Does software testing tool need to be 62304-compliant?

I work for a medical device engineering company in MA and we are not that familiar with the IEC 62304 process. We are self-teaching; there are many good resources I have to draw on but, since Elsmar Cove bit the dust, I miss having an interactive forum to ask and answer questions. This is my first post on this forum.

We have written a software test tool to test the calls for an API we are developing for a medical market. We are going to follow 62304 for the API but unsure if we need to follow 62304 for writing the test tool. We know that we need to validate the tool, but do we need to follow the entire 62304 process for writing the tool?

Can you refer me to any official document, i.e. a standard or guidance doc that addresses this scenario?

Thank you and I look forward to your response.

 

Simple answer is no, tools do not need to follow 62304, the just need to be recorded what tools were used.

With medical devices it is always good to keep in mind that the standards apply to the "medical device" only, which is the actual product that goes to market. Contrary to popular belief, 62304 does not even apply to prototype versions, just the final version that actually gets released to market (we will see if that statement generates some discussion ).

 

Hi Peter,

Thank you for your response; I certainly hope that you are correct. I do find it difficult to fathom that the development, risk analysis and testing of prototype versions would not be subject to 62304 requirements and constraints.

We are proceeding under the assumption that we can develop the software first using an Agile methodology and then go back and apply 62304 measures afterwards. This would be a huge time-saver for us.

What do you think?

 

The 62304 requirements are only on the final product. You should have some evidence that the tool works as intended and provides repeatability.

Please define "prototype" versions. If you are working a prototype in the lab no problem. If you are taking it for an outside trial, I might would be inclined to have some supporting documentation. I lean toward agreeing with Peter that it is only "required" for the final device, but depending on the risk and liability IMHO it would be a good idea to have some backing.

 

Peter, I don't disagree with the statement you made above; however, I think it's a risky way of thinking. Prototypes often morph into product. If you don't capture the design artifacts along the way, it's going to be difficult to create them when you move under design controls. Some forward-thinking and judicious application is not a bad idea even with prototypes.

My first question is always: what in the standard is not appropriate for generally good software development efforts? (Recognizing, of course, that the standard IS a bit dated, seemingly more waterfall-y than more modern, agile practices.)

Back to the OP's question: I agree that it's not required for tools (heck, it's not REQUIRED for product software!). However, if the tool will be validated, it *IS* expected that a defined software life cycle is applied to ensure the tool can remain validated across changes. So while full-blown 62304 (even class A) is likely overkill, there should be effort in the organization to ensure these tools are properly developed, maintained, validated, etc.