Improving Biometrics Processes with Six-Sigma

highpitchedteamSecurity

Nov 30, 2013 (3 years and 8 months ago)

72 views

PhUSE 2009


1

Paper
TS02



Improving Biometrics Processes with Six
-
Sigma


Andrew York, Covance
,
Crawley
,
UK


ABSTRACT

Since 2005, Covance Clinical Development Services has deployed Six Sigma as part of its process improvement
culture. The aim being to measure, improve a
nd monitor process efficiency leading to an increase in quality and a
reduction in the cost and time associated with running a clinical trial. In 2008, Biometrics completed its first Six
Sigma green belt project aimed at reducing the work time for writing
medical writing narratives, achieved through a
SAS
®

based programmatic approach. In 2009 we initiated further projects aimed, for example, at reducing the time it
takes to write a SAP or to reduce the work required to QC Tables, Figures and Listings. This
paper examines impact
of Six Sigma within Biometrics, the time commitment, the challenges faced. It will then present an overview of some
of our completed and ongoing projects and finally will look to the future and the lessons learnt along the way.

INTRO
DUCTION

Six
-
Sigma was first introduced into Covance in 2002 within our Early Development business. The success of this
programme led to its introduction into our Late Stage (Clinical Development Services) business in 2005. Although
Six
-
Sigma came into be
ing within a manufacturing environment, it was felt that there was sufficient potential as a
process development tool to adopt it for something other that the manufacture of “widgets”.


Adopting Six
-
Sigma was a complete culture change for Covance, processe
s needed to be
analyzed

and evaluated
statistically using applicable data and robust statistical method. Six
-
Sigma introduced an entire new language into
the company, as well as to introduce statistical methods to many staff who were not qualified statist
icians.


Many departments volunteered their best staff

to become black belts, committing

them to a minimum of two years as
a Black Belt during which they had to complete many weeks of training, three Black Belt projects and also to pass an
exam before they

were fully accredited as a Black Belt. These staff were all given a guarantee of a position with the
company at then end of their ‘stint’ as black belts and indeed many have gone onto critical roles with the company.


Green Belts likewise had to undergo
three weeks of training, complete two Green Belt projects and pass an open
book exam before they were accredited as Green Belts. Additional project team members, Process Owners and
Project Champions complete this list again with a significant training inv
estment in these staff.


Figure 1 (overleaf) demonstrates how a sig
-
sigma process can lead to fewer process failures (or defects) even when
compared to a marginally less robust process.


Note
: This paper will mostly focus on the DMAIC (Define, Model, Analy
ze, Improve, Control) model for Six
-
Sigma.

SIX
-
SIGMA METHODOLOGY

There are numerous references to Six
-
Sigma methodologies including LEAN and
DMAIC

models (the two used at
Covance), and some of these references can be found at the end of this paper for furt
her reading. As stated, Six
-
Sigma requires some knowledge of statistical inference and analysis and depending upon the type of data analyzed
can require fairly complex analysis. At Covance it is the role of the Black Belts under the guidance of the Mast
er
Black Belt to guide the Green Belts and Project Team Members through the analysis steps.


One way to describe Six
-
Sigma is that it is a measurable process which compares the Voice of the Process (VOP) to
the Voice of the Customer (VOC). Process improve
ments occur to (a) achieve the desired quality outcome and (b)
reduce variability in the VOP until it is as least as good as the VOC.


Figure 2 (overleaf) graphically illustrates and compares the Voice of the VOP versus the VOC. Although not shown
in th
is case, ideally the VOP should be contained within the VOC.


PhUSE 2009


2

Figure 1


The difference between processes that are 99% effective vs 99.99966%


99% Good

(3.8 Sigma)


99.99966% Good

(6 Sigma)


20,000 lost articles of mail per hour



Seven articles lost per
hour


Unsafe drinking water for almost 15
minutes each day


One unsafe minute every seven
months


5,000 incorrect surgical operations per
week


1.7 incorrect operations per week


Two short or long landings at most
major airports each day


One short o
r long landing every five
years


200,000 wrong drug prescriptions
each year


68 wrong prescriptions per year


No electricity for almost seven hours
each month


One hour without electricity every 34
years







Figure 2



Comparing the Voice of the Pro
cess vs Voice of the Customer



Defects

Voice of the Process

Inadequate
Design

Inadequate
Process
Capability

Unstable Parts &
Materials

Defects

Acceptable

LS
L

USL

Voice of the Customer

PhUSE 2009


3

D
EFINE,
M
EASURE

AND
A
NALYSIS

PHASES

The Six
-
Sigma Charter describes the aims of the project in ways that can be measured statistically (figure 3).
Baseline data hopefully already exists but if no
t this must be collected such that the current state of the process (or
VOP) can be measured and analyzed

statistically (figure 4).


Figure 3 : Project Charter for the
Medical Writing
Narratives Project


Project Objective:

Y: Reduce work time to header com
pletion for Draft 1 narrative by 50%


y1: Decrease work time variability of header completion for Draft 1 narrative by 50%.


Project Metrics:

Y: time to data entry of standard patient information in Draft 1 narrative header


y1: work time variability of h
eader completion for Draft 1 narrative


Defect Definition:

Y: time to enter data for header exceeding 8 min/header


y1: work time variability of header completion for Draft 1 narrative std dev. of >5 minutes




Figure 4 : Baseline Analysis of Work Time fo
r Narratives Header Completion




In addition to the baseline analyses, six
-
sigma tools such as
the SIPOC (Suppliers, Inputs, Process, Outputs,
Customers),
Process M
aps, C
ause and Effect (C&E)
M
atrix
, Fishbone D
iagrams are emplo
yed to determine and
evaluate cause of variability within the data. I
n statistical parlance these w
ould be your
independent variables (e.g.
the
x’s in when expressed in
a
regression
format:

y

=

0

+

1
x
1

+

2
x
2+ ... +

n
x
n
).


In the case of the narrative
s project, factors such as European vs US or Internal vs External medical writers were
amongst those considered when modeling the variables that contributed most towards the writing of medical writing
headers.

9
.
0
7
.
5
6
.
0
4
.
5
3
.
0
1
.
5
0
.
0
t
r
a
n
s
f
o
r
m
e
d

d
a
t
a
L
B
*
U
S
L
*
U
S
L
*
2
.
8
2
8
4
3
S
a
m
p
l
e

M
e
a
n
*
3
.
2
0
7
7
4
S
t
D
e
v
(
W
i
t
h
i
n
)
*
0
.
8
5
3
0
9
S
t
D
e
v
(
O
v
e
r
a
l
l
)
*
1
.
0
7
3
0
2
L
B
0
T
a
r
g
e
t
*
U
S
L
8
S
a
m
p
l
e

M
e
a
n
1
1
.
4
3
6
2
S
a
m
p
l
e

N
2
4
3
S
t
D
e
v
(
W
i
t
h
i
n
)
6
.
4
4
7
4
5
S
t
D
e
v
(
O
v
e
r
a
l
l
)
9
.
3
7
2
8
5
L
B
*
0
T
a
r
g
e
t
*
*
A
f
t
e
r

T
r
a
n
s
f
o
r
m
a
t
i
o
n
P
r
o
c
e
s
s

D
a
t
a
Z
.
B
e
n
c
h
-
0
.
4
4
Z
.
L
B
*
Z
.
U
S
L
-
0
.
4
4
C
p
k
-
0
.
1
5
Z
.
B
e
n
c
h
-
0
.
3
5
Z
.
L
B
*
Z
.
U
S
L
-
0
.
3
5
P
p
k
-
0
.
1
2
C
p
m
*
O
v
e
r
a
l
l

C
a
p
a
b
i
l
i
t
y
P
o
t
e
n
t
i
a
l

(
W
i
t
h
i
n
)

C
a
p
a
b
i
l
i
t
y
P
P
M

<

L
B
0
.
0
0
P
P
M

>

U
S
L
6
3
7
8
6
0
.
0
8
P
P
M

T
o
t
a
l
6
3
7
8
6
0
.
0
8
O
b
s
e
r
v
e
d

P
e
r
f
o
r
m
a
n
c
e
P
P
M

<

L
B
*
*
P
P
M

>

U
S
L
*
6
7
1
7
0
7
.
0
5
P
P
M

T
o
t
a
l
6
7
1
7
0
7
.
0
5
E
x
p
.

W
i
t
h
i
n

P
e
r
f
o
r
m
a
n
c
e
P
P
M

<

L
B
*
*
P
P
M

>

U
S
L
*
6
3
8
1
4
2
.
8
5
P
P
M

T
o
t
a
l
6
3
8
1
4
2
.
8
5
E
x
p
.

O
v
e
r
a
l
l

P
e
r
f
o
r
m
a
n
c
e
W
i
t
h
i
n
O
v
e
r
a
l
l
B
a
s
e
l
i
n
e

P
r
o
c
e
s
s

C
a
p
a
b
i
l
i
t
y
U
s
i
n
g

B
o
x
-
C
o
x

T
r
a
n
s
f
o
r
m
a
t
i
o
n

W
i
t
h

L
a
m
b
d
a

=

0
.
5
PhUSE 2009


4

I
MPROVE
AND
C
ONTROL
PHASE
S

Once the core proce
ss has been analyzed,
the project then

moves into its Improve Phase. The project team should
by now have some ideas for process improvements, helped by the statistical analysis of the independent variables
as part of the baseline analysis. Part of this p
rocess is to involve the Project Champion to assist with the ranking of
those factors which are most considered to be under the influence of the business to change or control (termed
High
-
Level and Mid
-
Level C&E).


In the case of the Narratives project, a
SAS
®

based solution was determined to be the optimum solution and was
subsequently put in place to replace the manual process of header creation. Further data was collected to assess
the performance of the new process in comparison to the baseline process
. The data was compared statistically to
assess if the process had been improved and in fact if the goals of the charter had been
achieved (figures 5 and 6

-

overleaf
).


Once the process improvement has been established, the project enters its long
-
term c
ontrol phase. Responsibility
for tracking the control metrics falls to the nominated Process Owner who will monitor the process until either it falls
out of control or else further improvements are conducted. Such improvements could be as the result of a

new Six
-
Sigma project or a “just do it” (i.e. no
-
brainer) project.



Figure 5 : Assessing the Process Improvement





OUTCOME

For the narratives project, both of the charter objectives were met and exceeded and the project was
hailed as a
success by the senior management at Covance. In addition the project is now cited as part of the Six
-
Sigma training
at Covance. Financially the project was also a success leading to savings for our Medical Writing department that
exceeded th
e investment due to increased programming time. Finally and importantly our medical writers could
spend more time focusing on the medical writing aspects of the narratives instead of manually entering header data
into the individual narratives.


Figure 7
(
overleaf
) illustrates the control chart for the Narratives project, the process improvement can clearly be
seen.

9
.
0
7
.
5
6
.
0
4
.
5
3
.
0
1
.
5
0
.
0
t
r
a
n
s
f
o
r
m
e
d

d
a
t
a
L
B
*
U
S
L
*
U
S
L
*
2
.
8
2
8
4
3
S
a
m
p
l
e

M
e
a
n
*
3
.
2
0
7
7
4
S
t
D
e
v
(
W
i
t
h
i
n
)
*
0
.
8
5
3
0
9
S
t
D
e
v
(
O
v
e
r
a
l
l
)
*
1
.
0
7
3
0
2
L
B
0
T
a
r
g
e
t
*
U
S
L
8
S
a
m
p
l
e

M
e
a
n
1
1
.
4
3
6
2
S
a
m
p
l
e

N
2
4
3
S
t
D
e
v
(
W
i
t
h
i
n
)
6
.
4
4
7
4
5
S
t
D
e
v
(
O
v
e
r
a
l
l
)
9
.
3
7
2
8
5
L
B
*
0
T
a
r
g
e
t
*
*
A
f
t
e
r

T
r
a
n
s
f
o
r
m
a
t
i
o
n
P
r
o
c
e
s
s

D
a
t
a
Z
.
B
e
n
c
h
-
0
.
4
4
Z
.
L
B
*
Z
.
U
S
L
-
0
.
4
4
C
p
k
-
0
.
1
5
Z
.
B
e
n
c
h
-
0
.
3
5
Z
.
L
B
*
Z
.
U
S
L
-
0
.
3
5
P
p
k
-
0
.
1
2
C
p
m
*
O
v
e
r
a
l
l

C
a
p
a
b
i
l
i
t
y
P
o
t
e
n
t
i
a
l

(
W
i
t
h
i
n
)

C
a
p
a
b
i
l
i
t
y
P
P
M

<

L
B
0
.
0
0
P
P
M

>

U
S
L
6
3
7
8
6
0
.
0
8
P
P
M

T
o
t
a
l
6
3
7
8
6
0
.
0
8
O
b
s
e
r
v
e
d

P
e
r
f
o
r
m
a
n
c
e
P
P
M

<

L
B
*
*
P
P
M

>

U
S
L
*
6
7
1
7
0
7
.
0
5
P
P
M

T
o
t
a
l
6
7
1
7
0
7
.
0
5
E
x
p
.

W
i
t
h
i
n

P
e
r
f
o
r
m
a
n
c
e
P
P
M

<

L
B
*
*
P
P
M

>

U
S
L
*
6
3
8
1
4
2
.
8
5
P
P
M

T
o
t
a
l
6
3
8
1
4
2
.
8
5
E
x
p
.

O
v
e
r
a
l
l

P
e
r
f
o
r
m
a
n
c
e
W
i
t
h
i
n
O
v
e
r
a
l
l
B
a
s
e
l
i
n
e

P
r
o
c
e
s
s

C
a
p
a
b
i
l
i
t
y
U
s
i
n
g

B
o
x
-
C
o
x

T
r
a
n
s
f
o
r
m
a
t
i
o
n

W
i
t
h

L
a
m
b
d
a

=

0
.
5
2
.
8
2
.
4
2
.
0
1
.
6
1
.
2
0
.
8
0
.
4
0
.
0
t
r
a
n
s
f
o
r
m
e
d

d
a
t
a
L
B
*
U
S
L
*
U
S
L
*
2
.
8
2
8
4
3
S
a
m
p
l
e

M
e
a
n
*
1
.
0
3
4
4
4
S
t
D
e
v
(
W
i
t
h
i
n
)
*
0
.
2
6
4
2
4
6
S
t
D
e
v
(
O
v
e
r
a
l
l
)
*
0
.
3
4
3
9
3
6
L
B
0
T
a
r
g
e
t
*
U
S
L
8
S
a
m
p
l
e

M
e
a
n
1
.
1
8
7
8
2
S
a
m
p
l
e

N
2
1
9
S
t
D
e
v
(
W
i
t
h
i
n
)
0
.
5
9
1
9
6
5
S
t
D
e
v
(
O
v
e
r
a
l
l
)
0
.
8
1
9
6
9
8
L
B
*
0
T
a
r
g
e
t
*
*
A
f
t
e
r

T
r
a
n
s
f
o
r
m
a
t
i
o
n
P
r
o
c
e
s
s

D
a
t
a
Z
.
B
e
n
c
h
6
.
7
9
Z
.
L
B
*
Z
.
U
S
L
6
.
7
9
C
p
k
2
.
2
6
Z
.
B
e
n
c
h
5
.
2
2
Z
.
L
B
*
Z
.
U
S
L
5
.
2
2
P
p
k
1
.
7
4
C
p
m
*
O
v
e
r
a
l
l

C
a
p
a
b
i
l
i
t
y
P
o
t
e
n
t
i
a
l

(
W
i
t
h
i
n
)

C
a
p
a
b
i
l
i
t
y
P
P
M

<

L
B
0
.
0
0
P
P
M

>

U
S
L
0
.
0
0
P
P
M

T
o
t
a
l
0
.
0
0
O
b
s
e
r
v
e
d

P
e
r
f
o
r
m
a
n
c
e
P
P
M

<

L
B
*
*
P
P
M

>

U
S
L
*
0
.
0
0
P
P
M

T
o
t
a
l
0
.
0
0
E
x
p
.

W
i
t
h
i
n

P
e
r
f
o
r
m
a
n
c
e
P
P
M

<

L
B
*
*
P
P
M

>

U
S
L
*
0
.
0
9
P
P
M

T
o
t
a
l
0
.
0
9
E
x
p
.

O
v
e
r
a
l
l

P
e
r
f
o
r
m
a
n
c
e
W
i
t
h
i
n
O
v
e
r
a
l
l
S
i
x

S
i
g
m
a

I
m
p
r
o
v
e
d

P
r
o
c
e
s
s

C
a
p
a
b
i
l
i
t
y
U
s
i
n
g

B
o
x
-
C
o
x

T
r
a
n
s
f
o
r
m
a
t
i
o
n

W
i
t
h

L
a
m
b
d
a

=

0
.
5
Change in Z score

Baseline:
-
0.35
σ

Improved: 5.22
σ

PhUSE 2009


5

Figure 6 : Analyzing the Improvement






Figure 7 : Control Chart for the Narrative Project






























I
m
p
r
o
v
e
d
B
a
s
e
l
i
n
e
8
0
7
0
6
0
5
0
4
0
3
0
2
0
1
0
0
D
a
t
a
T
i
m
e

t
o

W
r
i
t
e

H
e
a
d
e
r
s

f
o
r

N
a
r
r
a
t
i
v
e
s

(
m
i
n
)
S
i
x

S
i
g
m
a

P
r
o
c
e
s
s

I
m
p
r
o
v
e
m
e
n
t
s

i
n

N
a
r
r
a
t
i
v
e

W
r
i
t
i
n
g
Significant Improvement p<0.001

11.4 min per

narrative

1.2 min per

narrative

4
6
4
1
3
6
3
1
2
6
2
1
1
6
1
1
6
1
4
3
2
1
S
a
m
p
l
e
S
a
m
p
l
e

M
e
a
n
_
_
X
=
1
.
0
3
4
U
C
L
=
1
.
3
3
8
L
C
L
=
0
.
7
3
1
B
a
s
e
l
i
n
e
I
m
p
r
o
v
e
d
4
6
4
1
3
6
3
1
2
6
2
1
1
6
1
1
6
1
1
.
5
1
.
0
0
.
5
0
.
0
S
a
m
p
l
e
S
a
m
p
l
e

S
t
D
e
v
_
S
=
0
.
2
9
4
U
C
L
=
0
.
5
1
7
L
C
L
=
0
.
0
7
0
B
a
s
e
l
i
n
e
I
m
p
r
o
v
e
d
1
6
1
1
X
b
a
r
-
S

C
h
a
r
t

o
f

H
e
a
d
e
r
1

b
y

P
r
o
c
e
s
s
U
s
i
n
g

B
o
x
-
C
o
x

T
r
a
n
s
f
o
r
m
a
t
i
o
n

W
i
t
h

L
a
m
b
d
a

=

0
.
5
0
T
e
s
t
s

p
e
r
f
o
r
m
e
d

w
i
t
h

u
n
e
q
u
a
l

s
a
m
p
l
e

s
i
z
e
s
PhUSE 2009


6



FURTHER PROJECTS

The above is an example of just one project; Covance is currently running or has completed dozens of such projects.
Within Biometrics examples of new ongoing projects includes:




Reducing work time on SAP production



Reducing th
e work time for QC of Tables, Figures and Listings



Reducing the work effort for data transfers


Additional cross
-
functional projects are also being considered, including collaborations with Data Management and
Medical Writing. In total we have contributed

to this effort with one Black
-
Belt, five Green Belts and perhaps a dozen
Project Team members.

CONCLUSION

Six
-
Sigma represents a valuable tool for driving and measuring process improvements within Covance Biometrics
and across functions. Covance has prio
ritized
Six
-
S
igma as
a major engine

for change and as even in
this years’
harsh economic climate has prioritized the Six
-
Sigma efforts around the globe.


Challenges and roadblocks to Six
-
Sigma of course exist and these include a
lignment
of resources, resis
tance to
change, having to retrospectively collect historical data from
non
-
optimal
systems. Of course most Green Belts are
not
trained
statisticians and so
Black B
elt support is

a
crucial
success factor.


However a natural result of implementing Six
-
Sigm
a is an inherent ability to define and collect meaningful
performance metrics for those processes utilizing a six
-
sigma control plan. Needless to say both Covance and our
Clients will all benefit from improved processes in the long run.

REFERENCES




The Po
wer of Six Sigma

-

Subir Chowdhury (Dearborn Trade Publishing) 117pgs, $17.95: A good
introduction that uses a fictionalized tale of 2 individuals to explain how Six Sigma works.



What Is Six Sigma

-

Pete Pande and Larry Holpp (McGraw
-
Hill) 86pgs, $12.00: E
asy read, provides basics



Six Sigma For Dummies

-

Craig Gygi, Neil DeCarlo and Bruce Williams (Wiley) 318 pgs, $21.99: Contrary
to the title, the book is a thorough attempt to provide more detail on the DMAIC process, some good
examples, links methodology
to tools.



Six Sigma and Minitab

-

Quentin Brook (
www.qsbc.co.uk
) 180pgs, ~$35.00: A more in depth link between
DMAIC and Minitab, (similar to th
e Minitab cheat sheet idea), with good flow diagrams to point you to the
right tools.



Lean Six Sigma Pocket Toolbook



Michael L. George, David Rowlands, Mark Price, John Maxley,
282pgs, McGraw
-
Hill ($15)


Excellent in
-
depth guide to the tools and methodo
logy.

ACKNOWLEDGMENTS

With Special thanks to Grace Lee, Green Belt and to Jill Johnston, Black Bel
t on the N
arratives Project.

CONTACT INFORMATION

Your comments and questions are valued and encouraged. Contact the author at:


Andrew York

Covance CAPS Limi
ted

Manor Royal

Crawley

RH10 9PY

Email:

Andrew.york@covance.com

Web:

www.covance.com


Brand and product names are trademarks of their respective companies.