(at least one area in life where this is actually true)
I get asked some things all the time, and the number one thing is, "Do we really need x samples?" or questions of a similar nature.
My answer is always the same: "No, I am just kidding. I wanted to see the fear in your eyes."
Jokes aside, I understand a business's (=profit-oriented) intention to reduce the samples (=costs). A sample must be produced, and costs for the workforce, raw material, and subsequent testing add up, so the company wants to reduce it as much as possible.
Fine, I get it.
But think of opportunity costs – what will faulty products potentially cost you?
Process Variation: The more samples, the better?
The goal is to understand and secure the manufacturing process's output or the test method's output.
To do so, we want to take a couple of parts (samples) that reflect the population we want to understand.
These thoughts help us reach a well-known conclusion: the more samples we draw from a population, the better we understand it.
However, we want to consider economic factors, so the essential question arises:
How can you minimize the number of samples while still clearly understanding the true process variation?
Standards (e.g., ISO 11607) and guidelines (e.g., GHTF on process validation) in our industry already provide a good framework to answer these questions —but we must use it correctly without shortcuts.
GHTF states, "…many nonconformities are not the result of errors, instead they are the result of excessive variation and off-target processes.".
Understanding (and reducing) process variation becomes the number 1 goal when striving for a stable and consistent process.
But – how do you reduce process variation if you don't even know it exists in the first place?
Where We Go Wrong in Process Variation
Imagine you have a scenario, as shown in the figure below.
Not having enough information, as shown above, means you don't know that certain things cause process variation.
Let's look at an example that might sound familiar to you.
Let’s assume you are performing a process validation; your boss chases you to finish it this week for whatever reason (maybe someone chases him because of, e.g., market pressure).
Now, because you feel pressured, you decide to produce all samples for OQ and PQ with the same batch of raw material and with one operator only – because, come on, we all agree that these influences are negligible, don’t we? 😉
So, you finish on time and your boss is happy, duty done (and don't forget your promotion when you successfully complete this project).
I've may exaggerated that a bit, but unfortunately I've seen situations like these often enough in reality.
People tend to fall into the trap of short-term thinking, e.g., finishing the project quickly, getting a promotion quickly, reducing sample sizes and the number of PQ batches (which is essentially the same as reducing sample sizes), and other things that supposedly bring (short-term) benefits.
By doing that, we ignore the long-term consequences. We forget about WHY we are doing quality management – it should not be for some personal achievements or getting things done asap for your boss, it should be for patient safety.
I feel like many people are not aware that someone will have to carry the consequences for wrong decisions later – ideally it will “just” be a failed audit, worst-case it affects a patient.
Let’s Do It Right: Reduce Process Variation, Not Sample Sizes
If there is one thing you take away from this blog post, it should be this:
Please be thoughtful when reducing the number of samples.
Think long-term. Think process variation. Think quality.
Keep in mind that reducing the number of samples will also reduce the visibility of what is going on in the process.
If you think long-term, the cost of ensuring your quality will be worth it.
So, let's do it right from the start.
The journey towards full compliance with process validation requirements is both a challenge and an opportunity. It's a chance to redefine quality standards, enhance safety, and ultimately, build trust with patients and healthcare providers.
I invite you to explore the intricacies of process validation and other pivotal topics in the medical device industry.
We already addressed the topic of sample sizes and how to rationalize them in other blog posts, for example:
I also invite you to check out our YouTube channel (@SIFoMedical) where you will find answers to common questions, in a short and easy-to-digest video format.
If you have any burning questions, please comment below or email us at office@sifo-medical.com.
Let's navigate the complexities of regulatory compliance together, share knowledge, and elevate our industry standards to new heights.
Author: Simon Föger
コメント