Ensuring that a product’s design and materials will provide the required long-term performance is one of the most difficult challenges to overcome during product development. In most cases, the performance can only be assessed after widespread and extended use. While this issue is relevant to the design of all products, it is absolutely critical to the medical device market, where products are subject to lengthy and expensive approval processes and product liability can be substantial.

According to Dr. Eric Scribben, principal research scientist at Battelle, the development schedule may not contain a sufficient amount of time to conduct a complete assessment of long-term performance in all practical use environments. While the motivation and rationale for omitting such a study are understandable, consequences can be serious in cases resulting in product recalls.

Effects and risks of performance testing

In practice, engineers leverage several techniques to address the effects of environmental or use stresses that may reduce product life. One approach is to introduce safety factors into design parameters. These safety factors may be based on failure rates predicted from parts stress (or count) modeling with such limited data as the mean time between failures (MTBF) or mean time to fail (MTTF). Unfortunately, these parameters are often obtained through material or component level tests in use conditions much different than what the product may experience.

An alternative approach is to assess long-term performance by conducting a physics of failure analysis, which considers performance degradation from a more mechanistic point of view. Obtaining data for either approach can be time consuming. Accelerated tests are performed to speed up this process by increasing the severity of stresses with the intention of producing a proportional increase in the failure rate.

There are numerous risks associated with accelerated testing. Perhaps the most significant is ensuring that the applied degrading stresses produce an increase in the rate of degradation without fundamentally changing the nature of failure. Scribben describes an approach Battelle developed that provides for the identification of the primary failure mechanism and the stresses that influence the rate at which the failure proceeds. This information is used to determine if the measured response to accelerated conditions is, in fact, consistent with what occurs in real time. Verification of the scaling relationship between the magnitude or severity of the degrading stress and the degradation kinetics is of tremendous value in establishing confidence in the predicted life values. It also enables extension of the approach to a wide range of stresses and products.

Standard methods (e.g., ASTM D 3045 - Heat Aging of Plastics Without Load, ASTM F 1980 - Accelerated Aging of Sterile Barrier Systems for Medical Devices) are clear about limitations in scope and use. In the case of ASTM D 3045, the standard considers only forced convection, hot air environments and specifies that the service life obtained should be considered the “maximum expected.” ASTM F 1980 is applicable only to the ability for a sterile barrier system to withstand heat aging.

“We have been approached by numerous organizations with concerns that extend beyond heat aging and the performance of sterile packaging,” says Scribben, “so we have had to consider more versatile approaches.” A more versatile approach, he says, should include the following steps:

• Conduct a thorough review and inspection of the item, including all known aspects that may influence life performance (e.g., materials, construction process, packaging, environment, duty cycle). This activity is often performed in conjunction with a review of the open technical literature to support formulation of a hypothesis for the failure mechanism with identification of the degrading stresses.
• Conduct a series of relatively short-term scoping experiments to supplement the gaps in the technical literature. This includes preliminary validation of the failure mechanism hypothesis, the significance of combined stresses, the magnitude of stresses that may be applied, and the approximate rate of degradation that may be observed.
• Complete and conduct an experimental design for exposure and testing once the design parameters are bound.
• Analyze the data and formulate performance models to make predictions as needed.

An example of the prediction for a recent product that was found to have elongational properties that were sensitive to the combination of temperature and humidity is shown in the figure below.

Prediction of the current performance of an item after exposure to a known environment for approximately two years followed by a hypothetical environment for the remaining portion of the designated 15-year life.

Scribben recognizes that all approaches have limitations and warns of a more practical issue. Aside from a few minor technical limitations, such as products that exhibit a random or non-monotonic response, this approach can be resource intensive and, in its entirety, may not be applicable to every product. It is better suited for situations involving long time scales, limited product availability, or cost constraints associated with testing in great quantities. However, the scope of each activity may be customized to produce a program that is amenable to a much wider variety of products. The balance between confidence and risk lies at the core of characterizing life performance.

Battelle has utilized this versatile approach many times to predict changes in material performance,” says Scribben. “We have made life predictions based on changes in material properties for a wide variety of environments and products that are intended to function for periods ranging from several years to more than 20 years. We have been conducting this type of analysis for our customers for nearly 30 years and continue to be pleased with the results.”