behavioral economics and managment of pharmaceutical qbd 25 august 2016

Behavioral Economics and Management of Pharmaceutical Quality by Design

Ajaz S. Hussain, Ph.D.

8/25/2016 © Ajaz S. Hussain | Insight, Advice & Solutions LLC 1

Pharmaceutical knowledge pyramid

can be toppled easily!

Serendipitous intersection of Behavioral Economics & CGMP

The Context


Breach of Data Integrity: one of the most important and relevant topics currently

Could we have anticipated it?

..a lot going on underneath

Open question - effective CAPA?

Example observations

Symptom or cause?

Context

Publications:

Culture of Pharmaceutical Quality: Connecting the Dots. Biopharma Asia September/October 2015

Culture of Pharmaceutical Quality Management System. Biopharma Asia November/December 2015

Culture of Pharmaceutical Quality: Personnel Development. Biopharma Asia March/April 2016

LinkedIn Posts:https://www.linkedin.com/in/ajazshussain

Slideshare :http://www.slideshare.net/a2zpharmsci


An urgent request, in 2014, to develop and conduct a training on Culture of Pharmaceutical Quality for management and staff at facilities in India

FDA observations alleged serious breach in data integrity

The program aimed at addressing behaviors; it included case examples (a collection from my FDA experience) of blind-spots in understanding of CGMP and the potential for adverse consequence to patients when we deviate

Behavioral economics (per appreciation gained during a project at a tobacco company), and not Ethics/Morality, informed the design of course content; not to “point finger” and to discuss “uncomfortable” topics

Similar training conducted at several companies all across India, 6000+ individuals participating. Follow-up surveys . LinkedIn & Slideshare for updates/discussion.


https://www.linkedin.com/in/ajazshussain

http://www.slideshare.net/a2zpharmsci

Breach of Data Integrity

Testing into compliance, data manipulation, data deletion/ record destruction, misreporting, disregarding failing and/or questionable results, all leading to possible breaches in the integrity of critical data, has become one of the most important and relevant topics currently ..

500

1000

1500

2000

2500

3000

3500

2008 2010 2012 2014 2016

FDA CGMP Inspections Overseas Drug Facilities

?

Carmelo Rosa. DIA Multicenter International Data Integrity Workshop. 13-14, Nov. 2014. Bangalore, India.


"Data integrity [issues] really sounds off alarm bells for us ... if you see data integrity [issues] on the surface, there is likely a lot going on underneath.“

Thomas Cosgrove, FDA; on Observations related to Breaches in Data Integrity


Could we have anticipated it?

..a lot going on underneath..

FDASIA

New Office of Product Quality @ CDER/FDA

“One Quality Voice”

Will these changes get to the root-cause?


Latent factors and a globalized supply chain…

Variance and skewness of latent factors

The Gold Sheet. 28 March 2014.

A sharp increase in data integrity findings has raised questions about the quality of pharmaceutical manufacturing in India.

What’s happening and why, who’s responsible, and how can drug makers and the regulatory agencies that watch over them put a stop to it?


• Look Out for These Data Integrity Issues

• Focus on India Masks the Real Data Integrity Problems

• Data Integrity Requires Stronger Local Enforcement

Thakur Cahilly

Takahashi

These violations cause your drug products to be adulterated within the meaning of 501(a)(2)(B)…..

Your laboratory records did not contain all raw data generated during each test for finished drug products manufactured at your firm. Your quality unit relied on incomplete records to make batch release decisions in support of regulatory submissions to the Agency.

A QC analyst deleted original test method validation data and admitted plans to fabricate sample preparation data.

The trial injection was stored in the “trails” [sic] folder located on a personal computer. The release chromatogram identified injection (b)(4)141119009 as the sample. The trial and release chromatograms for (b)(4)141119009 do not match, and they identify different peaks.


Directive 2003/94/EC GMPC Withdrawn!


Evidence of data integrity issues within GMP documentations, buildings were also falsified to mislead……

WHO Prequalification Team - Inspection Services

Serious concerns regarding the integrity, reliability and accuracy of the data generated and available at your manufacturing site and on your ability to prevent contamination and crosscontamination of your products.

The company failed to adequately perform dissolution tests and may have manipulated dissolution test results

The laboratory was requested to perform, under observation, the dissolution tests for X

During the first test, one of the solution vials inside the auto-injector was switched, without notifying inspectors of what was being done

However, after injection, the refrigerated bracketing solution did not fall within system suitability acceptance criteria (x.xx % RSD) and the run was rejected by the company

Dissolution results were … did not comply …

The run was restarted overnight in absence of the inspectors and passing dissolution results …were obtained

The inspectors requested that the dissolution test be repeated, in front of them

Results of…..were obtained, which differed from the results obtained by the laboratory in the absence of inspectors


Breach – to act contrary to (law, promise)

BDI - a new trend and/or improved detectability?

Errors vs. malicious? Operations vs. Development? Generic vs. Brand? Domestic vs. Foreign?

Symptom or a cause?

Consequence (severity), likelihood, detectability of DI errors or breaches?

Individuals, their company, shareholders

Regulators (CGMP inspection, CMC review, Management, local authorities vs international)

Competing companies and the pharma sector

Insurance companies and payors

Public (within a country, global)

Patients and their family members

Degree of cognitive dissonance?

• Sufficient to motivate a change?

• Ample reasons to rationalize away dissonance?



Why attention to Behavioral Economics can improve our management of QbD work-streams?

How? What (benefits)?


https://www.pinterest.com/pin/313563192778012694/


Journey across India

Insights from CPQ Training

Reaction of participants to FDA WL’s vary

Difficult to comprehend why now; occasional suspicion of FDA’s intent

Regulator heterogeneity (US FDA Vs. MHRA, ..) difficult to reconcile

Quality = Meeting specifications and it is generally good; sincere belief

Growing recognition - importance of documentation and ALCOA in need for significant attention; malicious intent a rare exception

Good technical abilities; attention on communication & experience needed

For the paradigm “Quality = Conforming Test Results “

Ability to design/plan for effective SOP’s in need for attention

Hierarchal organizations, distance from supervisor and management

Over-optimistic expectations (e.g., RFT = 100%) common; can be rationalized via “FDA Approved/Inspected” and “Process is Validated”

“Looking & Being Good = Meeting Expectations”; How to “Do Good” ?


Dominant Paradigm

Ontology: What is pharmaceutical quality?

Epistemology: How do you know what you know?

Methodology: What is your plan of action (e.g., in case of OOS)?

Methods: What do you use to collect data?

Conforming to Test Results

Conforming to specifications per ANDA/USP

Consulted USP & followed instructions (SOPs); who wrote the SOP and how? Was it “validated” for your QC?

File a deviation report and follow SOP for investigation (Phase I then, if necessary Phase II)

Check the methods used before; resampling/testing in Phase II

A hypothetical case example to illustrate one challenge

Makes analysts (and supervisors) “accountable”

But we only define responsibilities for analysts and supervisors?

Types of human errors in QC plus the stress of inappropriate method reproducibility & repeatability?

Efforts needed for “Full Investigation” and its potential consequence?

Re-testing and Re-sampling?

Corrective and preventive action plans?



Head of Operations (the Boss)

B.Pharm/MBA Experience in

Pharma 20 Years

Calls the analyst and supervisor

On several previous occasion after a lengthy investigation it turned out to

be analyst’s error. “Complete the investigation quickly and release the batch on time!” The product is FDA approved and

process validated. The QC lab is not well managed! (thoughts)

Head of Analytical

Ph.D. (GC-MS)Experience in

Pharma 15 YearsX

When the CEO of the company will learn about this it will not look-good for

me and the Analytical Department. “Add as many resources you need; complete the investigation quickly!“

Supervisor(QC)

M. Pharm. Experience in QC 10 Years

SOP Training

Per SOP, Deviation

Report.

This is not a GC test. I don't know how to speedup OOS investigations. What if the root cause is unknown; again?

“We are looking bad!”

Amos Tversky and Daniel Kahneman. The framing of decisions

and the psychology of choice. Science. 1981 Jan 30;211(4481):453-8.

The psychological principles that govern the perception of decision

problems and the evaluation of probabilities and outcomes produce

predictable shifts of preference when the same problem is framed

in different ways.

Analyst (QC)

M. Pharm. Experience in QC 1YearSOP Training

Ooops! OOS!Modified Release Tab.

Dissolution Test

Hussain, A. S. Culture of Pharmaceutical Quality: Connecting the Dots. BioPharma Asia September/October 2015

On the framing of decisions and the psychology of choice

Start here START


Framing & “Testing into Compliance” Dissolution Testing : Many reasons to rationalize away dissonance

“Testing into compliance”

In general – low empowerment is a

significant challenge (low perceived behavioral

control); plus reasons to rationalize….

Attitude toward

performing the behavior

Process validation is

done so quality is good;

Test prone to error

“Batch failure means I made a

mistake”

Subjective norm (local)

Documentation not critical

Compendial testing

sufficient

Regulators collect & test samples – no

issue there


Searching for answers after failing!

“we can be blind to the obvious, and we are also blind to our blindness.” ― Daniel Kahneman, Thinking, Fast and Slow

July 10, 2015

RE: USP Prednisone Tablets Reference Standard (Catalog #1559505 Lot R001B0)

Dear USP Customer Our records indicate that you have recently

purchased USP Prednisone Tablets Reference…. Recent evaluation…..Lot R001B0 is no longer suitable for its intended … use. USP regrets issuing this notice after the valid use date has passed; please be assured that USP makes every effort not to do so. A new lot is expected to be available no later than July 31, 2015.


My personal realization of a “streetlight effect”

A slide from my presentation at the FDA’s ACPS Meeting, May 2005


Our cognitive biases: Can keep us busy reacting to ‘common cause’ variation


We keep using “market standards” as batch release tests; when there are alternative approaches

Critical importance of reproducibility and repeatability of our test methods - for physical quality attributes is widely underappreciated

Although we are starting to realize that we can no longer afford the inefficiency it imposes; our focus remains on the least common denominators

Outsourcing for lower costs with an expectation of efficiency is our collective “streetlight effect”

600 days*

Good Development to Good Manufacturing Practices

PAT, ICH Q 8-11, PV 2011

Management of QbD

Paradigm[Ontology][Epistemology]Methodology (ICH Q8; QbR)[Methods]

[Need specific attention]

Product & Process Development

• Stage 1: Process Design

Scale-up , Transfer, Validation

• Stage 2: Process Qualification

Commercial Manufacturing

• Stage 3: Continued Process Verification


10X, 3 batches?Monitoring & Control

Validated rangesProcess Performance

Explicit assurance of DI

Statistical ConfidenceState of Control; CapabilityCommon vs. Special Cause

Effective CAPAExplicit assurance of DIContinual Improvement

1 batch, 3 batchesPatient failure modes?

Manufacturability?

Complex generics: Totality of evidence

New Products New & Existing Products

Changing regulatory defaults & expectations: Heterogeneity during transition process

We must confront this irrationality; starting with how we establish regulatory specifications – before not after the product has been developed!

“Therefore, we propose that the sponsor’s specification of Q=80% at 60 min should be changed to a specification of Q=80% at 30 min.”

Was the product designed and developed with the FDA proposed spec. as the target? If not, why would you accept it – just to get the approval? How should you make the case for the right specification?


A slide from my presentation at the FDA’s ACPS Meeting, May 2005

Bla

ck-b

ox

(bef

ore

ICH

Q8

)

Process Capability and Measurement Process Capability and Measurement

Capability: Dissolution TestCapability: Dissolution Test

•• When we evaluate process capability by When we evaluate process capability by

measuring variability in the product measuring variability in the product

producedproduced

•• Total variability Total variability σσ 22TotalTotal

•• Assuming independent variable (if not independent Assuming independent variable (if not independent

for example interaction between measurement and for example interaction between measurement and

product a covariance term needs to be included)product a covariance term needs to be included)

•• σσ 22TotalTotal = = σσ 22

ProductProduct + + σσ 22MeasurementMeasurement

•• σσ 22MeasurementMeasurement = = σσ 22

RepeatabilityRepeatability + + σσ 22ReprodicibilityReprodicibility

© Light Pharma

Process Capability: If you can’t measure it, you can’t improve it

Process Capability Roadmap:

1

Has Measurement

System capability

been verified?

STOP!

Do not compute

Proc. Cap. statistics.

Improve the Meas. System.

No

2

Is the process stable

or unstable via SPC?

Yes

STOP!

Do not compute

Proc. Cap. statistics.

Investigate special causes.

Improve process stability.

3

Is the data normal

“enough” via the

Normality Test?

STOP!

Transform data.

No4

Compute

Cpk

Yes

Unstable

Stable

0

Challenge

Specs!

p-value < 0.05

p-value > 0.05

Gage R&R

& Calibration

SPC Charts

Scott Tarpley, UK Arden House 2004

Slides from my presentation at USP Annual Scientific Meeting "The Science of Quality“. September 26–30, 2004: Biopharmaceutics and Drug Product Quality: Performance Tests for Drug Products, A Look Into the Future. And, from a more recent visit (below)….


Management of QbD can be far from ideal

Unaccounted Cognitive Biases: Blind spots, Streetlight effect,….

Market failure

Information asymmetry

Decisions under risk

Many reasons to rationalize away dissonance

QbD a Paradigm• Origin “quality cannot be tested into products, it has to be

built-in by design” (FDA 1987 Process Validation Guidance), adopted in the FDA’s PAT Guidance (2004)

Ontology?

• What is pharmaceutical quality? What is Assurance? Why is CGMP & DI critically important?

• When a placebo effect is very significant, and generics are promoted as cheap – how effective will be the generic on substitution?

• From an IT perspective – what is the current status of our Inactive Ingredient Guide? Are excipients truly inactive? What about functionality?

Epistemology?• How do you know what you know? Regulatory defaults: 10X, 3

batches, file first and figure it out later – will you?

Methodology for QbD

• ICH Q8 outlined a methodology that facilitates regulatory communication of scientific understanding of critical factors and their risk-based controls in the CTD-P2 section of regulatory submissions

Methods?• Are our methods for physical attributes appropriately

validated for use in QC? Do we understand DOE?.......


Market failure

Information asymmetry

Decisions under risk

Unaccounted Cognitive Biases

Many reasons to rationalize away dissonance

Clinical & Quality disconnect; quality = specifications, lack of pharmacovigilence signal= quality is good; placebo effect

Risk-uncertain: CQA’s, CPP’s, CMA’s (certificate of analysis); SOP’s ≠ CPP; Training on SOP for 15 minutes; Validation = state of control; Rate of right first time often in 60-90% range; assumed to be 100%; , Market standards for batch release; OOS root-cause often “unknown”; Common cause –to- special cause confusion; CAPA for common cause change SOP/re-train; System = folder of policies and SOP’s written for regulators

Regulatory defaults: 10X scale-up factor, 3 batches, bio-batch, 80-125%, Pharmaceutical equivalence, File –first and figure it out later works!; Regulator knows best why should I do more?

Nudges and frames targeted towards FDA Approval, Passing GMP inspection, business targets with less than optimal focus on patients

Significant Framing Effects, recurring OOS with root cause unknown, time and $ spent in Operations = Costs (R&D $ = investment) & regulator heterogeneity


We are finally beginning to understand that irrationality is the real invisible hand that drives human decision making. The End of Rational Economics. HBR July-August 2009

Because self-interest of organizations does not assure protection of their own shareholders ($ -measurable); what about the interest of patients (which is not easy to measure; “market failure”)?


Despite the progress in regulatory & pharmaceutical science

Pharmaceutical knowledge pyramid can be toppled easily!

Why attention to Behavioral Economics can improve our management of QbD work-streams? Because…..“we can be blind to the obvious, and we are also blind to our blindness.” ― Daniel Kahneman, Thinking, Fast and Slow


Unaccounted Cognitive Biases & Many reasons to rationalize away dissonance

https://axispraxis.wordpress.com/2016/03/24/the-streetlight-effect-a-metaphor-for-knowledge-and-ignorance/

http://www.nbc.com/blindspot

http://dilbert.com/strips/comic/2000-02-03/


Why attention to Behavioral Economics can improve our management of QbD work-streams?

How? What (benefits)?




Illustration of a typical application of Behavioral Economics

Default Option: Generic Medication Prescription Rates

Patel MS, Day SC, Halpern SD, et al. Generic Medication Prescription Rates After Health System–Wide Redesign of Default Options Within the Electronic Health Record. JAMA Intern Med. 2016;176(6):847-848.


The overall generic prescribing rate increased significantly from75.3% during the 10-month pre-intervention period to 98.4%

There was less of an increase for levothyroxine representing a greater proportion of opt-outs (95% CI, −16.4 to −14.8; P < .001)

Levothyroxine

Human development: Ability for system or integrative thinking

Is your team at 3rd

Order of Consciousness?


Need to account for variance & skewness in the distributed understanding of pharmaceutical quality

Globalized supply chain, regulator heterogeneity, demographics: Is the current approach to education, training & experience adequate?


Elephant in the dark or men with blindfolds on?

Ex

pre

ssP

ha

rma

16

Ma

y 2

016

A 21st Century Fable about Pharmaceutical Quality and Preventing a Clash of Cultures

https://www.linkedin.com/pulse/21st-century-fable-pharmaceutical-quality-preventing-hussain-ph-d-

Integrating Big Ideas: Fisher, Shewhart, Deming, Juran, Kegan, Kahneman,..


Epistemology essential in human development (to higher Orders of Consciousness) and for overcoming Immunity to Change (Professor Kegan)

Human irrationality (Prof. Kahneman, Ariely and others)

“For instance, few realize that we are changing the brains of schoolchildren through medication in order to make them adjust to the curriculum, rather than the reverse.” ― Nassim Nicholas Taleb, The Bed of Procrustes: Philosophical and Practical Aphorisms

Advice & Solutions:Informed & Guided by Behavioral Economics

Re-building or strengthening

Culture of Pharmaceutical

Quality

• Training program with a focus on blind-spots and other prevalent biases, Score-card, Pledges, Nudges

• For several companies

Question based Development Management

• Breakthrough therapy orphan drug, complex generics & biosimilars

• Asking the right question at the right time; explicitly stating/tracking assumptions accepted, pre-defining the precision for the answers sought (stage-n-gate criteria)

Technology transfer,

knowledge management and

Process Validation per 2011 Guidance

• A framework to risk-assessment & mitigation for current products (without a comprehensive development report) and for new products


CPQ: Founded on the paradigm of Quality by Design

We do our best to develop medicines and the evidence needed to satisfy the needs of patients – we develop these products consciously, recognizing that quality cannot be tested into our products .

We know that nothing is perfect and there will be some errors in our design, systems and procedures, or we may make mistakes in following set procedures.

It is normal, easy and rewarding to work within our quality management system, without fear, to detect, correct and to learn from our mistakes.

In doing so we act consciously in the interest of patients – especially when no one else is looking, and we continually improve our quality by design and aim for right first time.



Framework: Culture of Pharmaceutical Quality (CPQ)

Culture –Pharma Quality

Quality is Normal

Quality is Easy

Quality is Rewarding

System-

QMS

Appreciate System

Theory of Knowledge

Knowledge of Variation

Psychology of Change

Practices-GXPs

Fear Removed

Mastery

Awareness

Environment Leadership Emphasis Message Credibility Peer Involvement Employee Empowerment

Connect to CultureConnect to Practice Quality by Design


Score-card; Key Areas of Improvement and Recommendations

Illustrative Score Card: For a company currently in cGMP remediation.

Stars & color codesGood progress, continue efforts already initiated

Blind-spots, current efforts need additional considerations

Blind-spot + new targeted projects/efforts needed


An illustrative impact of training (survey via Survey Monkey® 3 months after training)

Rigorous adherence to science with effective peer review would make a discussion on behavioral economics -irrational!

Current (regulatory) defaults can become cognitive biases which accumulate silently because there is no clearance mechanism -

eventually knowledge pyramid is toppled!

Specifically for breakthrough drugs (rapid development), Complex Generics, Biosimilars,…

Time-crunch and/or increased complexity!

ICH Q8 outlined a methodology. Certain gaps remain in our ontology (e.g., excipient functionality),

epistemology (e.g., blindly following FDA guidance/USP monographs) and methods (e.g., physical attributes)

Many aspects of our pharmaceutical & regulatory science is not rigorous!

Watch out! Some current regulatory defaults can become cognitive biases!


Additional Information

• http://www.slideshare.net/a2zpharmsci/complex-generics-developing-defensible-statistical-analyses-and-acceptance-criteria

Complex Generics Developing Defensible Statistical Analyses and

Acceptance Criteria

• http://www.slideshare.net/a2zpharmsci/biosimilar-development-eptm-2015

Biosimilar Development EPTM 2015

• http://www.slideshare.net/a2zpharmsci/breakthrough-designation-opportunities-challenges-ajaz-hussain

Breakthrough Designation Opportunities Challenges

AAPS 2014


http://www.slideshare.net/a2zpharmsci/complex-generics-developing-defensible-statistical-analyses-and-acceptance-criteria

http://www.slideshare.net/a2zpharmsci/biosimilar-development-eptm-2015

http://www.slideshare.net/a2zpharmsci/breakthrough-designation-opportunities-challenges-ajaz-hussain

We shall not cease from exploration, and the end of all our exploring will be to arrive where we started and know the place for the first time. T. S. Eliot


Between stimulus and response there is a space. In that space is our power to choose our response. In our response lies our growth and our freedom. Viktor E. Frankl

Thank you


Between regulatory query and response there is Design Space. In that space is our comparability protocol…

behavioral economics and managment of pharmaceutical qbd 25 august 2016

Leadership & Management