John Daley, quality assurance manager at AngioDynamics, Inc. (Queensbury, NY), explains how to fit verification and validation processes into a product development schedule.
My company understands the importance of performing verification and validation procedures, but how can we do it without slowing down our product development schedules?
The easiest way to perform design verification and validation activities in a timely manner is to remember that they can and should be performed continuously throughout product development.
Before describing how to schedule these activities, however, it's important to describe the differences between them. Design verification is the method used to determine that the device meets or exceeds the manufacturer's predetermined specifications. Design validation is the method used to determine that the device meets the user's needs.
Suppose you want to market a reusable tongue depressor. Before the project begins, you develop various design specifications. A few might be that the device must withstand 20 EtO sterilization cycles, weigh between 3 and 5 oz, and be between 5 and 5½ in. long. These are all internal specifications that can be verified throughout the design phase. Any engineer or technician can subject representative samples to 20 sterilization cycles, weigh them on a scale, and measure them with a ruler. If the results are acceptable, the design has been successfully verified.
Now let's say it's time to validate the design. You can do this by having samples reviewed by a cross section of the people who will eventually be using the finished devices. In this case, pediatricians, nurses, and general practitioners would be good choices. If these groups evaluate the device and feel that it would make a good reusable tongue depressor, then all you have to do is document this review. The design is now validated. Only in this way can you be sure that you are putting devices on the market that users really want. In fact, the tongue depressor mentioned above would probably pass verification but fail validation. Why? Most pediatricians, nurses, and general practitioners don't care if a device can withstand 20 EtO cycles. They want one that can withstand hundreds of autoclavings.
How do you fit these activities into your production schedule? Verification is nothing new to device manufacturers, so it should not affect most schedules. These procedures should be already built into the product development process.
The trick is to incorporate FDA's new validation requirements into your production schedule. To do this, you need to evaluate each aspect of the design as it becomes available. For example, if the contour of a device's handpiece is important, then give a model to end-users to evaluate. If they like it, then that portion of the design is validated. Now you can go to the next step of development. As another example, what if you want to find out if the luminescence of a cathode ray tube (CRT) display is adequate for use in a radiology suite? Mock up a CRT to the proposed luminescence and have it evaluated in a radiology suite. You have then validated the display.
The quality system regulation says that "Design validation shall be performed under defined operating conditions on initial production units, lots, or batches, or their equivalents." Ideally, you'll perform design validation on initial production lots. If you don't, you must be able to prove that the prototypes you're validating are the same as the devices you'll manufacture in the initial production lot.
To prove this, you must develop a system that ties the testing done at the various phases of the production process to the characteristics of the final device. This system doesn't need to be complicated. In general, you simply need to create a matrix that documents that design specifications are being adequately tested. Since those specifications have, one hopes, already been shown to meet the end-users' needs, the link is now complete. If your company follows good testing practice and issues protocol numbers to all relevant testing, the matrix can be as simple as a three-column chart with the following headings: "User Requirement," "Specification," and "Supportive Testing." In the tongue depressor example, the text under the columns might be "Must be lightweight," "50 g ± 2 g," and "Pr-001," respectively. As long as you have a system like this in place, showing equivalence should be a simple matter.