As always people will answer to me that it depends. Process Validation or Verification depends on the process, on parameters or on the product. So how to choose?
In this article I will give you all the keys for you to define if you should Validate your process or if you should consider verification.
The definitions I will write below are mainly coming from the GHTF Process Validation Guidance.
Medical Device School episode
In this episode of Medical Device School, Monir El Azzouzi and Stefan Bolleininger are helping you define the difference between a Process Validation and a Process Verification.
We hope you’ll like it. Don’t forget to subscribe to get notified when another episode is released.
Subscribe to my Youtube Channel to get notified of my next videos
Table of content
- What means Process Validation?
- What means Process Verification?
- Process Validation - FDA requirements
- Decision Tree
- Should I validate even if verification is OK?
- Podcast Episode 81
- Process Validation examples
- Example of processes covered by verification
- 7 steps to succeed on your Project
- Why should I perform a revalidation?
- What is an IQ OQ PQ?
- Question to answer during a Validation Project
- Which Statistical Method?
- Wrap up
What means Process Validation?
The definition is:
Process Validation: Establishing by objective evidence that a process consistently produces a result or product meeting its predetermined requirements.
To make it Easy, this is a method to prove that your equipment produces always good parts. But you’ll see that it can be more complex than that.
What means Process Verification?
On the other side, we have Process Verification.
It’s definition is:
Process Verification: Confirmation by examination and provision of objective evidence that the specified requirements have been fulfilled.
Again, to make it simple, this is the act to measure each product to confirm that they are ok.
Process Validation - FDA requirements
If you are looking for the FDA requirement for Process validation, you should look at the FDA QSR 21 CFR part 820.75.
Below is a copy of the content:
Sec. 820.75 Process validation.
(a) Where the results of a process cannot be fully verified by subsequent inspection and test, the process shall be validated with a high degree of assurance and approved according to established procedures. The validation activities and results, including the date and signature of the individual(s) approving the validation and where appropriate the major equipment validated, shall be documented.
(b) Each manufacturer shall establish and maintain procedures for monitoring and control of process parameters for validated processes to ensure that the specified requirements continue to be met.
(1) Each manufacturer shall ensure that validated processes are performed by qualified individual(s).
(2) For validated processes, the monitoring and control methods and data, the date performed, and, where appropriate, the individual(s) performing the process or the major equipment used shall be documented.
(c) When changes or process deviations occur, the manufacturer shall review and evaluate the process and perform revalidation where appropriate. These activities shall be documented.
This provides you with a recipe of what the FDA wants to see when they will be auditing you.
We will talk about some of those requirements in the next chapters.
The FDA also issued guidance on Process Validation that can provide you additional support to perform your own project. So don´t hesitate to look at it.
Share this article
Decision Tree
Let’s review that through a decision tree so you can know if your process should go through the validation route or the verification route.
On the graph below, we can start on the box A.
A. Is process output verifiable?
This means that you can check the result of your process. You can have a measure of the output and then confirm that this meets the product requirements.
If you answer Yes to that then you should go to B where is asks you
B. Is verification sufficient and cost effective?
You’ll then need to provide some justification to say that this is sufficient. The final result is that you can use Verification for your process.
C. Verify & Control the Process
But if the answer is “No” for A or B, then you need to decide if you
E. Redesign your product
and finally go back to A to test your new outputs or you decide to
D. Validate you process
as the outputs cannot be verified.
Should I validate even if verification is OK?
This is strange for people to see an aspect of cost on a Quality topic. But as a Medical Device manufacturer should be a sustainable business, the cost of all this process should be taken into account to make some decision.
But we should always put safety first. So you cannot justify to reduce safety because of high cost of the process.
Mathematics
When you will start to put some numbers together, you will maybe define that the verification process is cost effective. So why not just going through that. But if you are really thinking ahead, this process can be costing you a lot even if for you this is ok. There are the companies looking long term and the companies thinking short term.
If you invest in validation, you can for example reduce the volume of work for employees on inspection of each part. As you have not validated your process, you need to check one by one each critical features as you have no documentation proving that your process can produce good parts 100% of the time.
Draw me a Process Validation
Below is a figure summarizing this. First year this will cost you a lot more than if you start only with verification. But then the cost is reduced. Why?
Because:
- The process is proved to be stable so you produce less scrap.
- The employees control less so there is no need to hire more inspectors.
Volume or no Volume?
One thing to consider is also the volume you will sell. If your forecast is really low for one product, maybe it’s not worth performing a validation. Because to validate your process your will need to produce some parts to have some good statistical results.
Imagine now that to validate your process you need to produce 140 units of your product. But your yearly forecast is 20. So this means that for your validation you will make around 7 years of products (7 times 20 parts/year = 140). And some of them will be scraped.
So this should be carefully analyzed before you decide to continue with verification or switch to validation.
Podcast Episode 81
Within this episode, Adnan Ashfaq from Pharmi-med Ltd. is offering the Validation Toolkit. Download it from the button below.
Process Validation examples
To help you understand the concept of validation, I did choose some processes where you have no choice, you need to validate them. Why?
Because if you try to verify these processes, you finally will need to destroy all your production. Which economically is not really interesting for you.
When you produce a product that should remain sterile for a long period of time, the only way to verify if this is working is by opening the bag and check if the product inside is confirmed to be sterile.
But if you do that with all your products then you cannot sell them. So for this case, a validation is mandatory. You need to validate your process to prove that it will always sterilize your product correctly.
Same for this process. You cannot check if your room is clean at each moment. But you can prove by making a validation that all the processes in place permit to maintain the cleanliness of the room.
This is the same as for sterile process. You cannot test each product to confirm that the aseptic filling process is continuously working. So a validation is the best way to go.
If you should destroy your product to verify that it meets the requirements, then you should think Validation.
Example of processes covered by verification
So let’s now focus on verification. I will describe below some processes where verification can be done. This list is not exhaustive so please don’t hesitate to suggest me more on the comments.
First, when you hear manual, this can be a problem when we attach to it validation. Because the validation of a human is really a big problem we have. Don’t say it’s impossible but can raise a lot of questions. So for this kind of manual process, you can measure the cutting size and see if this meets the product requirements
This can be covered by verification, but I can say that it can also go to Validation. You just need to answer more questions that we will talk about later,
Same as before, this can be verified but you can also mix it with some validation. Still need to get more information to be sure.
7 steps to succeed on your Project
To start such a project, I would say that there are some critical steps to follow. I will share with you the X steps to get a project finished.
Step 1 - Team Building
Building a team should be your first action. This is not a 1 man job. For sure, you need to involve some subject matter expert on the process you are trying to validate.
I can be the best Validation Engineer, if I don’t understand the process I will go nowhere.
So a Project leader, a Validation Engineer and subject matter experts (depends on the complexity of the process). This can be a minimum of 3 people involved.
Step 2 - Plan your journey
Then you should plan the approach and define the requirements for your process or product. What is exactly your goal and how you are planning to reach it?
As said before, you will be able to do nothing if you don’t know how the process is working. So a meeting with some SMEs to describe to you the process would be great. Even better if you can see the process working live and ask your questions immediately.
You will then understand if this validation will need more or less time to be finalized,
Step 3 - Collect your parameters
You should consider your process as a lottery. You have maybe 4 or 5 parameters to play with so you can reach the right output and win the lottery. But for that you should understand those parameters and then define what are the impacts of it on your product.
Step 4 - Navigate the tree
Finally, with the decision tree described earlier, you should define if you go for a verification or validation. As mentioned before, there are multiple criteria to look at to take good decision.
If you go for verification, then you can stop here and implement a verification process within your production. If you are the happy winner of a validation project, I would recommend you to continue your marvelous journey.
Step 5 - Create a Masterpiece
When you start a validation, you first need to have a Master Validation Plan which will be the bible for your process validation. Inside this Master Validation Plan, you will describe the process and define your validation strategy.
In terms of tools, you should select the ones you will use for your validation. We will go more deeply after that on this topic.
Step 6 - Validation Choreography
And now that you are ready, you will go through a tango with your documentation. You will create a validation protocol and then a report and again a protocol and again a report and a protocol and a report…. I hope you like this choreography.
The term that we are using for this dance is the process validation IQ OQ PQ. Which means:
- Installation Qualification
- Operational Qualification
- Performance Qualification
I’ll tell you more about it don’t worry.
So you should perform the process validation IQ OQ PQ and create all the documents related to that. This should provide you with a positive result related to the validation. If not then you’ll have to restart the dance.
Step 7 - Monitoring
But even if your process is validated, you still need to monitor it. But less often than if it was not validated. You then need to define how will the monitoring be executed.
For example, for product cleanliness, every 3 months we will send some samples to a laboratory to confirm that our cleaning equipment still meets the requirement.
As soon as those criteria will be defined, I recommend you to write that on a procedure that should be followed by your employees.
Why should I perform a revalidation?
If it is already validated, why should I revalidate?
This depends on many things but the first one is that you have changed something on the process. In that case, you need to assess if this change is affecting your initial validation. If this is the case, there is no choice, you need a revalidation. You should define if an OQ and/or PQ need to be redone.
The other thing is the fact that you see a deviation on your monitoring. If this is the case, then something has changed and you need to review that. Maybe a parameter was not taken into account on your initial validation and then this is compromising it. So a revalidation will be needed unless you can justify it.
GAP ASSESSMENT TOOL
You need a ready to use tool to make a Gap Assessment of the MDR, MDSAP, ISO 13485. Click on the link below to see the offer
What is an IQ OQ PQ?
These terms are famous on the Medical Device Industry as this is mainly the process used for medical device validations. If you are knowing this vocabulary, this is something that can be appreciated during an interview. But if you know the complete method, this will be your key to enter the door. So let’s define the process validation IQ OQ PQ one by one.
This is the first step for a new validation. The definition is:
Installation Qualification (IQ): establishing by objective evidence that all key aspects of the process equipment and ancillary system installation adhere to the manufacturer’s approved specification and that the recommendation of the supplier of the equipment are suitably considered.
But as I know you prefer me to summarize it I would say that this step answering the question: Is my process correctly installed (Follow vendors requirements)
If the vendors says that you need a certain amount of electrical power, you need to verify if you have it. If the machine should work under a certain temperature, you need to check that there are temperature sensor to alert when this is not compliant anymore…
Again, what is the definition you will find on the books?
Operational Qualification (OQ): establishing by objective evidence process control limits and action levels which result in product that meets all predetermined requirements.
So, if I may try to summarize this, I would say that this during this phase you would ask “What are the limits of my machine?” or also “If I turn this button, until when my product will remain compliant?”. By doing this you are defining the worst case conditions.
If your products are good at the worst case conditions, then they should logically be good inside this range. So this is the objective of this phase.
So after that you should have a range for each of the parameters. If we continue with the cleaning process example, you can have a temperature limit for the cleaning bath. Maybe the detergent we will put inside will be efficient from this temperature level to this one. And then you need to test those limits.
So you’ll manufacture some products with the upper limit and some others with the lower limit. Maybe if you have too many parameters, you’ll need a DoE or Design of Experiment to find the right parameters.
Ok, let’s read the definition.
Performance Qualification (PQ): Establishing by objective evidence that the process, under anticipated conditions, consistently produces a product which meets all predetermined requirements.
So during this phase, you will answer the question “Can my process produce consistently good parts?”. So to do that you need to run many batches and test all of them.
Again for a cleaning process, we will clean 3 lots of products (Basket fully loaded – Worst case) and then send them to a laboratory to test them for our criteria.
If all the products are ok, then your process is considered efficient. Congratulation.
Question to answer during a Validation Project
When you are starting your project, there are a few questions that you should answer related to your process to help you. I list them below:
- What should you verify/measure?
- How should you do to verify/measure?
- How many should you verify/measure? i.e. Statistical significance
- When should you verify/measure?
- What are the acceptance/rejection criteria?
- What are the required documentation?
If your validation is answering those questions, then you are good to go. This is really important as this will define if you have covered all the important aspects of the project.
Let´s deep-dive on each of them so you can really understand with some examples of what information we are looking for.
What should you verify/measure?
To answer this question you need to understand 2 things:
- How your process is working
- Which output is important
For example, if we continue with the cleanliness example, we need to define what are the criteria to verify/measure and this can be for example
- a certain level of cleanliness on your products,
- and/or a certain quality of the water
- and/or a certain temperature of the bath
These things are the parameters that would be critical for your process.
You can also find this normally on your risk assessment. If you follow ISO 14971, you can use the risk you identified as an input for choosing the element to verify or measure. If you risk assessment identifies some elements that will not be taken into account on your validation, we can then imagine that an auditor can raise a Non-conformity unless you have a good justification.
How should you do to verify/measure?
Now that you know the criteria, you need to define your method to measure them. A calibrated thermometer for the temperature, or a sensor that measures water quality…
This is your choice and you’ll then need to explain why this is the best method for you.
How many should you verify/measure?
It is clear that the higher the number, the more accurate results you will get and this will be then easy to defend. So some statistics should be involved here to define how many samples, or tests or… you should execute to prove that your process is correct.
For cleaning, this can be the number of parts to send to laboratories but also the tests that should be performed. Provide some statistical proof and then you’ll be on the safe site. On the other side, if you just decide to take a random number, this will not be accepted by an auditor so be careful.
AQL for sampling
You can also define for certain processes, a sampling method to measure the parts if you decided to validate this type of process. And for that you can use a sampling method called AQL (Acceptance Quality Limit). This is a statistical methodology to help you define how much part you should inspect for a certain lot size. Check on this website to find an AQL calculator.
We need also to mention a vocabulary that is important. It´s “Critical to Quality”. What does it mean?
If one of your feature is critical to quality, it means that without this feature, the product cannot work properly. So the inspection of this feature should be well executed.
But some features are easily measurable without destroying the product and others are not. And here the concept of Verification and Validation is important.
If to verify your feature, you need to destroy the product, then you need to go for a validation. The usage of AQL for sampling can then be chosen.
But if your feature can be measured you have the choice to measure it at 100% which involved a lot of resources or validate your process.
This again depends on your volume.
When should you verify/measure?
This seems to be a strange question but if we take for example your full process for cleanliness. Will you measure before cleaning, immediately after cleaning or when the parts are on their final configuration for your customers…
If after cleaning your parts, they will be potentially contaminated by next processes, then it makes no sense to measure them immediately after cleaning. You should consider that your cleanliness level is the result for your customer not only for the process. But you can still make measures just after your cleaning process to see if this achieves the right objective and then try to reduce further contamination.
So when you will draw your process flow, you’ll need to justify why you choose to measure after a certain step.
What are the acceptance/rejection criteria?
It is clear that if you don’t define clear acceptance/rejection criteria, you cannot really complete a validation. On your protocols, you will need to define what are the results you expect and when this process will be considered valid. Measurable criteria are ideal as this can give you a clear limit.
On cleanliness, we say that the product should be without any discoloration or stains… But each person can interpret it in a different way. Someone will say that there is no discoloration when another one will say that this discoloration is due to the machining process and not the cleaning when another will not see the discoloration…
So when you have measurable criteria this will help you to avoid all those discussions.
What are the required documentation?
To complete your validation, you’ll need to define also the documents that you need. As mentioned, there are the validation documents needed for IQ, OQ, PQ, but there are also the laboratory reports, or the supplier requirements…
A list of all this should be defined in a procedure or directly on the protocols.
I publish Videos of my Podcast or also some Videos on specific Medical Device TOPICS.
Subscribe below
Which Statistical Method?
To perform a process validation, the team can decide to use many statistical tools. Here is a list below.
For each of those methods, I also provided a link to an external resource to help you understand it.
- Control Charts
- Capability Study
- Design of Experiments
- Tolerance Analysis
- Robust design methods
- FMEA Failure Modes and Effects Analysis
- Sampling plans (AQL for example)
- Mistake proofing
To prove that your process is validated, you will need to use some of those methods as they are factual.
The result of your tests should be justified by your statistic method.
Wrap Up
As you can see this is not an easy process to handle, this is why it’s sometimes a big discussion at some companies. Should we validate or verify our processes?
As mentioned, there are some processes were validation is mandatory. No choice. And others where you can say that it’s worth the money to validate.
But what is critical, is to not let this on the hand of only one person as we really need to have information coming from the SMEs regarding the process.
As you can see this is not an easy process to handle, this is why it’s sometimes a big discussion at some companies. Should we validate or verify our processes?
As mentioned, there are some processes were validation is mandatory. No choice. And others where you can say that it’s worth the money to validate.
But what is critical, is to not let this on the hand of only one person as we really need to have information coming from the SMEs regarding the process.
Medical Device expert. Monir founded Easy Medical Device to help Medical Device companies to place compliant products on the market. He proposes his consulting services so don’t hesitate to contact him at info@easymedicaldevice.com or +41799036836
My objective is to share my knowledge and experience with the community of people working in the Medical Device field.
