Transcript Center for Business Intelligence & Analytics

Process Analytical Technology
Solution Presentation
for Actionable Information
Center for Business Intelligence and
Analytics (C-BIA)
for Actionable Information
Mission
“Create business value for clients by enabling
superior performance through unleashing
hidden wealth in operational and external
data sources combined with innovative
Analytics.”
23-Jan-07
3
History
Cyrus Mehta, Ph.D.
Founder and President
Cytel Inc.
Nitin Patel, Ph.D.
Founder and Chairman
Cytel Inc.
www.cytel.com
C-BIA, a division of TechKnit was founded in 2004
23-Jan-07
4
Founders
Cyrus Mehta, Ph.D. - Founder and President , Cytel Inc.







An influential thought leader in the area of biostatistics
Dr. Mehta has concentrated his research activities on developing software and
innovative methods for flexible clinical trial designs and non-parametric exact
statistics.
Dr. Mehta has published over 65 papers in journals like JASA, Biometrika and
Biometrics.
He and his co-authors, Dr. Nitin Patel and Dr. Karim Hirji received the 1987 George
W. Snedecor Award from the American Statistical Association.
In 1995 Dr. Mehta was elected a Fellow of the American Statistical Association.
In 2000, Dr. Mehta was named the Mosteller Statistician of the Year by the
Massachusetts Chapter of the American Statistical Association.
In addition to his activities as President of Cytel Inc, Dr. Mehta has been a member of
the faculty in the Department of Biostatistics, Harvard School of Public Health since
1979.
23-Jan-07
5
Founders
Nitin Patel, Ph.D. - Founder, Chairman and Chief Technology Officer






Dr. Patel is a leading expert on the development of fast and accurate computer
algorithms to implement computationally intensive statistical methods.
He has published over sixty-five refereed papers in the areas of statistics, operations
research and computing.
He and his co-authors, Dr. Cyrus Mehta and Dr. Karim Hirji, received the 1987
George W. Snedecor Award from the American Statistical Association. In 2003,
Dr. Patel was elected a Fellow of the American Statistical Association.
Dr. Patel has been a member of the faculty at MIT's Sloan School and the Operations
Research Center since 1995.
Previously, he was a Professor at the Indian Institute of Management, Ahmedabad,
and held visiting positions at Harvard, the University of Michigan, the University of
Montreal and the University of Pittsburgh.
23-Jan-07
6
SAS
www.sas.com

SAS is the worlds largest Business Intelligence and
analytics software co.

Based out of Cary , NC, USA – SAS has world wide
presence across continents .

In India SAS has a marketing office located in Mumbai.

SAS has a R/D center in Pune with a strength of about
250.

SAS tools provide End to End solution across Enterprise
23-Jan-07
7
SAS Technology Layer and Products
The main technology platform provides the
following components –
 Data Quality
 Data Integration
 Data storage
 OLAP Server
 Friendly Interface
23-Jan-07
8
Focus on Pharma Companies
SAS has many years of experience in pharma reporting and analytics.

Clinical trial research reporting is done in SAS formats.

Base SAS is used by the lead pharma companies .

SAS STAT is a tool used by leading pharma companies.

SAS Graphs and STAT are industry acknowledged leaders in the area
of statistical analysis.

SAS has developed a special focus on regulatory reporting and
Pharmacovigilance reporting.

SAS compliments the stringent requirements of Pharma industries in
terms of Production processes and testing and trials.
23-Jan-07
9
SAS Pharma Focus
Business Subjects
Pharma Subjects

Field Force Incentive

Production Quality (PAT)

Sales Force Effectiveness

Clinical Data Management
(PheedIT)

Forecasting

Pharma Company
Vigilance

CDISC - Clinical Data InterChange Standards
Consortium

Production Dashboard

Sales Dashboard

Inventory Dashboard
23-Jan-07
10
Uniqueness and Expertise
for Actionable Information
People

Team of 50 people in Pune, consisting of:


Statisticians (Ph.D. and Masters in Statistics)
Statistical software developers (Masters in Statistics)





23-Jan-07
Microsoft
SAS
Data Analysts and Business Intelligence solution designers
and developers (MBAs and Masters in Statistics)
Data Managers (MCAs)
Information Technology managers (Engineers and MCAs)
12
Spirit of Research, Innovation and
Experimentation





Cytel is built on research
work of the Founders
Imbibed from the founders
“mind-set”
Collection of people built it
further
Vast repository of
methodologies and
software library
Witnessed in several
products, key amongst
them are:
23-Jan-07
13
C-BIA Management





Mayank Shah Chartered Accountant
Mayank has over 27 years' experience as Consultant, Executive and Academician in
the field of MIS and BI applications for business. Mayank is Consultant and
Executive Director of TechKnit and leads C-BIA.
Ajay Sathe PGDM, IIM, Ahmedabad
Ajay has over 17 years' experience in IT industry specializing in Software
Development and Technology Management. Ajay is Director of TechKnit and CEO
of Cytel India.
Shrikant Athavale, Industrial Engineer
Shrikant has nearly 36 years' experience in Industrial Engineering and Quality
Management. Shrikant is Executive Director of TechKnit and leads C-elt,
an e-Learning unit.
Vanaja Vaidyanathan, MBA and Cost and Works Accountant
Vanaja has over eight years of experience in Business Intelligence practice, including
work experience with A F. Ferguson & Co., Asian Paints, GE Capital and Satyam
Computer services and is in charge of delivery at C-BIA.
Dan Crowell, MSc. in Economics, London School of Economics
Dan is our associate based in USA, looking after business development and client
interaction. Dan worked with IFC, GBI Team during 2004 to 2006 to coordinate
activities in the field in South Asia. He first worked in South Asia in 1999 when he
was a Fulbright Scholar in India.
23-Jan-07
14
Experts Panel





Nitin Patel, Ph. D.
Dr. Patel is a leading expert on the development of fast and accurate computer algorithms to
implement computationally intensive statistical methods. Dr. Patel is Founder and CoChairman of Cytel Software Corporation, Cambridge, MA, USA and Visiting Professor. MIT
Sloan School of Management.
Suresh Ankolekar, Ph.D.
Dr. Ankolekar has over 23 years of academic and consulting experience at Indian Institute of
Management, Ahmedabad (IIMA) and Maastricht School of Management, Netherlands (MSM).
Prof. Ankolekar has developed commercial software to solve large-scale optimization
problems in transportation and has provided consulting in analytical software projects to Cytel
Inc., and others. Prior to his doctoral study in management at IIMA, he worked as industrial
engineer at Larsen and Toubro (Bombay).
Ashok Nag, Ph. D.
Dr. Nag, a former senior executive of the Reserve Bank of India, the central bank and the
monetary authority of the country, is a well-known expert in the banking and financial
analytics, data warehousing and data mining.
Sunil Lakdawala, Ph. D.
Dr. Lakdawala has over 20 years' of consulting experience in IT applications for business
including Data Warehousing and Data Mining. Dr. Lakdawala is a consultant in BI applications
and is visiting professor at S. P. Jain Institute of Management & Research.
V. Chandran, Aeronautical Engineering
Chandran has over 22 years' experience in technology functions, including CTO positions in
companies with sizeable software teams. Chandran is Vice President with Cytel India
heading Technology Management function besides managing consulting assignments.
23-Jan-07
15
C-BIA Partnerships






SAS: Silver Alliance partner…
Mastek: BI Solutions…
Cytel-Cognizant: Pharmaceuticals- clinical
trials…
Statistics.com…XLMiner marketing in USA
Syscon Infotech: BI solutions…
Intech Systems – BI Solutions…
23-Jan-07
16
Expertise





Data Warehousing

Manage large volume of data

Building data warehouse and ‘cubes’
Online Analytical Processes (OLAP)

Studying data patterns by slicing, dicing and drilling

Making inference
Data Mining

Manage large volume of data

Sampling

Building valid models

Making predictions – scoring
Statistical Analysis

Manage large volume of data

Data distribution

Pareto

Outliers

Trends

Correlations
Clinical Trial Reporting and Analytics
23-Jan-07
17
Technology and Infrastructure in Pune





6000 sq. feet of office space in Pune
Secured Network with high bandwidth
Connectivity
Windows and SAS Platforms with more than
six servers
Methodologies and SOPs for, BI solutions,
Analytics and Clinical Trials
Well established software development
practices
23-Jan-07
18
C-BIA Advantage

Focus and Expertise




Character





BI and Analytics
Multiple levels of expertise in BI
Understand business management issues
Entrepreneurial
Innovative
Quick to respond and deliver
Stickler for on-time-zero-defect delivery
Cost advantage
23-Jan-07
19
Selected Customers
ProfitLogic, US
Mastek, India
Trumac, India
Savita Chemicals, Ind ia
Tata Motors Ltd.,
23-Jan-07
Bharat Petroleum, India
Dainik Jagran, India
TAM Media Research, India
KPIT Cummins Infosystems
Ltd., India
India
20
Process Analytics Platform
for Actionable Information
Guidance for Industry
PAT — A Framework for Innovative
Pharmaceutical Development,
Manufacturing, and Quality Assurance

Encourages the right approach … measurement, data
integration, statistical modelling & process understanding based
on data.

Companies able to demonstrate process understanding will be
treated differently, e.g. be allowed to change processes without
revalidation if have data and models to backup decisions.

Essentially companies able to show they are doing the right
things will have relaxed regulation regarding CMC (Chemistry
and Manufacturing Controls).

http://www.fda.gov/cder/guidance/6419fnl.pdf
23-Jan-07
22
PAT Tools From FDA Guidance

Multivariate tools for design, data acquisition and
analysis

Process Analysers (at-line, on-line, in-line
measurement tools)

Process Control Tools

Continuous Improvement and Knowledge
Management Tools
23-Jan-07
23
The SAS Pragmatic-PAT Solution:
Data Integration, Modelling and Control for
Operations
7: Simulate
8: Improve
6: Effectiveness
Modelling
9: Verify
5: Efficiency
Modelling
10: Deploy and
Control
3: Cleanse
2: Integrate
4: Maintain
1: Data
23-Jan-07
24
SAS Pragmatic-PAT Solution Elements
Model Builder
Data Integrator
23-Jan-07
Model Deployer
25
Modelling Cycle: Drives Increased Process
Understanding and Operational Improvement
23-Jan-07
26
SAS Pragmatic-PAT Model Builder Capabilities:
Visual Modeling: Literally “see” and interact with the sources of variation to quickly understand the status quo.
Statistical Modeling: Easily use a wide repertoire of proven statistical technology to target:

Efficiency Models - Predict failures enabling corrective action and control prior to an adverse event (thereby
reducing your rejects and rework).

Effectiveness Models - Understand the root causes and drivers of problems
(thereby
enabling process and systemic improvement)..
Clarify The
Objectives
Extract Analysis
Ready Data
Visual
Modelling
Various Users,
with different skills
and capabilities
New
Information
Statistical
Modeling
Assess The
Findings
Deploy
Enabling Technology (respects wide range of Users and Data Types)
. . . Delivered in a way that respects wide range of user skills and capabilities.
23-Jan-07
27
Mapping of data analysis technology to process capability
and dependence on extent and relevance of measured
inputs:
23-Jan-07
28
Case Study 1: Mature Manufacturing with Few
Measured Inputs








23-Jan-07
Established tablet product manufactured at several doses.
Prior measurement systems based on storing finished material while
offline QA tests performed to assure the finished product meets the
performance specification.
Historically, 16% of production batches fail to meet the 60-minute
dissolution requirement of NLT 70%. QA investigations into lot failures
rarely found an assignable cause.
Team tasked with investigating process and dramatically improving sigma
capability.
Data-sparse situation typical of mature manufacturing; focused on the
process for tablets at single concentration.
Deployed effectiveness modeling techniques to cost effectively increase
process understanding:

Process Mapping to identify key metrics/variables

Retrospective data collection around key variables

Visual Exploratory Data Mining

Decision Trees
Identified and verified interim solution to increase sigma capability from
2.3 sigma to 3.1 sigma with a predicted defect rate of 5%.
Ongoing DOE investigations at reduced scale focused on generating
understanding required to gain further reductions in defects.
29
Mature Manufacturing with Few Measured Inputs
Key processes and inputs associated with excessive variation in
60-minute dissolution
Recursive Partitioning Decision Tree
23-Jan-07
30
Case Study 2: New Production Facility with Many
Measured Inputs









Inhaler product been in commercial production for a couple of years.
Extensive inline measurement systems designed into the facility.
Data-rich environment of 520 measured inputs.

V1 to V30 processing parameters of milling, blending and packaging

V31 to V100 properties of material 1

V101 to V170 properties of material 2

V171 to V520 properties of material 3.
The key performance metric is percentage of a given dose reaching stage 3-4 of a cascade
impactor test, which must be between 15% and 25%.
Prior to application of Pragmatic PAT, 240 commercial batches were manufactured, approximately
14% of which failed to meet the performance requirement of the cascade impactor test. QA
investigations rarely found assignable cause.
Deployed effectiveness and efficiency modeling techniques to increase process understanding:

Decision Trees

PLS

DOE
Variation in four key process variables responsible for batch failures.
DOE used to specify new controls on the four process inputs.
Result is increase in capability to 4.8 sigma with a predicted batch reject rate of 0.1%.
23-Jan-07
31
New Production Facility with Many Measured Inputs
Tree Map of PLS Model Coefficients
Recursive Partitioning decision tree identifies inputs most
strongly associated with variation in % at stage 3-4
DOE Summary Analysis
23-Jan-07
32
Project Approach
Describe
Capability
Credentials
Give examples of
potential benefits
Workshop
Discover Vision
Where does the
client want to be?
What are the
client’s
information
needs?
Pilot
Build a business
case
Calculate the ROI
Present findings
to project sponsor
Pilot
Assessment
Data Source
Review including
Quality
Assessment
Project Scoping
Time to
Information (as-is
and to-be)
Multiple
Processes
PROJECT MANAGEMENT METHODOLOGY
Capability
Presentation
Continuous
Improvement
PROJECT
DEFINITION
AD Assess and Define
PROJECT
PLANNING
AE Analyze and
Evaluate
PROJECT
EXECUTION
AQ Analyze Data
Quality
DE Design
DT Define Target
RQ Resolve Data
Quality Issues
CO Construct
CD Create Data Mining
Mart
LO Load
SE SEMMA
FT Final Test
DP Deploy Platform
IM Implement Model
Project Closure
PROJECT
SUMMATION
RV Review
Ongoing Maintenance
and Operation
23-Jan-07
33
Actions & Next Steps




Capabilities Presentation
 Capabilities, credentials & references
Workshop
 Discovery
 Where do You want to be?
 What are Your business needs?
Value Assessment
 Create the business case
 Calculate the ROI
Building Application
 Multiple Process
 Continuous Improvement
23-Jan-07
34
Thank You
Any Questions?