In the medical device industry, there are huge pressures and incentives to file FDA 510(k) or PMA submissions to gain approval for market entry. Start-up companies literally live and die by meeting these short-term milestones because they often are attached to funding rounds. From start-ups to even the most experienced, established OEMs, if the race to meet short-term milestones is run blindly, the achievement of them can forever hamstring a medical device timeline, limiting its production capacity and lessening the impact that device can have on patient lives. In this industry, like no other, it is important to avoid the trap of filing a weak design too soon. Locking down too early a design that dictates the methods of manufacture impedes constructive design iterations and improvements.
Design for manufacturability (DFM) and design for assembly (DFA) are well-known methodologies. They are just two design practices within a larger philosophy sometimes referred to as design for excellence (DFX) that aims to optimize a product’s design for ease of manufacturing. These tools have been used in the development of consumer products for a long time to reduce costs and increase margins.
In the medical device industry, DFX can be particularly powerful not just for cost reduction but, when used at the right time, it can be the most effective way to streamline the development timeline and smooth the road to stable production. When done correctly, it helps achieve the best possible quality and highest patient impact. To achieve these benefits, companies can implement a variety of DFM methodologies (see Figure 1). Every medical device engineer should know the definition of the acronyms important at the design stage.
Design for manufacturing (DFM) ensures that the part is designed specifically for the process that will be used to manufacture that part.
Design for assembly (DFA) ensures that the whole product can be assembled easily with high quality.
Design for reliability helps ensure that a product will perform its function over its expected lifetime.
Design for quality (DFQ) ensures that a product will work reliably.
Design for maintenance (DfM) ensures that a product can be serviced to extend it useful life.
Design for test (DFT) ensures that products can be tested effectively and efficiently.
Design for supply chain (DfSC) improves the overall logistical efficiency of a product.
Design for cost (DFC) minimizes costs associated with manufacturing.
Design for sustainability (DFS) emphasizes the well-being of people and the environment.
Many times, these concepts can get lost on the road to commercialization. The array of individual tools and methodologies encompassed by DFX can be dizzying and ever changing. While individual disciplines of DFX are important, the timing of when to use the philosophy of DFX is much more important. Far too often DFX is treated like a checkbox item somewhere between the end of the design phase and the start up of the production line — a line item on a Gantt chart to fit in the middle somewhere after the team sees if the prototypes work.
When should a designer start to evaluate a product with a mindset of DFX? The answer is: ‘earlier than you think.’ A typical medical device development timeline may allocate a very short time in concept and design only to suffer long periods during early builds implementing fixes and engineering change orders before beginning to ramp up to stable production. Because of the regulations inherent in medical device manufacturing, these changes become extremely difficult to implement as the program matures.
By intentionally expanding the conceptual phase of a device development timeline with a disciplined mindset and adherence to the tenets of DFX, companies can often get a jump start on their product commercialization, hurdling over much of the firefighting and change orders inherent in first article builds, pilot runs, and process validations (see Figure 2). Sounds great, but how?
By the Numbers: Metrics Matter
Development teams need to measure the viability of a particular design approach early and often to determine whether they are on the best path toward commercial production or whether they need to change direction. This is best done by using a set of metrics to gauge one design approach against another.
The single most important metric that can be used to evaluate a product design is also the simplest one — part count. While it is easy to understand how a device with 100 parts would be harder to manufacture than a device with 10, what may be underappreciated is the way in which the natural sources of manufacturing variation multiply geometrically over those parts. It is not just the 100 parts you need to plan for, it is the two or five critical specifications of each of those parts, multiplied together over all of those parts, that can become unmanageable. These variations are not always seen in short-run design validation builds unless intentional limit challenge builds are conducted.
Let’s imagine a product that has a frame design made of five parts, precision-machined out of aluminum. The CNC equipment available today is quite capable of holding very precise tolerances, down to 50 millionths of an inch, and certainly the parts can be specified with tolerances controlled tightly enough so that all the parts fit together at minimum and maximum material conditions that will result in a frame assembly that meets specifications. Now imagine the cost of all those precision-machined components, the pins and fasteners used to bolt them together, and the load they put on the quality systems (both external and internal) to ensure that they meet specifications over the lifetime of the product.
Now imagine that same frame redesigned to combine those five parts into one part made with a net shape process, like vacuum casting. With critical-to-function dimensions machined after fabrication by a net shape process, the total number of sources of variation are minimized geometrically. Instead of trying to manage the variation of CNC machining on five parts, the variation factors only affect one final machining operation.
The total number of parts in a device design has a larger systemic effect on the entire business. Fewer parts to design means fewer parts to procure, fewer parts to receive, inspect, inventory, and conduct quality operations upon. Further, fewer opportunities for those parts to fail to conform to specifications, fewer opportunities for people and process to damage or install incorrectly … the list of risk reductions is extensive.
Another metric by which designs can be measured for manufacturability is process — not only how many processes are required, but whether there are special processes that are hard to control or have long cycle times. For example, liquid adhesives are notorious for being problematic in manufacturing systems. Dispensing the exact right amount of adhesive, in the right place, is only half the problem. Waiting for an epoxy or adhesive to cure in the right conditions is the other half.
These are setbacks that medical device development teams may not consider early in device design. Well-intentioned teams, in a hurry to get a working prototype into a functional test to show results, sometimes cannot resist reaching for a tube of glue to stick parts together. If that design choice moves forward into production, it can present significant challenges to a manufacturing team validating the process. Further along, when customer demand rises, scaling up an adhesive dispense-and-cure process means lots more work in process (WIP), clamping fixtures, and considerable space requirements to house the operation and maintain order. The opportunities for error compound quickly.
There exist many industrialized processes for joining parts together without glue. Plastic injection molded parts can be joined with snap fits or can be heat staked. They can be welded with a number of robust processes like hot plate, infrared radiant welding, spin welding, linear vibration welding, and ultrasonic energy. Better still are options such as the compounding effect of utilizing the sub-techniques of plastic injection molding, such as overmolding, insert molding, and in-mold decoration to combine parts, features, and materials during the injection molding processes, where the economies of scale really start to add up.
Another metric that is sometimes overlooked is the number and type of threaded fasteners used in a design. Just seeking to standardize threaded fasteners is not enough. Consider whether the design really needs to utilize a threaded fastener at all. With plastic enclosures being more widely used as medical devices become smaller and more portable, the use of threaded fasteners becomes less than optimal. Overtorquing of threaded fasteners can easily crack even durable plastic resin cases, especially if they involve compression against a sealing surface. The bosses used to house threaded inserts can be prone to knit line formation during molding that can lead to crack formation when torque is applied.
Sometimes design simplicity is the best path forward. A simple plastic joining technique, Mattel Pins, which was created by a toy company to reduce cost, has become a fairly common but overlooked method of closure designed into small handheld or body-worn devices.
DFX Case Study: A Wearable Medical Device
A recent project taken on by Minnetronix involved a wearable continuous glucose monitor utilizing state-of-the-art optical technology. The customer had functional prototypes, but the design was unfit for high-volume manufacturing. The design was comprised of 50 individual parts, eight of which were threaded fasteners. There were machined parts, glass lenses, gaskets needing compression, and (most limiting of all) a set of optical windows created by dispensing a clear epoxy very carefully, without any bubbles, that then had to be cured for 16 hours. The assembly was not automation compatible, and building 600,000 devices per year was not feasible.
A cross-functional team of designers was able to reduce the number of parts in the design immensely by using a phase zero approach to quickly generate concept architecture, evaluate, and then iteratively redesign the device until all that was left were seven components for final assembly in a bottom-up stackable design. The design was automation compatible and could successfully reach production volumes of 600,000 devices per year.
The design was simplified greatly by using plastic injection molding and the subtechnologies of in-mold decoration and overmolding to combine parts and features into value added subassemblies. Most notably, the optics block involved the combining of 29 individual components into one through a series of optical molding, overmolding, and insert molding.
This result was only possible through the willingness of the team to back up a bit on the product development timeline and reenter the phase zero design segment with an open mind, creativity, and the resolve to take DFX to the limit.
DFM is a crucial philosophy that represents a set of techniques and, most importantly, a discipline that can help medical device companies get products to market faster by ensuring stable production ramp-ups and achieving the highest possible quality. Design teams can improve outcomes by inculcating a mindset that includes multiple iterations of concept architecture in a rigorous phase zero, considering the entire system by which the device will be manufactured, long before the commercial production phase is reached.
While the tools of DFX stand on their own, its true power is brought to bear when the entire design and development organization embraces the mindset and ruthlessly slashes design complexity — eliminating parts, adhesives, threaded fasteners, and special processes.
By implementing an early design evaluation process that encourages multiple iterations of concept architecture before starting detailed design and testing, companies can identify the most robust processes and design toward them to reduce costs and risks, and ultimately to improve patient outcomes.