This is clearly a gray area from my perspective and I am looking for some input:

Since IDE(Investigational Device Exemption) approvals exempt a company from all CFR but 820.30 Design Controls and GDP, does this give a company freedom not to validate a process even if it is manual and you are doing 100% verification of a build but sampling on the testing(using the device is destructive.)??

This is clearly a gray area from my perspective and I am looking for some input:

Since IDE(Investigational Device Exemption) approvals exempt a company from all CFR but 820.30 Design Controls and GDP, does this give a company freedom not to validate a process even if it is manual and you are doing 100% verification of a build but sampling on the testing(using the device is destructive.)??

Hi jscholen,

For a particular process to be validated / verified, I always would like to refer the GHTF document - Quality Management Systems - Process Validation Guidance (http://www.ghtf.org/documents/sg3/sg3_fd_n99-10_edition2.pdf)

You're right in that you are somewhat exempt from manufacturing controls, I believe validation is determined on a case-by-case basis. For example, we are developing a device that will go out as IDE initially and we know (I believe through discussions with the FDA) that the sterilization process will have to be validated. We will also have to show biologic compatibility. So clearly, FDA has authority to exercise some judgment in this area. So I think you'd be best served by discussing what FDA will expect with your FDA rep.

This is clearly a gray area from my perspective and I am looking for some input:

Since IDE(Investigational Device Exemption) approvals exempt a company from all CFR but 820.30 Design Controls and GDP, does this give a company freedom not to validate a process even if it is manual and you are doing 100% verification of a build but sampling on the testing(using the device is destructive.)??

Most Companies hope to support a PMA submission with data from the IDE. As part of the submission, your manufacturing processes will be audited. In my experience, having the processes developed and validated as soon as possible can help to avoid filing a PMA supplement because of a process change.

With your example, the fact that the process is manual is not relevant. If the output of the process is 100% verifiable as you say, then validation is not required.

I'm not sure what you mean by "sampling on the testing". Are you taking a sample to test?

Sounds like you may need to define your process outputs.

I'm not sure what you mean by "sampling on the testing".

820.75(a) stated that if "the process is not fully verified by subsequent inspection and test"...

since we are not able to do 100% testing since the testing is destructive, we sample test, but the device is 100% verified for correct assembly.

This is my latest view...since the device is investigational and exempt from validations except as noted by yodon, the device must be adequatley tested/inspected to meet user expectations, not specifications.....

what does everyone think of that statement?

820.75(a) stated that if "the process is not fully verified by subsequent inspection and test"...

since we are not able to do 100% testing since the testing is destructive, we sample test, but the device is 100% verified for correct assembly.

This is my latest view...since the device is investigational and exempt from validations except as noted by yodon, the device must be adequatley tested/inspected to meet user expectations, not specifications.....

what does everyone think of that statement?

It sounds as if you've made up your mind and are looking for others to support your conclusion.

I don't agree with the approach as it ignores the long term implications.

I suggest you document your conclusion and be prepared to defend it with FDA.

I'm too new for support, just looking for advice...

Since you don't agree with the statement, why does the FDA give companies IDE approval for a device with exemption from all but 820.30? I am assuming that a device specification is not fully verifiable in this situation, but the user expectations is....

I would argue that some validation is part of design control and risk mitigation which is also part of design control. I would not be quick to dismiss validation.

Regards,

John