Batch testing leaves much to be desired, especially for medical devices and components. A method that embraces the entire manufacturing process could lead to better and more cost-efficient products.
Medical device manufacturers face unique and challenging issues when it comes to product quality, regulatory and standards compliance, and defect tracking. Because the products made by these companies address human health and well-being, and are often implanted or used in the body, there is no room for device failure or noncompliance.
Facing ever-increasing regulation and scrutiny by FDA, medical device manufacturers need cost-effective ways to monitor their manufacturing processes and to test their products. The industry has made significant advances in recent years in the area of monitoring and improving process control. By automating processes that were once manual, many companies have seen their quality improve. And by monitoring the process, they’ve been able to ensure that control is maintained.
However, when it comes to product testing, the change has not been quite so forward-thinking. Today’s de facto quality test standard, batch destructive testing, is expensive, reduces yields, and, ultimately, fails to certify that every part conforms to standard.
There is a way to eliminate destructive testing, bring more accuracy to the manufacturing process, and reduce costs: process signature verification. For signature verification, every aspect of the manufacturing process that could impact product quality is captured as it happens. One accurate approach involves monitoring and recording key attributes in real time throughout the duration of the process, producing what is called a process signature. The shape of the characteristic curve contains detailed information about the quality of the manufacturing process for each individual part. By recording and analyzing these process signatures, it is possible to identify key features in the signatures that are correlated with final product quality. These features can then be tracked and tested against limits to determine pass or fail on a part-by-part basis.
Rather than subjective post manufacturing quality control approaches, such as visual testing, or statistically based tests, like destructive testing practices or auditing, comprehensive process signature verification provides actionable insight into manufacturing processes as well as a wealth of data that can prove compliance with internal directives and FDA regulations. This approach ensures product quality in real time on a part-by-part basis. Automotive and industrial manufacturers have used the process for several years but it is also beneficial to the medical device industry.
Medical devices are typically manufactured in batches or lots, with quality checks through an end-of-line test that determines whether the products meet the quality standards dictated by both internal policy and regulatory standards. Using destructive testing, a representative sample of parts from each production batch is tested both from a function and performance perspective and for reliability and durability.
In most scenarios, the sample parts are destroyed during this testing process. Based on the results of these sample tests, the failure rate of the balance of the batch is then estimated. Should the estimated failure rate fall above acceptable parameters, the entire batch is removed from production, quarantined, and most often, scrapped entirely, regardless of the fact that there are probably a considerable number of the units that were up to standard. The good units get scrapped along with the bad because there is no way to determine which are which, let alone identify exactly what caused some units to fail.
Although the sampling method is the de facto standard, it is a costly approach. At a minimum, manufacturers diminish their yield by the sample size or, worse yet, the entire batch.
The significant labor and capital costs associated with the performance of the test contribute to its expense. The effectiveness and cost-effectiveness of this approach relies heavily on how well the process is controlled through regular and effective maintenance, and strict adherence to procedures and protocols. In short, the less controlled the process, the larger the required sample size, and the higher the cost of the testing. As a result, maintaining a well-controlled process is critical to controlling costs.
Because batch destructive testing relies on a representative sample rather than on information from each unit produced, it provides little or no direct evidence of product quality for the parts that are actually shipped to customers. Instead, it is assumed that because parts are manufactured in the same batch, and ostensibly pass under the same conditions, the rest of the parts in the batch are also good. This supposition jeopardizes customer safety because the manufacturer cannot assert that it indeed inspected and determined the quality of the device that was used by the patient.
Another inherent problem with sample testing is the lack of actionable data available based on results. It’s an end-of-line methodology that takes place after the entire batch is through manufacturing and whatever variables changed during processing are already completed. If a sample fails at this point, the manufacturer has two options: institute an expensive part-by-part manual inspection, which still won’t catch some defects that are impossible to determine by inspection alone, or reject the entire lot so that the 5–10% that are of poor quality do not get shipped to the customer.
Should manufacturers opt for the former, the test process may be lengthy and take days or even weeks to complete. If defects are detected, the long delay between manufacture of the product and the conclusion of the testing can mean that several other batches, and perhaps thousands of devices, have been manufactured with the same potentially defective process parameters and will also need to be quarantined. The costs associated—in dollar terms and in terms of a company’s reputation with prospects, customers, and FDA—are in some cases incalculable.
An in-process strategy is based on the principle that all manufacturing defects are the result of deviations in one or more process inputs, including variations in component characteristics, process station parameters, or environmental factors. It is important that all potential variables during critical manufacturing processes are monitored, from a compliance standpoint and also for traceability and risk mitigation. What was the humidity in the plant at the time? Which operator oversaw a particular process on the line? What was the temperature in the plant at the time of manufacture? By monitoring and measuring all of these inputs, it is easy to assess when a variable has changed and act quickly to rectify it. It can be quickly determined whether the problem is with the process, training, or component quality. This kind of insight is tremendously valuable when faced with proving compliance to FDA.
|Figure 1. A press-fit process signature (force versus distance). A-1 is the alignment work, Dy is the differential force, Y(x1) the maximum force required to align the part, X1 is the maximum distance traveled by the part, Dx is the differential force to measure the press retract travel.|
What constitutes best practices for in-process testing? Process signature technology is the capture of the unique process signatures created by each manufacturing process as the process is carried out. Figure 1 shows an example of a process signature and highlights the sorts of data points that can provide manufacturers with actionable insight. The area represented by A1 indicates a higher than normal value, meaning that a great deal of work is needed to align the parts for the pressing process. Two possible outcomes could result in poor part quality, such as a mismatch between the subcomponents being fitted or a misalignment of the parts press. Based on this information, the manufacturer can audit part geometry to identify whether parts are mismatched or audit the tooling station to determine whether misalignment is the problem.
By capturing and, in turn, analyzing these signatures for critical manufacturing processes, manufacturers can quickly identify any deviation from the ideal and, because they’re capturing all inputs, then manage the problem.
This is a global marketplace and when it comes to medical device manufacturing, there has been significant adoption of offshoring and outsourcing to contract manufacturers in Asia and elsewhere. To compete with the savings offered by these lower-cost options, U.S. companies looking to keep research and development close to home must streamline and find ways to reduce costs, while never sacrificing quality. In-process testing is a viable solution.
As a participant in the global marketplace, U.S. medical device manufacturers are in a highly competitive environment where cost is a significant factor in purchasing decisions. Process signature technology lowers costs for manufacturers because it increases yield, improves process efficiency, and facilitates continuous improvement to manufacturing, all in real time. This enables manufacturers to make better products for less money, savings that can then be passed on to the customer in the form of lower prices.
Perhaps the biggest driver for process signature technology adoption is compliance, both to internal standards and to FDA regulations. FDA’s introduction of title 21 CFR part 820 quality system regulation (QSR) instituted multifaceted manufacturing and process-measurement regulations.
The mandate of the QSR is to ensure that variations in the device manufacturing process are understood and minimized to produce low-risk, high-quality products. The stringency of FDA’s regulatory framework requires that systems be put in place to protect the consumer, even at the expense of a manufacturer’s bottom line. Process signature analysis supports compliance with these regulations.
In its recommended quality framework for pharmaceutical manufacturing, called process analytical technology, FDA outlines process signature capture and analysis as a sound methodology for quality control. “For certain applications, sensor-based measurements can provide a useful process signature that may be related to the underlying process steps or transformations. Based on the level of process understanding, these signatures may also be useful for process monitoring, control, and end point determination when these patterns or signatures relate to product and process quality.” The same principles apply for medical device manufacturing and the same benefits can be achieved.
Before a manufacturer is permitted to market a new medical device, it must produce a premarket approval (PMA) submission to FDA. A successful PMA or 510(k) submission must be accompanied by sufficient data for FDA to judge the safety and effectiveness of the new device. Such a task is more easily accomplished when there is a comprehensive dataset demonstrating that the critical processes are well understood, and the causal relationships between the process controls and product attributes have been established. Again, in-process testing and process signatures support these efforts.
If, despite a manufacturer’s best efforts, a defective product makes it into circulation, manufacturers using signature process methodology can quickly find and fix the problem, prove how it has been corrected, and get the production lines running again. Recalls are limited only to those products that are defective and the manufacturer has a comprehensive and detailed history record for its products that provide actionable proof as to which units were implicated. Intensive analysis can then be done to find out which variable caused the defect and why. Meanwhile, FDA can be assured that the product is well understood and a solution is in progress, limiting if not eliminating downtime on the line.
Environmental initiatives such as lean manufacturing also support the adoption of signature process methodology. By eliminating destructive testing, manufacturers also eliminate the scrap produced by the destroyed units. Rather than these destroyed samples or, worse yet, the whole batch, ending up in landfill, problems are eliminated at the root. Further, manufactured products adhere to requirements and can be shipped to the customer with the manufacturer confident of the product’s integrity.
Of course, the most important reason is the very real improvement in consumer safety that comes when 100% of all devices are individually checked for quality. When using process signature analysis, medical device manufacturers can provide solid assurances to their customers that each unit meets internal standards and FDA regulations. This ability provides a far greater level of assurance than relying on testing just a representative sample. Given the huge focus on quality in the marketplace, consumers are demanding more stringent and reliable quality processes, and there’s no situation in which quality is more important than when human life is at stake.
|Table I. Comparison of end-of-line destructive testing and real-time release, made possible through process signature technology.|
When medical device manufacturers want to make even slight changes to their manufacturing processes, they are required to submit the change to FDA for approval. When adopting a new technology approach, it can be met with initial skepticism, especially when faced with the seemingly pie-in-the-sky prospect of driving down costs by deploying a new quality control system. Thankfully, this technology is proven to be effective and cost-effective, as shown in Table I.
When building the case for adoption of signature analysis, it is important to emphasize to regulatory agencies a number of key points. It’s obvious that product quality improves when test coverage increases from a representative sample to 100% of all parts. A manufacturer’s case is be even stronger if it seeks to institute quality control of each critical manufacturing process rather than just several points on the line, such as during welding or crimping. Without such a holistic approach, 100% product quality cannot be confirmed.
Furthermore, using in-process testing, data acquired during manufacturing can be correlated to the precise step where a defect was created, providing valuable feedback for optimizing and maintaining the manufacturing processes. This ensures that the quality of the manufactured product is controlled and maintained on a continuous basis. Finally, by consolidating and storing all of the in-process test data and process signatures associated with each part, it becomes a vital component of the device history record. These are compelling arguments for adoption.
The current approach to quality in medical device manufacturing is inadequate, expensive, and, worst of all, does not ensure patient safety. Because industry is so heavily regulated, and rightly so, making the transition to a holistic process might be met with skepticism. But a significant number of leading medical device manufacturers have adopted in-process testing as an alternative and this should pave the way with the regulatory bodies for widespread adoption of the technology.
A clear understanding of the cost, environmental, business process, and consumer safety benefits can help a medical device firm adopt process signature analysis. The method is a comprehensive approach to in-process testing, one that captures and analyzes the process signatures of all critical manufacturing processes, that arms the manufacturer with a wealth of information. The process signature method not only supports regulatory compliance, but provides visibility into processes and enables complete product life cycle traceability. Root causes are readily determined and plans can be quickly put in place to rectify any problems that arise.
It’s not enough, however, to merely capture these data. An important step in streamlining manufacturing processes is stringent analysis of the collected data. This analysis provides manufacturers with insight into how to avoid future problems and identify issues that are impacting product quality. Such challenges could range from poor component quality to training issues, from temperature to humidity, to when and where the process was completed. Problems can be avoided in the future, which again saves both the manufacturer and its customer money.
Ron Pawulski is director of sales, medical for Sciemetric Instruments (Ottawa, ON, Canada).