Hi Roberto,

After doing some research, I was able to find that the FDA does not demand specific tests, but establishes guidelines for developing design controls, design validation, and design verification. Looking through the site briefly it seems that much of the information provided is definitions and guidelines to allow for an engineer to develop the appropiate tests for their project.

The information I found can be accessed at: http://www.fda.gov/RegulatoryInformation/Guidances/ucm070627.htm#_Toc382720788

-Andrew Nashed

If quality by design is used in the development of the product, a simple way to come up with tests for verification/validation is to just use characterization techniques that have been used to measure the critical quality attributes all through development. As long as the testing technique is reproducible, scientifically sound, and not complicated to do repeatedly (for quality control purposes), the you should have a good test.

I actually just wrote and performed both verification procedures as well as validations for the current company I work at. FDA does not mandate a specific way verifications are performed, they just mandate in their 510k or PMA submissions that in verifications, design outputs meet the design inputs. Verification must verify each requirement and pass in order for the requirement to be transferred. The FDA does not view any of the verifications, they usually stay in house in the company as documentation that provides evidence that yes the design input and output was successfully verified. However, in a FDA audit before the transfer and manufacturing phase they would certainly vet through procedures, documentations, and verifications to note if the company is ready or not to being manufacturing a device.

Hello Roberto,

I have done some research like my classmates, it was interesting to read that the FDA points out to verification not only in terms of tests. They list guidelines such as analysis and inspections. I agree with my classmates that they don't have a specific list, but they provide major explanation and overall explanation of how to conduct the verification activities in their design control guidance. The link attached has the complete guide specific for medical devices, and it explains in details all the needed information to conduct a successful verification process.

http://www.fda.gov/downloads/MedicalDevices/.../ucm070642.pdf

- Lamiaa

As others have stated, the FDA does not give mandatory tests that need to be done. However, they do enforce that design verification and validation very very strictly (as well as process/software etc.). Here, verification is done to show that the product meets its design specification (it is what its supposed to be) whereas validation is done to show that the product meets the customer needs (is performs its intended use). As a guideline, there are ISO and ASTM standardized tests (e.g. Instron pull tests) but ultimately, engineers just design a test to show prove V&V. For example, the design specification of a balloon is that it can fully inflate to 100cc. To test this, you can put the balloon in a known volume of water, inflate it, and measure the water displacement. Furthermore, they then have to prove that the test is reliable as part of the V&V process (using Gage R&R and other statistics). So they will perform the test across a determined sample size, record the results for each sample, and hopefully have statistically significant test results that prove the the test can be used by any operator with any any sample.

To your point, the ASTM and ISO standard test methods are very relevant to this question. They also can be modified for different applications after purchasing the test. The engineer does make up the test but they can use standards as guidelines and there are widely accepted statistical methods to analyse results. Ultimately the decision comes to the project team to mark the tests valid for the application and that the results do prove the device was successful.

Verification and Validation go hand in hand. They cover every phase of product development and it is always considered from the very beginning.
Material performance, Package quality tests, Bio compatibility of materials, Sterilization tests are the most commonly performed tests. After the tests are conducted, inspection and analyses is carried on.
Verification process requires only one unit testing in each category where as validation is conducted to mainly meet design specification of the product.

As many have mentioned before verification and validation often come hand in hand, and they sometimes are used interchangeable, but they are different. Combined verification and validation of engineering designs within the product lifecycle directly influence production performance, define product functionality, customer perception of the product, and robustness of the process. The essential framework for design verification and validation, which is lifecycle based consist of four stages: preliminary design stage that sets requirements, digital design domain, product and process development and prototyping phase, and lastly design of production system and a method for the realization of complex processes and products. I attached an article that goes into detail regarding the design verification and validation in product lifecycle

Chris
http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.175.5629

I agree with most of that the FDA doesn't demand specific test and to me this would be too much of an overhead and additional unwanted burden for them. I have some experience outside of BME with verification and validation, however, there is not much difference here. For each input requirement a test has to be created, and how to create the test will come from different source. One way is from either previous experience, or full understanding of the product from reading design doc and how to test it, or talking to the folks who wrote the requirements or the team that implemented the requirements. There is also understanding the standards and using it as a guideline to help in creating the test, and there are several other ways to approach this. In some cases you may need more than one test to verify a requirement.
So to your question, the test will depend on many factors including some that you mentioned like, biology, chemistry and physics. There should be a requirement specified to verify all aspect of these areas. In terms of validation the tests are more high level and one test can cover multiple functionality.

Adrian

As we have learned and many have already mentioned, verification tests come about from the design specifications in the DSD. Each specification can have a test linked to it and will be explored during verification. However, not all specifications have to undergo design verification. Based on my experience, in the earlier stages, the quality representative/team will approve which inputs are "Critical to Quality" (CTQ). These inputs and specs are explored in the follow verification tests. The FDA does not choose these tests or suggest them. The design team develops the tests and the appropriate heads (like the Quality Engineers) will approve or give suggestions.

The verification and validation reports provide evidence for a company that their product inputs will directly lead to the product output. In order to confirm this is true you must design a test to validate the claims. These tests that Dr. Simon mentioned are going to differ according to different projects. As our classmates have discovered through their research, the FDA does not have any regulation to how these tests are conducted. The FDA leaves you a guideline as to what these tests must accomplish. If these tests do not accomplish these guidelines they will not suffice. Therefore the design verification tests must ensure that all design outputs meet the design input requirements. The design verification will include identification of the design, methods, the date, and individuals performing the verification. This information will be documented in the Design History File.

As my classmates have mentioned the FDA does not demand specific tests for design verification or design validation. The FDA ensures good quality assurance practices with the 21 CFR Part 820.30 that Dr.Simon mentioned in his lecture. This regulation does not describe the practices that must be used but establishes the framework that manufacturers must use when developing and implementing design controls. The framework gives manufacturers the guidance and flexibility to design tests that fit their unique product while also meeting the regulations.

Hey Roberto,

I agree with what everyone on this thread has said so far. However, I would also like to add that when "making up tests" it is often beneficial to read up on publications that have done similar clinical or preclinical testing. Although the product being tested might be novel, there is always a good chance you may be able to find something very similar. From there, you should be able to get a good idea on how you will dictate an experiment to test a product. This is not only done in industry, but also Academia. Much research is done and ideas are borrowed quite frequently. This is because it is easier to follow a protocol that you know will work rather than start from scratch. Hope this answers your question!

-Zohour

Roberto this is an excellent topic of discussion and as you will find in the field, often a topic of contention. The FDA provides guidelines on how tests should be executed. Take a look at the recent changes to the guidelines for LAL testing and you will see that the FDA does not claim specifics. Another example would be the guidelines to establish cleanroom monitoring. Fortunately there are ISO standards that come and serve the need of the company. By keeping abreast with the ISO/ASME standard, the company is able to mitigate risk and have some guidance by the subject matter expert in that field. Once enough literature is collected, the skilled engineer should initialize a design of experiments or protocol based on the information they have that pertains to their product.