive
or negative returns.
Despite all of this, however, there is ample evidence that
process analytics can be implemented with an expressed
goal of improving efficiency and profitability so long as the
new technology’s impact on process quality assurance is
positive (as detailed in advance, e.g. by a project compa-
rability protocol). Furthermore, as noted in the Agency’s
PAT guidance document [2]:
…(PAT) is intended to support innovation and efficiency
in pharmaceutical development, manufacturing, and
quality assurance.
The guidance continues with:
…(efficient pharmaceutical manufacturing) is a critical
part of an effective U.S. health care system. The health
of our citizens depends on the availability of safe,
effective, and affordable medicines.
Thus, process efficiency is an important consideration
for patients and employees as well as shareholders; the
potential for economic impact should be afforded ample
consideration when considering the merit of investments in
PAT. According to the PAT guidance, gains in quality,
safety and/or efficiency are likely to come from:
&Reducing production cycle times (C/T) by using on-, in-,
and/or at-line measurements and controls
&Preventing rejects, scrap, and re-processing
&Real-Time-Release (RTR)
&Increasing automation to improve operator safety and
reduce human errors
&Improving energy and material (resource) use and
increasing capacity
While the list is accurate, it is neither exhaustive nor
instructive; the Agency has necessarily left it to companies
to determine for themselves the financial value potential of
PAT projects.
Process analytics do not inherently add value to an
operation. Rather, PAT methods and tools enhance man-
agement strategies such as Lean manufacturing and Quality
management which are designed to maximize yield and
efficiency, thereby allowing companies to retain a greater
share of the economic value added by operations. Hence,
the investment value of PAT depends upon the degree to
which it is integrated with corporate quality and Lean
initiatives.
Quality Management and the Total Cost of Quality
Within the context of this work, quality management refers
to practices used to systematically identify and suppress
operational and design factors which contribute to product
quality variation. Based on the description provided by the
FDA guidance, the objectives of PAT systems are closely
J Pharm Innov (2007) 2:38–50 39
analogous with popular quality management systems, such
as Six Sigma [8] and Total Quality Management (TQM)
[9]. Investments in quality management systems generate
financial returns by lowering the total cost of quality
(CoQ). The “P-A-F”CoQ model, established by Armand
Feigenbaum in 1956, groups quality costs according to
elements of Prevention, Appraisal, and Failure (internal and
external), each of which is described in the following
paragraphs [10–12].
Failure costs are the price of operating imperfect
processes. Internal failure costs are incurred as defective
units (i.e. batches of product) are identified prior to release
for distribution, resulting in expenditures for investigation,
documentation, rework (and re-inspection), or write-off and
disposal. The financial burden of internal failure costs for an
operation are a direct function of the ability of the process to
maintain product quality within specified limits, as charac-
terized by process capability indices such as the Cpk:
Cpk ¼min USL m
3s
;
mLSL
3s
ð1Þ
where, μand σare the mean and standard deviation, and
USL and LSL are the upper and lower specification limits,
respectively, for a product quality measurement. The
process capability index, Cpk, is directly related to the so-
called “process sigma”such that a 6σprocess corresponds
to a Cpk of exactly 2.00, or 2.0 defective parts per billion
(PPB), assuming ∼N(0,σ) quality variance distribution (an
alternative calculation for process sigma estimates 3.4
defective parts per million for a 6σprocess; see “Appendix”
for discussion on the calculation of process sigma) [8].
External failure costs are incurred as defective units are
identified following shipment to distributors or consumers,
resulting in expenditures for complaint investigations,
product recall, and penalties. External failures can be the
result of finite process capability, as well as product design
(e.g. degradation or lack of performance). Many CoQ
models include intangible and opportunity costs related to
(internal and external) failures; such costs might include
erosion of brand image, loss of confidence by supply chain
partners, or, perhaps, increased regulatory scrutiny. Pharma-
ceutical manufacturers are required by law to provide
assurance that products released for distribution meet strict
quality criteria; thus, the expected value of external failure
costs should be zero. Since the actual capabilities of
production and inspection systems are finite, however,
external failure costs should be built into operating CoQ
and risk models.
All other quality costs are attributable to either preven-
tion or appraisal activities. Appraisal costs are operating
expenses associated with determining the degree of confor-
mance to quality requirements, including inspection of raw
materials (RM), work-in-process (WIP), and finished goods
(FG), as well as the cost of maintaining inspection systems
(calibration, maintenance, depreciation, labor, etc.), whether
conducted by the quality unit, production, or an external
laboratory [10]. The labor costs associated with quality
control and assurance (QC/QA) can account for more than
two thirds of operating labor, or approximately 10% of
revenues of pharmaceutical manufacturers. The significant
financial burden imposed by such extensive quality units is
the result of operating for many years with quality systems
which were geared more toward “quality by inspection.”
Prevention costs, which are incurred in keeping failure
and appraisal costs to a minimum [10], can be more
difficult to evaluate since many standard protocols, such as
process and product design, new product review, and
supplier evaluations have built-in aspects of defect preven-
tion. There are many defect prevention investments,
though, for which costs can be more directly appraised,
such as quality training, quality risk assessment, and
process controls. Investments in defect prevention, such as
quality by design (QbD), continuous improvement, and
PAT-enabled controls, which are emphasized by FDA’s
twenty-first century cGMPs, reduce total CoQ by increas-
ing process capability, which both reduces failure costs and
enables reduced spending on inspection activities.
Lean Manufacturing
Lean manufacturing describes a management philosophy
and associated practices concerned with improving profit-
ability by systematic identification and suppression of
activities which contribute to waste. Wasteful activities (or
muda) which must be minimized include overproduction,
waiting, transportation, inappropriate processing, unneces-
sary inventory, unnecessary motion, and production of
defective units; all of which have the effect of increasing
the proportion of non-value-added (NVA) activities and,
therefore, process cycle time. Processes which have a high
percentage of NVA activities will by definition have
relatively high C/T, and will correspondingly have a low
rate of inventory turnover. Inventory turnover scales
directly with the inverse of C/T, and is estimated by
calculating the ratio of cost of goods sold (COGS) to the
average value of inventories over the same time period.
Total inventory turnover rate, along with the rates of RM,
WIP, and FG turnover, and process value-added ratio
(VAR) [13], are key performance indicators (KPI) of supply
chain and working capital management effectiveness.
Low inventory turnover can also be an indicator of
hidden problems in production; operations plagued by
frequent batch failures, for example, utilize overproduction
to maintain a “safety stock”of WIP or FG inventories to
40 J Pharm Innov (2007) 2:38–50
avoid downtime or reduced customer service (e.g. delayed
shipment of orders). As process capability degrades,
variance in C/T increases, or if there are many products
which must share the same production flow path, it is
commonplace for some managers to build even greater WIP
queues to maintain high capacity utilization rates [14].
While RM, WIP, and FG queues ensure that “no people or
machines sit idle,”they also ensure that plenty of people are
busy transporting, documenting, and monitoring the inven-
tories. Moreover, the value of inventories can consume a
significant portion of financial working capital, rendering it
unavailable to be deployed for potentially more valuable
investments.
Investments in Lean manufacturing systems generate
financial returns by reducing average inventory levels,
thereby reducing both carrying costs, and working capital
[15]. Perhaps as a consequence of FDA’s definition of PAT,
which is necessarily focused on quality management, Lean
manufacturing has not typically been considered as an
integral component of PAT implementations. This is
unfortunate, however, since the deployment of PAT is
critical to achieving the maximum benefit of Lean
manufacturing initiatives. Specific PAT projects enable
Lean operation by enabling real-time-release (RTR), by
reducing the C/T of operations, and by improving the
predictability (i.e. reducing the variability of) C/T.
In addition to overhead and working capital savings
related to inventory management, the value of improved
customer service is often overlooked as a driver for
implementing Lean manufacturing in the pharmaceutical
industry. While branded drug manufacturers are less
threatened by the availability of alternative suppliers for
their products, manufacturers of generic and over-the-
counter (OTC) drug products have significant motivation
to maintain low order lead-times as a way of differentiating
their products (and avoiding late-delivery penalties) in a
make-to-order marketplace. Alternatively, considering
make-to-stock scenarios, such as an inventory buildup in
advance of a new product launch or period of market
exclusivity, Lean manufacturing can be an important factor
in assuring that a sufficient, timely supply of product is
available (during the challenges of initial operation of a
new process).
Pharmaceutical Manufacturing Benchmarks
Cost of Quality
Unfortunately, even though CoQ is a well-accepted concept
in operations management, few companies have imple-
mented rigorous CoQ systems; as a result, management
tends to drastically underestimate the total magnitude of
quality costs paid by their organizations [11]. Consequently,
no specific benchmark data for CoQ in the pharmaceutical
industry was found. There are, however, many different
published estimates of process capability for the pharma-
ceutical industry. Some recent estimates of percent right-
first-time (%RFT), or yield, range from 85–95% [4,16],
90–95% [6], 2–3σ(96–99% RFT) [5]. A recent industrial
benchmarking study conducted by Macher and Nickerson
(with the cooperation of the FDA) estimated the average
process yield for a sample of pharmaceutical manufacturers
to be approximately 96% [17], which corresponds to Cpk
of ∼0.7, or ∼2.0σ. Despite the lack of published quantitative
data for the relative magnitude of individual cost compo-
nents of internal failure for the pharmaceutical industry, it
should be apparent that internal failure costs are greater
than simply the 4% of production which must be either
reworked or scrapped.
Lean Manufacturing
Estimates from the Macher and Nickerson report indicated
average WIP C/T of approximately 27 days, which implies
approximately 13.5 WIP inventory turns per year [17].
Despite the fact that the observed variance of reported C/T
data was quite extreme (the standard deviation of C/T was
greater than 28 days), the published estimates are corrob-
orated by WIP turn rates calculated from the industrial
financial data shown in Table 1and Fig. 1.
Since the total time to complete the most common unit
operations required for solid dosage manufacturing is far
less than 27 days, it is apparent that pharmaceutical
production operations are plagued by quite high levels of
NVA activities, or, alternatively, operate at very low VAR
[13]. Turnover rates for the companies reporting inventory
data are shown in Fig. 2.
It is sometimes opined that pharmaceutical manufac-
turers are limited in their ability to increase inventory
turnover due to the need to maintain large safety stocks of
FG inventories. The fact that non-pharmaceutical manufac-
turers held substantially greater FG safety stocks as a
proportion of total inventories tends to contradict this
theory.
The total inventory turnover rates for mid-sized and
generic manufacturers are roughly similar to the turnover
rates shown for larger companies on an absolute basis
(Fig. 2). After adjusting for the relationship between
inventory turnover and gross margins, however, it can be
seen that generic drug manufacturers are unique in having
particularly low rates of inventory turnover (Fig. 3). A 2006
report by Supply Chain & Logistics Canada indicates that,
surprisingly, pharmaceutical supply chain management
performance has actually degraded relative to other manu-
facturing industries in recent years. While North American
J Pharm Innov (2007) 2:38–50 41
manufacturers in all industries have increased their aggre-
gate supply chain efficiency by 20% since 1992, inventory
turnover has actually decreased by 36% for pharmaceutical
manufacturers in the same survey [18].
The Cost of Inventory Management
The annual cost of poor C/T performance can be estimated
by considering the carrying costs associated with main-
taining excess RM, WIP, and FG inventories. Inventory
carrying costs include two major components––weighted-
average cost of capital (WACC), and overhead.
WACC is a company-specific measure of the cost of
allocating capital to internal investments which includes
components related to financing costs (e.g.––interest rate on
debt), as well as opportunity costs incurred by consuming
working capital to finance inventories (e.g.––the expected
return on equity from alternative internal investments). De-
spite the fact that the cost of debt financing and risk premi-
ums are near all-time lows, WACC for the pharmaceutical
industry is still quite high due to the opportunity costs of
capital allocation. Estimates based on January 2007 data
indicate WACC for drug companies of 11.97% [19].
Overhead costs associated with inventory carry include
items such as expiration loss, cost of facilities, insurance,
paperwork, transportation and physical handling, spillage/
damage, and theft/pilferage. While published data for
inventory overhead costs for the pharmaceutical industry
were not found, a survey on inventory carrying cost
estimates suggests the combination of WACC and overhead
is likely to be greater than 25% [20]. The Supply Chain &
Logistics Canada report indicates, however, that 20% is the
generally-accepted standard inventory carrying cost rate
assumed by pharmaceutical industry management profes-
sionals [18]. Hence, based on the more-conservative
estimate of inventory carrying costs, companies can expect
to return 20¢ in ongoing cost savings for every dollar’s
worth of inventory reduction. In addition to carrying-cost
savings, however, every dollar of inventory reduction yields
an equivalent one-time cash return of working capital
which can be applied toward debt reduction or more
profitable alternative investments in growing the company
(e.g. R & D).
Case Study
Description of Operations
Consider the situation for a mid-sized company which
manufactures small-market branded, generic, and OTC
solid oral pharmaceutical dosage forms having a market
capitalization of ∼$1 billion; which is roughly the size of
Table 1 Tabular comparison of market cap, breakdown of revenues, components of inventories, and inventory turnover rates for major branded, mid-sized and generic, and non-pharmaceutical
process-based manufacturers
Major branded pharma Mid-sized/generic pharma Non-pharma manufacturers
Median ($Bil.) High ($Bil.) Low ($Bil.) Sample (n) Median ($Bil.) High ($Bil.) Low ($Bil.) Sample (n) Median ($Bil.) High ($Bil.) Low ($Bil.) Sample (n)
Market Cap. 120 190 67 10 5 37 0.95 9 30 200 0.50 14
Components of revenue Median (%) High (%) Low (%) Sample (n) Median (%) High (%) Low (%) Sample (n) Median (%) High (%) Low (%) Sample (n)
COGS 26 44 16 10 51 71 22 9 52 85 26 14
R & D 15 21 10 8 14 4 <1 8 0
SG & A 32 37 27 21 35 15 30 63 3
Operating margin 27 34 9 8 29 −23 13 19 −6
Components of inventory Median (%) High (%) Low (%) Sample (n) Median (%) High (%) Low (%) Sample (n) Median (%) High (%) Low (%) Sample (n)
Raw materials (RM) 21 31 8 7 35 63 22 9 22 47 <1 8
Work-in-process (WIP) 35 68 24 15 33 1 9 25 4
Finished goods (FG) 41 54 27 49 64 36 65 84 38
Inventory turn rates Median (turns) High (turns) Low (turns) Sample (n) Median (turns) High (turns) Low (turns) Sample (n) Median (turns) High (turns) Low (turns) Sample (n)
Raw materials (RM) 13 19 6 7 8 12 2 9 23 4153 11 8
Work-in-process (WIP) 7 14 2 15 8 5 36 218 15
Finished goods (FG) 6 8 5 4 7 2 8 30 3
All data were gathered from 2006 annual reports from publiclytraded companies. Non-pharmaceutical companies were chosen randomly from manufacturing industries having similar processes to
pharmaceutical manufacturers (chemical manufacturing), and from food manufacturers who are expected to face similar requirements for cGMP production and inventory controls.
42 J Pharm Innov (2007) 2:38–50
the smallest public drug company surveyed for this
research. While the company utilizes only a single
technology platform based on high-shear wet granulation,
dozens of SKUs are scheduled through the flow path to
generate approximately $450 million in annual sales. The
average C/T map for the company is shown in figure
(Fig. 4). Only 24% of the average WIP cycle is related to
value-added activities, and only 3.7% of the total C/T is
related to value added (VAR = 0.037). It is important to
distinguish between value-added and “required”activities;
while multiple QC activities are required for this operation,
they add nearly 20 days of NVA to the average C/T. The
company reports inventory cycle times for RM, WIP, and
FG of 46, 25, and 91 calendar days, which translates to
inventory turnover rates of approximately 8, 14.6, and 4 per
year, with an overall inventory turn rate of 2.25 per year. In
line with its peers, the company maintains process
capability of ∼0.7, or ∼2.0σ. Since there is no formal
CoQ accounting system in place, however, the total
expenditure on maintaining quality for each unit of
Fig. 1 Graphical comparison of
(top to bottom) the breakdown
of revenues, components of
inventories, and inventory
turnover rates for major, brand-
ed (a), mid-sized and generic
(b), and non-pharmaceutical
process-based manufacturers.
The vertical bars in the lower
graphs correspond to maximum,
median, and minimum turnover
rates
Fig. 2 Graphical comparison of
total inventory turnover for ma-
jor branded, generic and mid-
sized, and non-pharmaceutical
process-based manufacturers.
The heights of the segments
within each bar correspospond
to the portion of total invento-
ries for each company accounted
for by raw materials, work-in-
progress, and finished goods.
Cross-hatched bars indicate
companies reporting incomplete
inventory data; negative inven-
tories correspond to accounting
adjustments
J Pharm Innov (2007) 2:38–50 43
production is unknown. Based on the reported breakdown
of revenue components shown in Fig. 5, the company
maintains a 10% operating margin, and is valued at
approximately 22× EBITDA (earnings before interest,
taxes, depreciation, and amortization).
Project Strategy
Facing an increasingly competitive marketplace and relent-
less pressure on margins from its biggest customers, the
management team decides to undertake a manufacturing
performance campaign focused on reducing CoQ and
overhead expenses related to low supply chain velocity. In
order to maximize the impact of their initiative, the
company incorporates selected PATs into their plans which
will allow them to achieve real-time-release (RTR) of
finished products while simultaneously improving the
efficiency of their quality operations. The company’s
initiative included three main elements: planning and
assessment, IT infrastructure, and process analytics.
Planning and Assessment
Within the context of this example, strategic operational
planning and assessment encompasses all non-technology,
project management aspects of the initiative. The project
Fig. 3 Cluster diagram illustrat-
ing the statistical correspon-
dence between total inventory
turn rate and gross margins
Fig. 4 Graphical illustration of the transformation of C/T before and
after deployment of PAT and lean flow path management (FPM). The
length of each segment corresponds to the relative portion of total C/T
consumed by the operation or activity (as denoted by the cross-
referenced number in the inset table). Implementing PAT and FPM
enable compression or elimination of non-value-added (NVA) activ-
ities, thereby increasing the value-added ratio (VAR), or the portion of
total C/T devoted to value-added processes
44 J Pharm Innov (2007) 2:38–50
begins with organization of an internal team to champion
the initiative, establishment of preliminary goals, and
initiation of communication between the project team,
company management, and the FDA. Assessments are
ongoing throughout the performance initiative, and are used
to prioritize individual projects to improve efficiency,
reform NVA activities, and mitigate risks to quality. Project
design and management should follow a logical structure,
such as the DMAIC (Define, Measure, Analyze, Improve,
Control) cycle [9]; it is important that the plan incorporates
sufficient flexibility to address the needs specific to the
company. Besides prioritizing investment activities, the
planning and assessment functions within the initiative
can have a dramatic impact on performance by facilitating
redesign of operating procedures (e.g. cleaning and change-
over protocols).
Process Analytics
With regard to analytical instrumentation, many solid
dosage manufacturing operations will benefit from three
major types of installations. In many cases the first PAT
installation should be focused on developing more rapid,
comprehensive analytical capabilities for raw materials
analysis within the dispensary. Deployment of enhanced
solid-state characterization tools for raw materials identifi-
cation is a very low-risk way to introduce new technologies
while building a critical understanding of the true variabil-
ity of materials which may impact downstream processes.
Additionally, the sensor data collected within the dispensa-
ry may ultimately be important as a basis for efficient
calibration of downstream sensors [21].
The second phase of instrumentation involves developing
control systems for critical unit operations such as granula-
tion, blending, or coating. In some cases control models can
be built using existing process data (e.g. air temperature,
torque, etc.); in many cases, though, technologies such as
near-infrared spectroscopy [22] and in-line particle size
analyses will be necessary to develop effective process
controls. While not specifically considered a production
operation, significant efficiency gains may be realized by
introducing new analytical devices to compress downtime
operations, such as deployment of ion mobility spectros-
copy (IMS) to reduce delays for cleaning verification.
The third phase of instrumentation is related to rapid
quality analysis. Rapid (or real-time) quality analysis is
critical to the success of continuous improvement by
providing timely characterization of process capability at
higher levels of statistical confidence than traditional
release testing protocols. Finally, the integration of raw
material characterization, unit operation control data, and
finished product quality analyses will enable RTR of
finished products, which, in many cases, will be the
most important factor in achieving the financial benefits
of PAT.
Production and Supply-chain Management Systems
In addition to deployment of new sensor technologies, co-
deployment of operations management and scheduling
systems capable of capitalizing on the operational changes
related to PAT (e.g. RTR) have a synergistic effect in
accelerating production flow. For example, without updat-
ing process schedules, elimination of QC hold times
might simply result in new WIP accumulation queues.
Additionally, without deploying systems to measure the
impact of operational changes on process performance
(including WIP velocity and CoQ) the actual ROI from
individual projects may be understated or attributed to
other factors, thereby putting future investments in
advanced manufacturing technologies at risk. The value
of PAT projects is enhanced significantly by including
Fig. 5 Graphical illustration of
the approximate breakdown of
revenues before (a) and after (b)
deployment of PAT and flow
path management (FPM) for a
hypothetical pharmaceutical
manufacturing company. The
pre-PAT data (a) are based on
average data for generic and
mid-sized pharmaceutical com-
panies; breflects the expected
impact of operational perfor-
mance enhancements enabled
by PAT
J Pharm Innov (2007) 2:38–50 45
systems for Lean manufacturing, such as flow path
management [14].
Results
After assembling a project team and performing the first
process risk and efficiency assessments, the team agreed to
work toward three stretch goals: increase average Cpk from
0.7 to 1.0, double the rates of RM, WIP, and FG inventory
turnover, and reduce labor components of CoQ by 15%.
Additionally, the team agreed to implement technological
or procedural changes to address the highest-priority risks
to quality identified during operational assessments.
Process Capability Enhancement
For many solid dosage manufacturing operations, increasing
Cpk to 1.0, or process sigma to 4.3, is a stretch goal well
within feasibility. The financial benefit of process capability
enhancement was calculated by reducing each component of
COGS by 3.3%, which corresponds to the portion of
production recovered by reducing the rate of batch failure.
Based on the distribution of revenue components shown in
Fig. 5, process capability enhancement reduces COGS by
∼$8.2 million annually. The estimated savings do not include
components related to savings on disposal fees, opportunity
costs of failure, or potential for increased sales due to
recovered production volume, all of which would compound
the savings due to improved quality performance.
Flow Path Management
Acceleration of the process and inventory is achieved by
three factors: elimination or minimization of inspection hold
times, reduction of NVA activities, and optimization of
production scheduling. For this case study the improvement
goals included doubling the rate of inventory turnover,
which merely brings performance up to the median rate of
turnover observed for the non-pharma companies surveyed,
which is still well below the level of “world class”supply
chain efficiency. By achieving RM, WIP, and FG turn rates
of 16, 30, and 7.6, respectively, and assuming 20% inventory
carrying costs (ref KPI survey), more than $13 million
annual overhead savings would be realized. Additionally, the
50% reduction in total inventories carried would free up
more than $66 million in cash from working capital.
Reduced Cost of Quality
The opportunity to significantly reduce CoQ beyond the
direct effects of improving Cpk is a major financial benefit
of implementing PAT, yet is often not identified as a
strategic goal. This is not to suggest that traditional QC/QA
functions must be completely displaced by PAT sensors or
process models. Rather, the implementation of real-time
analytical methods enables parallel operation of important
QC activities to be maintained out of the C/T critical path,
thereby facilitating more efficient scheduling of off-line
quality operations. It is important to also consider indirect
effects on CoQ related to Cpk improvement such as the
reduction of laboratory and root-cause investigations which
often divert laboratory resources from more productive
activities, such as R & D support. Indeed, the quality unit
will have an important role in establishing and maintaining
PAT sensor systems. Such activities can be managed outside
of the C/T critical path, however, and in many cases can be
outsourced to external experts. For the fictitious company
described in this case study, 15% reduction of CoQ labor
related to routine inspection and failure investigations
would yield more than $6 million in annual savings.
Ultimately, as companies continue to gain internal expertise
in deploying improved analytical and control technologies
it is reasonable to expect that significantly greater reduc-
tions in CoQ will be realized.
Discussion
Discounted Value of Savings
Ignoring implementation costs, and assuming gains are first
realized in the third year of the initiative, the 10-year
discounted (WACC ∼12%) value of the savings described
in the preceding paragraphs is more than $175 million
(including $66 million in working capital savings). The
operational transformations responsible for the savings
described in Figs. 4and 5are the expected direct results
of three main factors:
&Process capability was improved by deployment of
PAT-enabled process controls and by implementing an
effective continuous improvement system
&Process and supply chain velocity increased following
elimination of QC/QA hold times, the achievement of
RTR, and by optimizing inventory levels through deploy-
ment of a real-time flow path management system.
&Implementation of PAT for real-time quality analysis
enabled reductions in labor and resources required for
operation of the quality unit.
In addition to the direct effects of these factors there are
many indirect savings which may add to the value of the
PAT + Lean initiative. For example, some companies may
be able to realize significantly greater CoQ savings by
46 J Pharm Innov (2007) 2:38–50
levering PAT installations to further reduce fixed and
variable costs of monitoring compliance. Furthermore,
studies in other process-based industries have shown that
overhead costs are more directly related to the number of
transactions than production volume [23]; hence, manufac-
turers with more complex flow paths (e.g. generic and
OTC) can expect outsized gains from improved turnover.
Other potential long-term gains from PAT + Lean include
more efficient process development and scale-up, technol-
ogy transfer, reduced regulatory burden, and increased
production flexibility.
The Synergy of PAT and Lean
Of the $27.7 million in total estimated annual cost savings,
only the $8.2 million related to Cpk enhancement would be
classified as being related only to PAT. The remaining 60%
of savings require aspects of both PAT and Lean manufac-
turing. For example, while real-time analytics and controls
obviates the need for QC holds, improved scheduling is
required to adjust to the faster pace of production. It is
difficult to accurately quantify all of the synergistic aspects
of co-deploying PAT and Lean using only a hypothetical
example. There is, however, a fundamental relationship
between process variability and C/T which demonstrates the
strategic value of deploying process controls to manage
variability. Hopp and Spearman [24] described the “cor-
rupting influence”of variability on process C/T; they
developed a mathematical relationship which describes
queuing time (CT
q
) as a function of the coefficients of
arrival and process time variability, c
a
and c
e
, capacity
utilization, u, the number stations, m, and the mean process
time of a job (i.e. unit operation), t
e
:
CTq¼c2
aþc2
e
2
uffiffiffiffiffiffiffiffiffiffiffi
2mþ1ðÞ
p1
m1uðÞ
!
teð2Þ
Hence, as described by Eq. 2, queuing time, the amount of
time which WIP must wait for equipment availability for
processing, increases with the square of both arrival- and
process-time variability. Furthermore, process variability
tends to cascade through operations; the variability in
process time for one operation becomes the arrival time
variability for the next, leading to additional scheduling
problems and delays. While process buffers can mitigate c
a
,
enabling higher capacity utilization, process acceleration
requires true mitigation of process variance.
Indeed, failing to mitigate variation before initiating a
lean manufacturing initiative will often result in counter-
productive results. It is for these reasons that the projects
described for this case study were planned such that process
capability and quality management were addressed first
through deployment of PAT, followed by active manage-
ment of production to minimize inventories.
Reality Check
Admittedly, some simplifying assumptions were made
within this example which may reduce the potential value
of savings in a real implementation scenario. Additionally,
the magnitude of investments required in terms of capital,
labor, and time related to actually implementing the changes
were not discussed. It is important to realize, however, that
the financial gains were generated not by bringing a
“benchmark average”pharmaceutical manufacturer up to
world class, but rather by bringing operational performance
up to nearly the median performance of process-based
manufacturers in other industries. Indeed, given the current
state of manufacturing performance in the pharmaceutical
industry, only relatively modest improvement in process
capability and supply chain velocity are needed to realize
outsized gains. Based on these relatively conservative
estimates only a fraction of the estimated gains would be
required to repay the cost of most any PAT system which
would be implemented on such a scale. Furthermore, most
branded and generic drug manufacturers are much larger and
have much greater throughput than the company used for the
case study, and therefore have greater potential for gains.
As noted earlier in this manuscript, the tremendous
opportunity for pharmaceutical manufacturers to reap
significant cost savings through deployment of PAT and
Lean has been described many times before. Despite these
data, the rate of investment in advanced processes and
controls among pharmaceutical manufacturers still lags far
behind most other process-based industries. For a compa-
ny to invest significant capital in new manufacturing
technologies the risk-adjusted return on investment
(RAROI) must be worthwhile relative to available alter-
natives (e.g.—additional staff, increased warehouse space,
R & D spending). Ordinarily this would imply that the
risk-to-return profile of PAT investment is less favorable
in these sectors.
Five years ago the real and perceived technological and
regulatory risks associated with implementing process
analytics for real-time monitoring and control of production
processes were relatively high. The regulatory reforms that
have been deployed in conjunction with FDA’sPAT
initiative [2,25] have significantly reduced the regulatory
risks associated with incorporating process analytics into
established operations. Furthermore, the published experi-
ences of early-adopters of PAT in the pharmaceutical
industry, academia, and technology suppliers, as well as
many years of collective experience in process analytical
chemistry (PAC) from other industries, have ironed out
J Pharm Innov (2007) 2:38–50 47
many of the technical issues involved with deploying PAT.
Most of the significant technological risks associated with
investments in PAT have been addressed. As the potential
for significant financial rewards become more apparent, and
as the regulatory and technical risks and uncertainty have
faded, one would have to wonder why more management
teams don’t choose to invest in PAT. As cited by Vernon,
Hughen and Trujillo [7]; Nelson and Winter [26] demon-
strated empirically that “it is not uncommon for industries
to evolve more slowly to capitalize on such inefficiencies
than orthodox economic theory would predict.”On the
other hand, though, it may be that early attempts at PAT
projects have left some companies feeling burned after
experiencing many new costs with few tangible benefits
stemming from insufficient planning and lack of internal
experience with PAT.
Summary
The purpose of this manuscript has been to describe in
detail some of the opportunities for financial returns on
investments in PAT. Investments in process analytical
sensors and controls generate such returns indirectly by
removing operational, technological, and regulatory road-
blocks which have historically prevented management
from achieving maximum process efficiency. A hypothet-
ical case study based on publicly-available benchmarks
was used to illustrate the magnitude of savings due to
PAT, quality management, and Lean manufacturing which
could be expected for a small manufacturer of pharma-
ceutical solid dosage forms. By achieving conservative
levels of operational performance improvement in terms
of process capability enhancement, flow path manage-
ment, and reduced cost of quality, it was shown that a
company having revenues of $450 million could reason-
ably achieve nearly $27.7 million in annual cost savings,
and could return more than $66 million in cash from
working capital. If it can be assumed the cost savings are
not invested elsewhere in the operation, the company
would increase operating margins by more than 600 basis
points from 10% up to 16%, ultimately having a dramatic
effect on both the magnitude of earnings and the
predictability of operating costs.
Appendix
On the Calculation of Process Sigma
Due to its widespread success, and the proliferation of
consultants it spawned, competing approaches to Six Sigma
have evolved from the original process developed by
Motorola, leading to considerable confusion about the true
intent and spirit of Six Sigma [27]. Consequently, the true
method for calculating the sigma level for a process, and its
relationship with Cpk, is often misunderstood. According to
Keki Bhote, who helped develop the Six Sigma process
during his 42-year career at Motorola [27], the sigma level
of a process corresponds directly to the sigma used in the
denominator for calculation of Cpk (for a centered process)
[8]. Thus, based on Eq. 1, the sigma level for a centered
process will be equal to three times its Cpk score. In other
words, the distance between the control limits for a process
centered at zeros will be ±6σ.
It is widely acknowledged that a six sigma (6σ) process
will produce approximately 3.4 defects per million oppor-
tunities (DPMO). By applying simple statistical calcula-
tions to Eq. 1, however, it can easily be shown that,
according to the original definition of process sigma, a 6σ
process is expected to produce approximately 0.02 DPMO,
or 2.0 defective parts-per-billion (PPB). The startling
disparity between the figures is the result of major differ-
ences in how they are calculated, and a “fudge-factor”
known as the “1.5σshift,”or the “shift and drift”offset.
Calculation of the “shifted”sigma is achieved by
evaluation of the inverse standard normal cumulative
distribution, or probit, function for a ∼N(0,1) process at a
given %RFT [28]. There is, unfortunately, no direct
equation for calculation of the probit function; hence,
practitioners often rely on web-based calculators, third-
party software products, or manual interpolation from tables
Fig. 6 Graphical illustration of the correspondence between Cpk,
defect concentration, Process Sigma, and the “shifted”process sigma.
Most benchmarks indicate that pharmaceutical manufacturers operate
within the shaded region near Cpk ∼0.7
48 J Pharm Innov (2007) 2:38–50
relating process sigma values to DPMO or %RFT to
estimate the sigma level of their processes. The shifted
process sigma estimate can be calculated using a modified
version of an Excel
®
spreadsheet formula described by
Pyzdek [28]: ‘=NORMSINV(X)+ 1.5′, where Xis %RFT.
While there are numerous arguments attempting to
support the 1.5σdrift as a correction for such things as
long-term “natural drift”of processes, or as an adjustment
for differences between customer perception and internal
control limits, a brief review of the available Six Sigma
literature did not yield a truly rigorous statistical reason for
the more-complicated alternative calculation. Given the
historical origins of the probit function, one might speculate
that perhaps the shifted sigma is more linear as an input for
regression studies of process capability. An examination of
the relationship between Cpk, defect rate, process sigma,
and the shifted sigma, as shown in Fig. 6and Table 2,
however, would suggest that the alternative calculation is
an easy way of increasing the apparent performance of
operations and the 1.5σshift is simply the offset required to
prevent difficult-to-explain negative sigma values.
References
1. Pharmaceutical CGMPs for the 21st century—a risk-based
approach, final report. U.S. Department of Health and Human
Services, Food and Drug Administration; September 2004.
2. Guidance for industry: PAT—a framework for innovative phar-
maceutical development, manufacturing, and quality assurance.
U.S. Department of Health and Human Services, Food and Drug
Administration; September 2004.
3. Arlington S, Barnett S, Hughes S, Palo J, Shu E. Pharma 2010:
the threshold of innovation. IBM Business Consulting Services.
NY: Somers; November 2002.
4. Arlington S, Browne H, Buhle L, Byrne F, Curle P, Drayton S,
et al. The metamorphosis of manufacturing. IBM Business
Consulting Services. NY: Somers; May 2005.
5. du Pre Gauntt J. Quality manufacturing: a blockbuster opportu-
nity for pharmaceuticals. White Paper, The Economist Intelli-
gence Unit. www.eiu.com; September 2005.
6. Abboud L, Hensley S. Factory shift: new prescription for drug
makers: update the plants. In: Wall Street J. New York, NY; 2003. p. 1.
7. Vernon JA, Hughen WK, Trujillo AJ. Pharmaceutical manufac-
turing efficiency, drug prices, and public health: examining the
casual links. Drug Inf J 2007;41:229–239.
8. Bhote KR. What is Six Sigma? In: The ultimate Six Sigma. Vol 1.
New York: AMACOM; 2002. p. 9–14.
9. Marash SA, Berman P, Flynn M. Fusion management: harnessing
the power of Six Sigma, Lean, ISO 9001:2000, Malcom
Baldridge, TQM and other quality breakthroughs of the past
century. Fairfax, VA: QSU Pub.; 2004.
10. Malchi G, McGurk H. Increasing value through the measurement
of the cost of quality (CoQ) - a practical approach. Pharm Eng
2001;21:92–95.
11. Schiffauerova A, Thomson V. A review of research on cost of
quality models and best practices. Int J Qual Reliab Manage 2006;
23:647–669.
12. Williams ART, van der Wiele A, Dale BG. Quality costing: a
management review. Int J Manag Rev 1999;1:441–460.
13. Shannon P. The value-added ratio. Qual Prog 1997;30:94–97.
14. Gerecke G, Knight TP. Improving performance and reducing
cycle time using flow path management: a case study. Pharm Eng
2001;21.
15. Lewis NA. A tracking tool for lean solid-dose manufacturing.
Pharm Technol 2006;30:94–108.
16. Benson RS, McCabe JDJ. From good manufacturing practice to
good manufacturing performance. Pharm Eng 2004;24:1–6.
17. Macher J, Nickerson J. Pharmaceutical manufacturing research
project: final benchmarking report. Georgetown University,
McDonough School of Business & Washington University, Olin
School of Business; September 2006.
18. Logistics and Supply Chain Management (SCM) Key Performance
Indicators (KPI) analysis: a Canada/United States pharmaceutical
sector supply chain perspective. Industry Canada and Supply Chain
& Logistics Canada; http://strategis.ic.gc.ca/epic/internet/indsib-
logi.nsf/en/h_pj00282e.html,2006.
19. Damodaran A. Cost of capital by sector. Vol 2007. New York,
NY; 2007.
Table 2 Quantitative relationship between Cpk, % defective units,
PPM, the true Process Sigma, and the “Shifted”Sigma
Cpk % PPM Process
sigma
“Shifted”
sigma
Sigma
error
0.01 98 976,067 0.03 −0.48 −0.51
0.05 88 880,765 0.15 0.32 0.17
0.50 13.4 133,614 1.50 2.61 1.11
0.55 9.89 98,943 1.65 2.79 1.14
0.60 7.19 71,861 1.80 2.96 1.16
0.65 5.12 51,176 1.95 3.13 1.18
0.70 3.57 35,729 2.10 3.30 1.20
0.75 2.44 24,449 2.25 3.47 1.22
0.80 1.64 16,395 2.40 3.63 1.23
0.85 1.08 10,772 2.55 3.80 1.25
0.90 0.693 6,934 2.70 3.96 1.26
0.95 0.437 4,372 2.85 4.12 1.27
1.00 0.270 2,700 3.00 4.28 1.28
1.05 0.163 1,633 3.15 4.44 1.29
1.10 0.0967 967 3.30 4.60 1.30
1.15 0.0561 561 3.45 4.76 1.31
1.20 0.0318 318 3.60 4.92 1.32
1.25 0.0177 177 3.75 5.07 1.32
1.30 0.00962 96 3.90 5.23 1.33
1.35 0.00512 51 4.05 5.38 1.33
1.40 0.00267 27 4.20 5.54 1.34
1.45 0.00136 14 4.35 5.70 1.35
1.50 0.000680 6.8 4.50 5.85 1.35
1.5484 0.000340 3.4 4.645 6.000 1.35
1.55 0.000332 3.3 4.650 6.005 1.35
1.60 0.000159 1.6 4.80 6.16 1.36
1.65 0.0000742 0.7 4.95 6.31 1.36
1.70 0.0000340 0.34 5.10 6.47 1.37
1.75 0.0000152 0.15 5.25 6.62 1.37
1.80 0.0000067 0.067 5.40 6.77 1.37
1.85 0.0000029 0.029 5.55 6.93 1.38
1.90 0.0000012 0.012 5.70 7.08 1.38
1.95 0.0000005 0.005 5.85 7.23 1.38
2.00 0.0000002 0.002 6.00 7.39 1.39
J Pharm Innov (2007) 2:38–50 49
20. Stock JR, Lambert DM. Strategic logistics management. 2nd ed.
Homewood, IL: Irwin; 1987.
21. Cogdill RP, Anderson CA. Efficient spectroscopic calibration
using net analyte signal and pure-component projection methods.
J Near Infrared Spectrosc 2005;13:119–131.
22. Cogdill RP, Drennen JK. Near-infrared spectroscopy. In: Brittain
HG, editor. Spectroscopy of pharmaceutical solids. New York,
NY: Taylor and Francis Group; 2006. p. 313–412.
23. Banker RD, Potter G, Schroeder RG. An empirical analysis of
manufacturing overhead cost drivers. J Account Econ 1995;
19:115–137.
24. Hopp WJ, Spearman ML. Factory physics. Foundations of
manufacturing management. Chicago: Irwin; 1996.
25. Guidance for industry: comparability protocols—chemistry,
manufacturing, and controls information, draft guidance. U.S.
Department of Health and Human Services, Food and Drug
Administration; February 2003.
26. Nelson R, Winter S. An evolutionary theory of economic change.
Cambridge, MA: Harvard University Press; 1982.
27. Galvin RW. Forward. In: The ultimate Six Sigma. Vol 1. New York:
AMACOM; 2002. p. xxi–xxii.
28. Pyzdek T. The 1.5 Sigma shift. In: Quality digest. May 2001.
50 J Pharm Innov (2007) 2:38–50
