Click here to read this Lean in Supply Chain presentation.ppt

Turning Science into Caring

Supply Chain: Adding Customer Value

Anthony Orzechowski Director, R&D Quality Engineering

09-22511 v1.0

Presentation Supply Chain CapabilityDate: June 25, 2009

2 Company Confidential© 2009 Abbott

ConclusionWe can improve our customers’ ability to be competitive by providing means to translate quality measures into business impact

As suppliers, we should always want to help build our customers’ competitive advantage.

We can help them do this by …

– Relating the capability of key process outputs to safety and the financial impact on their business.

– Helping drive expanded business for the customer will in turn lead to expanded business for ourselves.

• This presentation illustrate one example of this translation by looking at a creative and helpful view of process capability.



50o 3’N

8o 21’E

Putting the Focus on the Customer

Family of integrated systems

High first-pass acceptance

Expanding assay menu

Pre- and post-analytics

Reliability

Scientific Leadership

Standardization

Consolidation

Workflow optimization

Results

Safety

Labor efficiencies

Cost pressures

Error reduction

Complexity



ARCHITECT “Family Commonality…”

Universal ARCHITECT

Sample CarrierRSH

Identical Software

Common Immunoassay

Reagents

i1000SR

Immunoassay

Immunoassay

i2000SR Immunoassay

c8000Clinical Chemistry

Common Clinical Chemistry Reagents

c16000Clinical Chemistry

Chemistry

Integration

ci8200Integrated

ci16200Integrated



Assessing Performance in a Clinical LaboratoryThe Essential Question …

“What amount of medical harm due to analytical error is it OK to let go undetected?”

Dr. Frederick A. Smith Children’s Hospital - Chicago

Dr. Frederick A. Smith - Children’s Hospital - Chicago



Measuring CapabilityMeasuring a Suppliers ability to Meet Customer Needs is a well accepted practice

• As suppliers we want to deliver solutions that will consistently be capable meeting the need posed by this essential question.

• This requires us to understand several key parameters …

– The Quality Specification (Total Error Allowable (TEa)

– Our expected Deviation from Target

– Our expected Variability about the process mean

• One traditional means to assess these together is a Capability index such as Cpk or Ppk.



Process Centered Capability Index — CpkCpk accounts for both inherent variation and a shift in mean.

kLSLUSLkCC ppk

16

1

LSL USL

Cp > 1.5

Cp = 1.0

Cp < 0.75

Target

USLLSL y

Cp is about 1.0, butCpk is about 0.5.



Typical Translation of Capability

Sigma Level = Cpk * 3

– But what does this mean to the customer relative to business impact?

– Is 12 Sigma twice as valuable as 6 Sigma to the customer?

– How do we help optimize value for the customer ?

Sigma Level Cpk DPMO Performance1 Sigma 0.33 690,000 Poor2 Sigma 0.67 308,000 Poor3 Sigma 1.00 66,800 Marginal4 Sigma 1.33 6,210 Good5 Sigma 1.67 230 Excellent6 Sigma 2.00 3.4 World Class



Visualizing the Impact of Capability Customer Quality Control

Process with Capability = 1.5

USL = Upper Specification Limit (In-House)LSL = Lower Specification Limit

Test MethodUncertainty

True Variabilityof the Product

Beta RiskBeta Risk

LSLLSL USLUSL

Process with Capability = 0.8

USL = Upper Specification Limit (In-House)LSL = Lower Specification Limit

Test MethodUncertainty

True Variabilityof the Product

Beta RiskBeta Risk

LSLLSL USLUSL

Visualizing the Impact of Capability on False Rejection and Acceptance



Translating Capability to Business ImpactMeasurement or improvement of capability should be related to tangible business results

• Simply supplying a measure of capability however does not directly translate into a quantified financial or business outcome for the customer.

• To help in this translation from a QC standpoint,

– Relate it to the Essential Question … What level of error is OK to let go undetected?

– Understanding how this question is assured at our customer will help to understand the value they place on higher levels of capability.



Relationship of Sigma to Cpk

Sigma is simply a different view of Capability

The Sigma Level is directly analogous to measuring capability through Cpk.

Sigma = (TEa – Bias)/(SD)

The difference between Cpk and Sigma Level is:

– Sigma Level measures the deviation from target (truth) rather than the Spec Limit and,

– The value is normalized by dividing by the variability, not 3x the variability.

– Sigma Level = Cpk * 3

USL (TEa)LSL (TEa)

Sigma is about 1.5Cp is about 1.0 but,

Cpk is about 0.5

Bias

%CVDefectsT

rue

Val

ue

See:• Six Sigma Quality Design and Control, Second Edition and • Westgard.com



Application of Sigma Concepts and Metrics for QC Selection Application of Sigma Concepts and Metrics for QC Selection Higher Sigma Levels allow Larger Shifts to go undetected and still produce a safe result

-7s -6s -5s -4s -3s -2s -1s 0s 1s 2s 3s 4s 5s 6s 7s

Critical Shift Limit risk of a “bad”test result

to 5% (this can be tuned

to the situation)

1.65s

Risk of 5% means quality requirement cuts the error distribution at 1.65s Risk of 5% means quality requirement cuts the error distribution at 1.65s from mean. from mean.

Shifts greater than this must be detected. Shifts greater than this must be detected.

See: Six Sigma Quality Design and Control, Second Edition

The higher the Sigma Level The higher the Sigma Level Delivers higher Allowable Critical Shifts Delivers higher Allowable Critical Shifts



Relationship of Sigma to SEcritHigher Sigma Levels allow Greater Shifts to go undetected and still produce a safe result

Critical Systematic Error (SEcrit)

– Index used to describe size of error that needs to be detected by QC procedure

SEcrit = [(TEa – Bias)/SD] – 1.65

– Sigma = SEcrit + 1.65

– Can relate SE to rejection characteristics of QC rules and numbers of QC measurements using known power curves

SigmaLevel

USLLSL

Sigma is about 1.5Cp is about 1.0 but,

Cpk is about 0.5

Bias

%CVDefectsT

rue

Val

ue

See:• Six Sigma Quality Design and Control, Second Edition



Method Decision Chart - Evaluating the System Capability

Method Decision Chart

0.0%

4.0%

8.0%

12.0%

16.0%

20.0%

24.0%

0.0% 2.0% 4. 0% 6.0% 8.0% 10.0% 12.0%

All

ow

ab

le I

na

cc

ura

cy

In a method decision chart, allowable error is plotted with the systematic error component on the Y-Axis and the random error component on the X-Axis.

Allowable Imprecision (%CV)





0.0%

4.0%

8.0%

12.0%

16.0%

20.0%

24.0%

0.0% 2.0% 4. 0% 6.0% 8.0% 10.0% 12.0%


All

ow

ab

le I

na

cc

ura

cy

Example: Total Allowable Analytical Error (TEa) = 20%

If a particular analyte could allow up to 20% (at 95% Confidence the allowable random and systematic error could be plotted as shown below

Total Allowable Error bound at 95% Confidence (i.e.. 2

Sigma)




Sigma Value = (TEa – BIAS) / (SD)

Sigma Value = (TEa – BIAS) / (SD)


0.0%

4.0%

8.0%

12.0%

16.0%

20.0%

24.0%

0.0% 2.0% 4. 0% 6.0% 8.0% 10.0% 12.0%


All

ow

ab

le I

na

cc

ura

cy

Zone A Less than 2 Sigma

Zone B

Example: Total Allowable Analytical Error (TEa) = 20%Zone A < 2 SigmaB > 2 Sigma

Zone A < 2 SigmaB > 2 Sigma





0.0%

4.0%

8.0%

12.0%

16.0%

20.0%

24.0%

0.0% 2.0% 4. 0% 6.0% 8.0% 10.0% 12.0%


All

ow

ab

le I

na

cc

ura

cy

Zone I Less than 2 Sigma

II

V

Example: Total Allowable Analytical Error (TEa) = 20%

IIIIV

Zone I < 2 SigmaII 2 to 3 SigmaIII 3 to 4 SigmaIV 4 to 5 SigmaV > 5 Sigma

Zone I < 2 SigmaII 2 to 3 SigmaIII 3 to 4 SigmaIV 4 to 5 SigmaV > 5 Sigma

Error Budget WindowWill the design meet requirements? With what confidence?

Sigma Value = (TEa – BIAS) / (SD)Sigma Value = (TEa – BIAS) / (SD)



Relationship of Sigma Level to QCAs the Sigma Level Rises, the amount of Customer QC drops and the financial impact to the customer is improved making the Customer more Competitive

Reference: Westgard Workshops - Quality Assessment from Test Outcome Data: Use of PT Data to Estimate Quality of Lab Tests, 2006 – Madison, WI

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.0 1.0 2.0 3.0 4.0

1.65 2.65 3.65 4.65 5.65

13s/2of32s/R4s/31s/6x 0.07 613s/22s/R4s/41s 0.03 412.5s 0.04 412.5s 0.03 213s/22s/R4s 0.01 213s 0.00 213.5s 0.00 213s 0.00 1

Pfr N

Pro

bab

ility

fo

r R

ejec

tio

n (

P)

3 4 5

DesirableError

Detection

DesirableFalse

Rejection

Systematic Error (SE, multiples of s)

Sigma Scale



Why High Analytical Capability Is Important to a Laboratory Director

• Lower performance leads to increased costs

• 6 sigma performance is our ideal benchmark for practical reasons

Sigma Metric

QC Impact Example QC(at 90% AQA based upon OpSpec Charts)

FalseRejects

6 Sigma Any rule will do 3.5 SD Rule with 2 controls

~0%

5 Sigma Relatively simple QC Rules 13s/22s/R4s

with 2 controls~1%

4 Sigma Complex QC 13s/22s/R4s/41s

with 4 controls(e.g. 2 reps at 2 ctrl levels)

~3%

<3 Sigma Extensive QC. QC alone will not assure quality

13s/22s/R4s/31s/6x with 6 controls (e.g. 2 reps at 3 ctrl levels)

~7%

Hig

her C

osts an

d C

om

plain

ts



Value Stream MappingA Flow Chart with Data to Help Visualize Problems

1Suppliers Customers2 3 4

Sub-Optimal Process:

1Suppliers Customers2 4

Future State Process:> 99% FPA

65% FPA

Redo

QC QC QC QC

99%98%90%

Rework

Scrap

Delay



Independent Analysis Confirms Six SigmaSten Westgard independently reviewed data from Abbott’s 2007 AACC ARCHITECT c16000 technical poster

• "7 out of 9 methods are world class” - “It is rare to find so much good news in a method evaluation study

< 3 Sigma = Poor Performance

3 – 6 Sigma = Good Performance

6 Sigma = World Class Performance

http://westgard.com/qcapp45.htm - QC APPLICATIONA Site Evaluation of Abbott Architect c16000



How do our competitors perform with Six Sigma?

• An evaluation of several common, high volume assays on a competitive chemistry system is published on www.westgard.com

• 4 out of 8 assays are < 6 Sigma (50%)

• 2 out of the 4 assays are < 3 Sigma = Poor performance!

0

1

2

3

4

5

6

7

8

9

10

Sodium Potassium Chloride BUN Glucose CK Total Protein HbA1c (%)

Abbott Competitor

6 Sigma



Error Budgeting Designing a Safe Product that will make our Customers more Competitive

ConventionalRequirements

Approach

Analytical QualityPlanning Budget

StableImprecision

StableInaccuracy

StableImprecision

StableInaccuracy

QC Safety Margin

Traditional Error Budget Approach

TotalError

Allowablein a

Patient Result




Assessing Capability Impact to the Business Lower Sigma Levels will result in Significant QC and Costs



Assessing Capability Impact to the Business Higher Sigma Levels will result in Reduced QC and Costs




Assessing Capability Impact to the Business Difference in Sigma Levels can be Used as a Competitive Advantage due to Differences in the Impact to the Business - Safety, Process Flow and Cost




Assessing Capability Impact to the Business Difference in Sigma Levels can be Used as a Competitive Advantage due to Differences in the Impact to the Business - Safety, Process Flow and Cost




Conclusions Benefits of Using Quality Specifications

• Clinical Laboratories

– Directly links assay performance requirements to medical utility and to Laboratory workflow impact due to Quality Control.

– Provides a clear unambiguous means for clinical laboratories to communicate their performance needs to diagnostic manufacturers.

• Diagnostic Manufacturers

– Provides a means for effective development and control of performance specifications

– Provides a means for diagnostic manufacturers to clearly discuss on-going Quality performance with their clinical laboratory customers.

Thank You !!!

A Promise for Life

Click here to read this Lean in Supply Chain presentation.ppt

Documents

accepted literature practice

5bias cvdefectstrue valuesee

method decision chart

competitive advantage due

test outcome data

allowable analytical error

sigma quality design

wiassessing capability impact