Quality Risk Management Maria Teresa Cruañes Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc. West Point, PA 19446 USA
Manufacturing Initiative for the 21st Century Sindusfarma, Sao Paulo, Brazil 15 June 2010
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
Key references International Conference on Harmonization (ICH) Guidance for Industry http://www.ich.org/cache/compo/363-272-1.html
Q8: Pharmaceutical Development (2005) Q8 (R2): Pharmaceutical Development Revision (Nov 2009) Q9: Quality Risk Management (2005) Q10: Pharmaceutical Quality System (2008) – Also: “Questions & Answers on implementation of ICH Q8, Q9, Q10 (2009)” • clarification, ongoing inclusions of additional Q&A’s Other FDA’s Guidance for Industry on Process Validation: General Principles and Practices (Draft Guidance in progress) Project Management Body of Knowledge (PMBOK), Chapter 11 (on Risk Management) The Lean Six Sigma Toolbook, 2005 SixSigma Advantage Training Material, 2007 Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
2
Outline •
Introduction to Risk Management – Definitions – Steps – Example
•
Pharmaceutical Quality Risk Management (QRM) – ICH guidance • Steps of QRM • Definitions
– QRM through product lifecycle – QRM Methodology • Some Tools and Examples » » » »
• •
5 Why’s Ishikawa diagram Cause and effect matrix Failure Mode and Effect Analysis
Concluding remarks QbD case study (if time allows)
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
3
Risk and Risk Management • Risk is an uncertain event or condition that, if it occurs, has a negative impact on – Project objectives- product quality and product availability – Project timelines, cost, resources
• Risk Management: process of identifying, analyzing and responding to project risks, thereby reducing their impact on the project. – Goal: maximize positive outcomes and minimize negative ones – Philosophy: deal with known risks proactively before they occur
• There is no ONE way to manage risk. There are various methodologies available. Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
4
Risk and Risk Management (continued) • Risk components are – – – – –
The harmful event/condition or hazard The cause(s) The likelihood or probability of occurrence of risk (P or O) The impact or severity of the effects of the event (I or S) The detectability of the cause of the event (D)
• Risk Priority Number = RPN = P x I – Detectability may be included Æ RPN = P x I x D (or O x S x D)
• One possible scale for computing risk: 1-5 – P or O: 1-Very Low, 2-Low, 3-Moderate, 4-High, 5-Very High – S or I: 1-Very Low, 2-Low, 3-Moderate, 4-High, 5-Very High – Here RPN ranges from 1 to 2
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
5
Risk and Risk Management (continued) •
Risk management steps are – Initiate process • risk statement, scales and acceptance criteria – Risk and cause identification • what should I be concerned about? • what is the cause(s)? – Risk Analysis • how much should I be concerned? – how probable of occurring? how impactful/severe ? RPN?
– Risk Response/Control • what should I do about it? Use RPN to prioritize responses • possible response strategies: – AVOID or eliminate effect or cause – MITIGATE or reduce impact/severity – TRANSFER impact to another party (shared responsibility) – ACCEPT risk-actively (with a contingency plan) or passively (do nothing) Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
6
Risk Management Example Project objective: Attend Final World Cup Game (Brazil vs Argentina) on 11 JUL in South Africa Book Flight
Drive to Airport
Board Airplane
Fly
Take Taxi
Arrive to Stadium Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
7
Risk Analysis Example Project objectives: Attending the Final World Cup Game (Argentina vs Brazil) on 11 JUL in South Africa Risk
Cause
Effect
If …
…due to
…then
I arrive late to Johannesb urg
Flight delay
My car breaks on the way to airport and I miss flight
I miss the game
Probability Impact Scale 1-5 Scale 1-5 (of risk due to given cause) 3
2
5
1
RPN (risk #)
Risk response
15
?
10
?
5
?
All flights that week canceled due to weather
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
8
Probability of occurrence
Risk Response GuidanceAn example of Probability x Impact Matrix Very high
5
D
A/B/C
A
A
A
4
D
A/B/C
A/B/C
A
A
D
A/B/C/ D
A/B/C
A/B/C
A
2
D
A/B/C/ D
A/B/C
A/B/C
10 A/B/C
1
D
D
A/B/C/ D
A/B/C/ D
5 A/B/C
1
2
3
4
5
3
Very low
Very low
15
A - Avoidance B - Mitigation C - Transference D - Acceptance
Very high
Impact/Severity (of Effect/Consequences) Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
9
Risk Analysis and Response Example Project objectives: Attending the Final World Cup Game (Argentina vs Brazil) on 11 JUL in South Africa Risk
Cause
Effect
If …
…due to
…then
I arrive late to Johannesburg
Flight delayed
I miss the game
Probability (of risk due to cause)
Impact
RPN (risk #)
Risk response
15
Avoid: Book flight departing three days earlier
2
10
Avoid: do not drive, take shuttle service Reduce: Car inspection and maintenance a week earlier
1
5
Avoid: cancel Transference: buy travel insurance (mitigates financial impact) Accept: 10 Do nothing
3
5 My car breaks on the way to airport and I miss flight
All flights that week canceled due to weather Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
Why Perform Risk Management in the Pharmaceutical Industry? •
To be proactive – Enhancing quality, minimizing delays and costs
•
To employ best practices – Reduces subjectivity, documents rationale for decision and cost-benefit trade offs and harmonizes communication
•
To allow prioritization (ranking of risks) – Understand our business and what matters most – Allocate resources accordingly
•
To better fulfill our mission and meet customer needs – Develop and deliver quality medicines to people that need them quickly and consistently
•
It is a regulatory expectation – Quality Risk Management (ICH Q8, Q9 and Q10)
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
11
Quality Risk Management (ICH Q9, Nov 2005) A systematic process for the assessment, control, communication and review of risks to the quality to the drug product across the product lifecycle.
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
12
Steps of Quality Risk Management (ICH Q9) 1. 2. 3. 4.
5.
Risk Management Process Initiation Risk assessment consists of the identification of hazards and the analysis and evaluation of risks associated with exposure to those hazards Risk control includes decision making and implementation of actions to reduce risks to acceptable levels. Risk communication is the sharing of …the output/result of the quality risk management process [which] should be appropriately communicated and documented. Risk Review A mechanism for monitoring events [on an on-going basis]… and to take into account new knowledge and experience.
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
13
Quality Risk Management- ICH Q9 (continued) Initiating a Risk Management Process: Quality risk management should include systematic processes designed to coordinate, facilitate and improve science-based decision making with respect to risk. Possible steps used to initiate and plan a quality risk management process: • Define the problem and/or risk question, including pertinent assumptions identifying the potential for risk; • Assemble background information and/ or data on the potential hazard, harm or human health impact relevant to the risk assessment; [need experts!] • Identify a leader [and/or facilitator] and necessary resources [team]; • Specify a timeline, deliverables and appropriate level of decision making for the risk management process. •[Identify a methodology, rating scales and risk acceptance criteria]
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
14
Quality Risk Management Steps ICH Q9(continued) 1.
Risk assessment • • •
Risk identification Risk analysis Risk evaluation
Key questions: – What might go wrong? – [What are the potential causes?] – What is the likelihood (probability) it will go wrong? – What are the consequences (impact or severity)? – Output is a quantitative estimate (eg, RPN) or a qualitative description of range of risk (eg, low, medium or high) – How does the risk estimate compare against given risk criteria?
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
15
Quality Risk Management Steps ICH Q9 (continued) 2. Risk Control The purpose of risk control is to reduce the risk to an acceptable level. The amount of effort used for risk control should be proportional to the significance of the risk. Key questions: – How is an acceptable level of risk defined? –
– – – – – – –
Criteria for risk acceptance will vary with organization, type of hazard, product and lifecycle (general rules under methodologies later)
Is the risk above an acceptable level? How are the risk acceptance and control decisions documented? What can be done to reduce or eliminate risks that are above acceptable levels? What are the possible risk reduction scenarios? What is the appropriate balance among benefits, risks and resources? Output: risk response/control strategy Are new risks introduced as a result of the identified risks being controlled?
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
16
Quality Risk Management Steps ICH Q9 (continued)
3. Risk Communication Sharing information about risk and risk management between decision makers and stakeholders – – – –
Industry, regulators, patients and/or within the company About Output of the QRM process And at any time needed during the QRM process [Output depends on the stage of lifecycle (see later)]
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
17
Quality Risk Management Steps ICH Q9 (continued) 3. Risk Review A mechanism for monitoring events [on an on-going basis during product lifecycle]… and to take into account new knowledge and experience. –
Key Question: Is the process in state of control? •
–
Mechanism to review and monitor risk events. • •
– –
Are the control actions fully implemented and working effectively and as expected Planned events, e.g., results from product review, inspections, audits, change controls, etc. [monitoring plans, check lists] Unplanned events, e.g., root cause from failure investigations, etc.
Frequency is commensurate with risk level May include reconsideration of risk acceptance decisions
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
18
A Depiction of Quality Risk Management across Product Lifecycle
Product Development
Process Development
Process Scale-up & Tech Transfer
Manufacturing Implementation
Product/prior Knowledge
Process Understanding
Risk Assessment
Risk Assessment
Risk Control
Risk Review
Excipient & drug substance design space
Process design space
Product quality control strategy
Continual improvement
Product Dev
Process Dev
Process History
Manufacturing Site
Risk Management Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Source: Moheb Nasr, FDA, 2007 Whitehouse Station, N.J., U.S.A. All Rights Reserved.
19
Output of Quality Risk Management Depends on the stage of product lifecycle: –
Product and process development: Inform the development plan to define CQA’s and design space • ...identifying which material attributes and process parameters can potentially have an effect on product CQAs (ICH Q8) •
–
An R&D example provided earlier in previous talk
Process scale up and qualification: Control Strategy •
… implement controls to stay within design space –
–
Many kinds of controls are possible: In process test via PAT, personnel training, equipment maintenance, process parameter controls, specifications, etc.
Manufacturing: Continuous verification and improvement •
Monitor outcomes, update knowledge, re-evaluate risk, revise control strategy, identify early warning signals of process out of control
R&D: research and development Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
20
QRM Methodology •
Non-exhaustive list of available tools: – Risk and Root Cause Identification • • • •
5 Why’s Ishikawa diagram Cause and Effect Matrix Failure Mode and Effect Analysis (FMEA)
– Risk Analysis/Evaluation • Risk Analysis Table • FMEA • Many Other
– Risk Control Tools • • • • • • • •
Statistical tools/models/Simulation Training Procedures Equipment Procedures Process Controls In-process testing End product testing Specifications Etc.
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
21
Risk Identification: 5 Why’s
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
5 Why’s • Risk Statement/Problem: Tablets show slow disintegration (Sometimes it takes several layers of questions to find the root cause, 3, 5 or more)
– Why? • The active ingredient dissolves slowly
– Why? • The active ingredient’s solubility is lower
– Why? • The active ingredient has converted to less soluble hydrated form (there is a known form conversion at 45 % RH)
– Why? • The tablets were exposed to elevated ambient humidity
– Why? • The relative humidity in the room where tablet drums are held is > 70 % RH
Root cause: malfunctioning environmental controls Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
23
Risk Identification: Ishikawa Diagram
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
Risk and Root Cause Identification: Ishikawa (Fishbone) Diagram Milling Screen size (s)
Blending time Blender Type
Screen Type Mill Speed Temperature Relative Humidity Operator Training
Plant Factors
Lubrication Blender Speed
Method
Feed Frame Compressing
Tablet Dissolution
Roll Pressure Roll Speed
Feeder Speed
Tooling
Analyst Sampling
Precompressing Main Compressing Press Speed Punch Penetration Depth
Analytical
Drug Substance Age PS Process Conditions
Roll Type
Diluents
Roll Chiller Roll Gap Feed screw speed Vacuum De‐ aeration
Moisture Potential
PS
Moisture
Other Lubricant Disintegrant
Roller Compaction
Binder
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
Raw Materials
critical quality attributes and process variables in red 25
Risk Identification and analysis: Cause and Effect Matrix
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
Cause and Effect Matrix Output: Product Quality Attributes Input: Variables
Content Uniformi ty
Assay
Tablet Hardness
Dissolution
Tablet Appearance
Output Importance (1-10 scale): Assume 10 for all (depends on TPP )
Correlation of Input to Output (blank: no correlation; 1: remote correlation; 3: moderate correlation; 9: strong correlation) API
Raw Materials
Environment
Process
Particle Size
3
3
9
3
1
Moisture content
1
1
3
3
9
Temperature
1
1
1
1
1
3
3
9
Relative Humidity Roll Pressure
1
3
9
9
3
Compression Force
3
3
9
9
9
Æ Team to study the input and output relationships with the highest scores FIRST to determine if these variables are critical quality attributes/process parameters 27 Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
Risk Identification, Analysis/Evaluation: Failure Mode Effect and Analysis
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
Failure Mode and Effect Analysis (FMEA) • There is NO one single methodology to manage risk • FMEA is one possible and common method • It is a structured approach to – Doing risk assessment – Planning, prioritizing and tracking risk control actions – Risk review after implementation of actions • It is completed by cross-functional group of experts and stakeholders • Requires pre-determined scale criteria to rate • Severity (Impact), Occurrence (Probability) and Detectability • Will vary with industry, company, organization, stage of lifecycle, etc. Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
29
FMEA Inputs and Outputs
Inputs •Process Map •Knowledge •Experience •Procedures •Pre-determined rating scales
Outputs
FMEA
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
•Action Plan •Action Rationale •Actions history •Re-assessed risk
30
Failure Mode and Effect Analysis (FMEA) FMEA Steps 1. Identify the steps and ways in which a process can fail (failure mode) Æ flow diagram 2. Identify the potential effect of step failure on product performance/customer satisfaction. 3. Identify the potential causes of the failure Æ materials, personnel, process & environmental variables
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
31
A Pharmaceutical Process Map API and intragranular excipients Lubricant
Blending
IntraGran. Lubrication
Roller Compaction
Milling
Extra granular excipients Lubricant
Extra-granular Blending
Lubrication
Compression
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
32
Process/Product FMEA Example Process Step
Potential Potential S Failure Mode Failure E Effects V
What is the In what ways What is process Step does the Key the impact /Input under Step/Input go on the investigawrong? Customer tion? Requirem ents?
Compression
OverSlow compression drug (too high release/ force Delayed produces too onset of dense or action hard tablet)
E R I T Y
O C C U R A N What causes C the Key E Step/Input to go wrong?
Potential Causes
Current Controls
What are the existing controls and procedures (inspection and test) that prevent either the cause or the Failure Mode?
Incorrect press setting
Operator 1 entry and Operator 2 check
Equipment malfunction
Fill Depth Gauge Annual Calibration
Die overfill due to variable powder flow
Visual inspection
D E T E C T I O N
R Actions P Recommend N
ed
What are the actions for reducing the occurrence of the cause, or improving detection?
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
Resp.
S E V E R I T What are the Y completed actions taken with the recalculated RPN?
Actions Taken
O C C U R A N C E
D R E P T N E C T I O N
33
Failure Mode and Effect Analysis (FMEA) FMEA Steps 1. Identify the steps and ways in which a process can fail (failure mode) 2. Identify the potential effect of the failure on the customer 3. Identify the potential causes of the failure 4. Estimate the risk of the failure mode = RPN – S = Severity of the effect on the customer (or impact) – P = Probability that the cause will occur (and result on failure) – D = Detectability of the cause of failure – RPN = Risk priority number = S x P x D (or I x P x D)
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
34
FMEA: Example of Scales (adapted from Six Sigma Literature; each organization may adopt their own) Severity of failure B a d
G o o d
Occurrence of cause (probability)
Detectability of failure or cause
10
Injure to customer or employee
10
At least once a day
10
No detection of cause or failure
9
Illegal
9
More than once a week
9
Occasional test or inspection
8
Complete loss of performance and/or major process disruption
8
Once per week
8
Systematic sampling and inspection of units
7
Extreme Customer dissatisfaction
7
Once per month
7
Manual inspection of all units
6
Partial loss of performance and/or process disruption
6
Once every 3 months
6
Manual inspection with mistake proofing modifications
5
Customer complaint
5
Once every 6 months
5
Statistical monitoring of process
4
Minor loss of performance and/or minor process disruption
4
Once per year
4
Statistical monitoring with immediate reaction to out of control conditions
3
Minor nuisance
3
Once every 1-3 years
3
Above and 100 % inspection surrounding out of control
2
Noticeable with no effect on performance
2
Once every 3-6 years
2
All units are automatically inspected
1
Unnoticeable: no effect on 1 Once every 6-100 years 1 performance Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
RPN scale range: 1-1000
Failure of defect is obvious and can be kept from 35 reaching customer
Process/Product FMEA Example Process Step
Potential Potential Failure Mode Failure Effects
What is the In what ways What is process Step does the Key the impact /Input under Step/Input go on the investigawrong? Customer tion? Requirem ents?
Compression
OverSlow compression drug (too high release/ force Delayed produces too onset of dense or action hard tablet)
S E V E R I T Y
Potential O Causes C
8
Incorrect press setting
Current Controls
C U R A N C What causes What are the E the Key existing Step/Input to controls and go wrong? procedures (inspection and test) that TPP (product prevent either performance the cause or requirement) the Failure Mode?
3
Operator 1 entry and Operator 2 check
D E T E C T I O N
R Actions P Recommend N
Resp.
Actions Taken
ed
What are the actions for reducing the occurrence of the cause, or improving detection?
What are the completed actions taken with the recalculated RPN?
S E V E R I T Y
O C C U R A N C E
D R E P T N E C T I O N
2 48
Equipment 4 Fill Depth 8 256 Gauge Annual malfunction Calibration
Variable die fill 6 Visual 7 336 due to poor inspection powder flow Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
36
Failure Mode and Effect Analysis (FMEA) FMEA Steps 1. 2. 3. 4.
Identify the steps and ways in which a process can fail (failure mode) Identify the potential effect of the failure on the customer Identify the potential causes of the failure Estimate the risk of the failure mode = RPN – RPN = Risk priority number = S x P x D (or I x P x D) – S = Severity of the effect on the customer (or impact) – P = Probability that the cause will occur (and result on failure) – D = Detectability of the cause in time to prevent the effect
5. Identify and prioritize future actions to control the risk – Deal with highest ranked risk first • • •
General rule: Deal with all RPNs >100 (if pre-determined risk acceptance is RPN 100 with highest ranked risk first – Assigned actions to responsible parties and specify timing for implementation 6. Later, at specified time(s): Revise risk estimation after actions are implemented – Revised RPN Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
39
Process/Product FMEA Example Process Step
Potential Potential Failure Mode Failure Effects
What is the In what ways What is process Step does the Key the impact /Input under Step/Input go on the investigawrong? Customer tion? Requirem ents?
Compression
OverSlow compression drug (too high release/ force Delayed produces too onset of dense or action hard tablet)
S E V E R I T Y
8
Potential O Causes C
Current Controls
C U R A What causes N What are the C the Key existing E Step/Input to controls and go wrong? procedures (inspection and test) that prevent either the cause or the Failure Mode?
Incorrect force setting
3
Operator 1 entry and Operator 2 check
D E T E C T I O N
R Actions Resp. P Recommended N
2 48
What are the actions for reducing the occurrence of the cause, or improving detection?
Continue Operator Training
S E V E R I T What are the completed Y actions taken with the recalculated RPN?
Actions Taken
O C C U R A N C E
D E T E C T I O N
R P N
Ron
On going
Ian
On going every week
1 7 56
7 336 Monitor/control See Control via prior step: blend NIR blending next spectroscop Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., y
2 3 48
Monthly Equipment 4 Fill Depth 8 256 calibration and malfunction Gauge Annual Calibration press maintenance variable die fill due to poor powder flow
6
Visual inspection
Whitehouse Station, N.J., U.S.A. All Rights Reserved.
8 3 2 48
40
Process Step
FMEA Steps Potential Failure Mode
Potential Effect of Failure
Potential Cause Current Control
Determine Severity
Determine Occurrence
Determine Detectability
Compute Risk Priority Number
Recommend and Implement Actions Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
41
Process/Product FMEA Example S E V E What is the In what ways What is the R process does the Key impact on the I T Step /Input Step/Input Customer Y under go wrong? Requirements investiga? tion?
Process Step
Extra granular Blending
Potential Failure Mode
Potential Failure Effects
Potential Causes What causes the Key Step/Input to go wrong?
O Current C Controls C U R What are the A existing N controls and C procedures E (inspection and test) that prevent either the cause or the Failure Mode?
D E T E C T I O N
R Actions Resp. P Recommend N
Actions Taken
ed
What are the actions for reducing the occurrence of the cause, or improving detection?
Incomplete •Variable 10 Change in 6 Measure 6 360 PAT to mixing Assay/Sub particle size particle size monitor potent of extraof every blending and medication granular raw excipient statistics to materials (but correlate to •And poor unknown particle size powder flow what a non for potential leading to acceptable particle size variable range ) specification tablet weight and hardness
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
What are the completed actions taken with the recalculated RPN?
Liz
S E V E R I T Y
O C C U R A N C E
D E T E C T I O N
R P N
Equipment 10 6 5 300 retrofitted for NIR probe. Developed NIR method. Initiated routine NIR measuremen ts.
42
Process/Product FMEA Example S Potential E Causes V E What is the In what ways What is the R What causes process does the Key impact on the I the Key T Step/Input to Step /Input Step/Input Customer Y under go wrong? Requirements go wrong? investiga? tion?
Process Step
Blending
Potential Failure Mode
Potential Failure Effects
O C C U R A N C E
Incomplete •Variable 10 Change in 6 mixing Assay/Sub particle size potent of extramedication granular raw materials •And poor powder flow leading to variable tablet weight and hardness
D E T E What are the C T existing controls and I procedures O (inspection N and test) that prevent either the cause or the Failure Mode?
Current Controls
R Actions Resp. P Recommend N
Actions Taken
ed
What are the actions for reducing the occurrence of the cause, or improving detection?
What are the completed actions taken with the recalculated RPN?
S E V E R I T Y
O C C U R A N C E
D R E P T N E C T I O N
One Year LATER
Measure 6 360 PAT to Eva NIR 10 1 3 30 particle size monitor measurement of every blending and of N lots excipient statistics to completed and (but correlate to statistical unknown particle size correlation what a non for potential complete and acceptable specification particle range is) specification set.
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
43
Process/Product FMEA Example Process Step
Potential Potential Failure Mode Failure Effects
What is the In what ways What is the process Step does the Key impact on /Input under Step/Input go the investigawrong? Customer tion? Requirement s?
Compression
S Potential E Causes V E R I What causes T the Key Y Step/Input to go wrong?
O Current Controls D C E C T U E R C A T What are the N I existing controls C O and procedures E N (inspection and test) that prevent either the cause or the Failure Mode?
R P N
Actions Recomme nded What are the actions for reducing the occurrence of the cause, or improving detection?
UnderTablet with 7 Incorrect 3 Operator 1 entry 2 42 Continue compression poor press and Operator setting Training (too low appearance Operator 2 (patient may force check produces not take) Equipment 4 Fill Depth Gauge 8 224 Monthly weak tablet) Annual calibration malfunction Calibration and press maintenan ce variable die fill due to poor powder flow
6
Visual inspection of powder flow during tableting
7
Resp.
294
Monitor/control prior step: blending
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
S E V E R I What are the T completed Y actions taken with the recalculated RPN?
Actions Taken
O C C U R A N C E
D R E P T N E C T I O N
Ron
On going
7 3 2 42
Ian
On going every week
1 7 49
See blending
Control via NIR (PAT)
1 3 21 44
Resulting Control Strategies (subset) •
For blending operation: – PAT to monitor extent of mixing (note: mixing time and revolutions may be flexible); rational particle size specification for extragranular excipients rejects problematic lots – Product CQAs: dissolution, assay and content uniformity
•
For tableting operation: – PAT to monitor mixing; frequent equipment calibration and maintenance and operator training – Product CQAs: tablet dissolution and appearance and hardness – Raw material CQAs: particle size of extragranular excipients
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
45
Conclusion •
Quality Risk Management supports a scientific and practical approach to decision making
•
It is meant to provide documented, transparent and reproducible methods to manage risk using available knowledge to – Identify risk and causes – Assess the probability, severity and, sometimes, detectability of the risk – Prioritize risks • In R&D: development plans to elucidate CQAs and define Design Spaces • In Manufacturing: plan and implement strategy to control variability important to product performance
– Communicate with internal and external stakeholders using common framework/language – Monitor and review risk as new knowledge and experience emerges46
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
Final Remarks •
Quality Risk Management exercises – Require (and enable) discussion & brain-storming – Require access to relevant information • Internal or external data bases • Expert opinions; literature • Knowledge – First principles, Statistics, Statistical models, Correlations, Experimental data, R&D and manufacturing reports, records, trendanalyses – Exhaustive cause and effect review – Require patience and disciplined execution – Require a facilitator and/or team leader – Take lots of time! – Success depends on • Effective discussion and communication • Implementation of response actions • Monitoring and review of result of actions • If not, it is just another piece of paper
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
47
Acknowledgments • • • • • • • • • • • •
Eric Ahuja Celia Cruz Brett Duersch Freddy Martinez Guzman John Lepore Mark Mowery Adam Procopio Sharon Rogin Larry Rosen Iván Santos Gert Thurau Sindusfarma, especially, Jair Calixto
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
48
Back Ups
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
Probability X Impact Decision ToolAnother Example
Probability
Very High 90% High 70% Moderate 50% Low 30% Very Low 10% Low
one level (E.g., Yellow to Red)
10% increase in cost or time; minor scope or quality effects
Moderate
20% increase in cost or time; minor scope or quality effects
Impact
Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
High
30% increase in cost or time
Risk Assessment ExampleRisk Analysis/Response • •
The team could then rank the risk and variables by rating the Probability and Impact to obtain a risk number, RPN = P x I, based on prior knowledge The team needs to agree on a scale for P and I, say, 1-5, where 1 is very low and 5 is very high Risk If …
In vitro dissolution slows down
Cause (variable) …due to
Change in API particle size Roll Pressure
Effect (consequence on patient) …then
Probability (of risk due to given variable)
Impact (of effect as per TPP)
4 (low solubility compound) In vivo release slow down (if assume in vitro model is biopredictive)
3 (data base)
RPN (risk #)
20
5 (reduced in vivo exposure)
15
Over compression
3 (data base)
15
Lubrication blend time and speed
1 (prior experimentation)
5
Risk Response
Experimen tal study to understan d the relationshi p between these three variables and dissolution None at the moment
API = Active Pharmaceutical Ingredient = Drug Substance Copyright © 2010 Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Whitehouse Station, N.J., U.S.A. All Rights Reserved.
51
