Answers are grounded in this paper and cite their sources.
Development of medical devices is highly regulated [1–3]. The U.S. Food and Drug Administration (FDA) provides a guidance stating that reliability should be part of the device performance requirements [4]. Yet the role of reliability testing and analysis varies significantly from one manufacturer to another. In addition, the FDA requires tracking of all medical device failures so ignoring reliability can lead to significant risks for the medical device manufacturer. However, traditional reliability testing can be expensive for complex electromechanical medical devices. With typical volumes of less than 1000 devices a year and large device costs, it is not feasible to build and test large sample sizes. Using a mixture of reliability analysis alongside traditional stress testing can increase confidence in the device and allow for smaller sample sizes and shorter test times.In this paper, a methodology will be presented for cost-effective analysis and testing of complex electromechanical medical devices during the design process and as part of production. As the design transfers into manufacturing, the role of stress testing changes. In manufacturing, stress testing [5] can be used to screen for infant mortality failures [6]. Ongoing analysis of the results of stress testing during manufacturing can be used to reduce the cost of stress testing during manufacturing. While others have presented practical reliability testing methods [7], few have shown actual examples. In this paper, actual case studies will be used to demonstrate this approach.A reliability plan should be included as part of the development of a medical device. A typical reliability plan lists the reliability goals along with the plans for testing and analysis. Goals for reliability can include typical mean time between failure (MTBF) goals or capture of early failures using stress testing. Testing to traditional MTBF goals can be expensive during the design process as sample size and test time are dependent on the MTBF goals and required confidence levels. Oftentimes for a complex medical device, other stress tests and analysis methods may be used to reduce the cost of testing. A reliability plan can be used to improve the design, as shown in Fig. 1.Given a goal for an MTBF of 40,000 h and a confidence level of 90% could lead to required test times of approximately 10,000 h for ten devices [8]. Although acceleration factors can be applied to reduce test time, it is still expensive and requires a large number of devices.Analysis methods and accelerated testing can be used to significantly reduce the time and cost of reliability testing during development. Analysis tools uncover weaknesses in the design and when supplemented with short-term accelerated testing can replace expensive MTBF testing.As the design moves into manufacturing, stress testing can be used as a means to remove the devices with infant mortality problems from the device pool. Devices typically fail according to the bathtub curve with many failures early in life, followed by a few failures until the device starts to wear out. Manufacturing stress testing is particularly important for Class III medical devices, such as life support equipment. Design of the stress test for manufacturing must be set up carefully so as to remove the true infant mortality failures without overstressing the equipment.The stress testing during manufacturing should be set up to collect data regarding the faults that are observed. This data can be assessed for trends and to troubleshoot problems, which can be used to improve the design.Two case studies are presented to show this approach.The following case study shows the results of reliability planning and testing for a medical grade single board computer that is a component of a blood analysis device. The requirement was a stated reliability goal as follows: “An MTBF of 5 years of approximately 20,000 h of continuous use will be achieved.”To reduce the test time and sample sizes, a mixture of reliability analysis and short-term testing was used instead. The reliability plan included the following steps:Step 1 stress analysis: Using a tool capable of doing FEA, it is possible to predict electromechanical weaknesses in the design of an electronics subsystem. By adding in information about the circuit board configuration and mounting and individual electrical component characteristics, it is possible to predict the MTBF. The analysis is much less expensive than traditional testing.Step 2 HALT testing: A HALT test was performed on a sample of three units to increase the confidence in the analysis. The small sample size for the HALT test was justified since the testing was done to confirm the MTBF analysis. Vibration in the range of 5–45 grms (root mean square acceleration) and temperatures from −10 to 90 °C were applied to the subsystem while it was operating. These limits were chosen due to the limits of the supporting test equipment.Results of the FEA analysis are shown in Figs. 2 and 3. In Fig. 2, the MTBF prediction was determined to be below the required value. In Fig. 2, a vibration analysis shows an area of the circuit board susceptible to failure in red. Updates were made to the design based on this initial analysis to strengthen critical areas. The HALT testing was performed on the updated design to confirm the design approach. No operational failures occurred during the HALT test. X-ray inspection of the circuit boards showed no solder joint issues or component issues.The following case study shows the results of reliability planning and testing for a small wearable Class III (life support) device as part of the manufacturing cycle; environmental stress screening provides a means to stress a device to screen out early failures prior to shipping the device.Cost-effective screening during manufacturing requires tracking results. By tracking and analyzing failures during screening, the effectiveness of the testing can be assessed and updated to be more efficient.The initial reliability plan for the device was the following. The manufacturing flow is shown in Fig. 4.The results of an analysis of the environmental stress screening showed that it was effective at removing failures from the build population of devices. Typical failures detected included solder failures, open vias, and failed components. The number of thermal cycles to failure for a sample of failed units is shown in Fig. 5. Over time it was shown that the random vibration of the circuit boards resulted in few failures and that all failures were detected during thermal cycling. The random vibration screen was later removed to reduce cost during manufacturing.
No takes yet. Share an insight, caveat, or question.
Lucke et al. (2014) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: