Quality, Speed, Price: Pick Three

A keynote address at the 7th Annual Orthopaedic Manufacturing & Technology Exposition and Conference (OMTEC) in June 2011 forecasted several changes in FDA regulatory policy and a concomitant evolution in the medical device market. The trend toward “negative innovation” is slowing the adoption of cutting-edge products—to the ultimate detriment of patient, doctor, hospital and manufacturer.

This increased emphasis on comparative effectiveness necessitates an expansion of data collection. This has some subtle implications. To counter increasing bureaucratic control, manufacturers need not just clinical outcome data, but also a comprehensive understanding of their own products’ geometry, the details of the results of their manufacturing process.

As W. Edwards Deming forewarned, foreign competitors throughout precision manufacturing—especially, but not exclusively, “Brand China”—are quite receptive to new ways of improving part quality while still decreasing cost, complexity and time to market. As we’ve seen in other industries, once we’ve lost the lead to foreign manufacturers, we’ll have a hard time getting it back.

Globally, the medical device industry remains competitive and promising. Even with the recent slowdown in Asia, global annual growth in the industry is projected to lie between five and 10 percent throughout the coming decade. This is still a great business to be in!

The challenge is to adapt to the changing times and find the best new ways of doing things. That’s only possible if manufacturers understand the available modern technologies.

Best practices have crystallized over the years and decades, but what might have been a best practice even five years ago might be dangerously antiquated now. American manufacturing firms, many of which are now struggling to survive, can’t afford to ignore opportunities to reduce new product development time by 30% or more—and improve product revision program speeds by upwards of 50+%. To prosper, they will have to change.

The Project Triangle: A 20th Century Problem

The traditional approach to design is borrowed from the Triple Constraint model of project management. Quality, schedule and cost are usually assumed to exist in tension. Manufacturers have generally attempted to optimize two of the three, and have given up on the third. (See Exhibit 1.) This is no longer necessary for those who are willing to apply advanced tools to their enterprise’s continual improvement.

Exhibit 1: The Project Triangle

We’ll use “quality” to mean excellent fit, form and function—in addition to the absolute requirements of safety and effectiveness. For too long, the standard practice in the medical device industry has been to optimize only cost. Manufacturers waste time by struggling half-blindly through an inefficient iteration sequence until they stumble onto a passable part. Conjecture is risky when one is working with inadequate decisionable information.

For new products being brought to market, some orthopaedic manufacturers have estimated the cost of each day’s delay at $75,000 (and more) in lost profit. Time is, indeed, a lot of money! Speed is of paramount importance. That has never been more true than now, when foreign manufacturers are demonstrating greater adeptness than ever before.

In the 21st century, manufacturers who take advantage of advanced technological tools can secure the needed quality, while improving both schedule and cost. Manufacturers can thus optimize all three of the seemingly elusive objectives of quality, speed and price, while actually increasing confidence.

The Inadequacy of 20th Century Metrology Data

I have a saying: CMMs fool you—optical comparators lie. Most users and managers understand this painful reality, but have too long accepted this status quo.

Coordinate measuring machines (CMMs) are still seen by some in the medical device industry as the gold standard in inspection. This belief persists despite the fact that FDA allows no special privilege of place for CMMs. FDA’s ultraconservative ethos requires quality departments to verify CMM data with handheld calipers and micrometers. The failings of CMMs far outweigh the few merits. At best, CMMs provide inadequate information, demand too much effort, require many questionable assumptions and yield scant decision support value.

One failing, which obliterates claims of accuracy, is vulnerability to human judgment. A CMM operator will always be susceptible to human error, regardless of experience. (One CMM programmer who had more than 30 years of experience in the medical device supply chain frankly admitted this in my office earlier this year.) And in our correlation studies of advanced metrology to CMMs, we discover latent CMM programming errors more often than you would like to believe.

Even more disturbing than the CMM’s human component is its inherently primitive treatment of exceptionally limited data. The shoulder dome in Exhibit 2. was inspected both with a CMM and a modern, 21st-century White Light Scanner (WLS). In this color plot visualization, green areas show surfaces that are within the tolerance band, while blue areas indicate minus material beyond the allowable tolerance band. (This deviation spectrum illustration is very typical, but configuration options are practically limitless.)

Exhibit 2: White Light Scanned Shoulder Dome, with CMM Points Superimposed

The CMM program presented 186 touch points, while the high-resolution 3D scanner captured 200,000 surface scan points—more than 1,000 times more. The CMM entirely missed the ring-shaped surface anomaly. Had any of the CMM points landed inside this anomaly, the deviations might have been mischaracterized at best…dismissed as flukes, at worst.

Optical comparators (OCs) are even more deceptive than CMMs. A “classic” OC casts the part’s shadow against an opaque screen, allowing the operator to view the largest profile of the part in silhouette and make an educated guess about the overall dimensions. A newer, dressed-up OC casts the same shadow, but does so digitally. Superimposed against a cross-sectioned CAD overlay, it permits a similar conjecture to that possible with an analog OC. This “measurement” process is completely subjective, and has led manufacturers to incorrect conclusions, with no confidence.

Too Slow for Modern Manufacturing

Another problem with 20th-century inspection is its lack of speed. During prototype and manufacturing process development, each iteration has to be inspected with an expensively-programmed CMM and/or expensively-fabricated fixed gauges and/or completely subjective handheld micrometers and/or (worst, because it lies) an optical comparator.

If an iteration is substantially different from its predecessor, a whole new PC-DMIS program will have to be written. This can require dozens of paid staff hours and stall the process for valuable days. It can then take 45 minutes or longer for the “comprehensive” (per se) program to gather only a few hundred points.

These development speeds are unacceptable in the 21st century medical device industry, in which every extra day’s delay can cost $25 million or more in lost revenue. (See Exhibit 3.) Development is expensive—delays prohibitively so. Now, using better technology, manufacturers can reduce development costs and avoid launch delays.

Exhibit 3: The Project Timetable and Associated Costs

You Get What You Pay For

There are at least two competing cost considerations that make even a comparatively inexpensive CMM far more expensive than it might first appear.

Even if we ignore the frequent problem of CMM programming errors, sparse CMM data can be responsible for costly errors during the design process. Engineering misjudgments resulting from inferior data can balloon costs and development time. One reason that the medical device industry’s scrap rates are orders of magnitude greater than those in automotive and aerospace is the stubborn adherence to such outdated, cost-ineffective inspection technologies.

Even worse in the long run is the opportunity cost inherent to slow product development. A trial and error approach to form and function can result in dozens of redundant and ineffective iterations where one or two strongly-supported iterations could have been sufficient. In this, too, the American medical device industry lags behind aerospace and automotive as confirmed by those who have transferred into this industry from the outside.

If the part even passes dimensional testing, it is subjected to a bench testing cycle that amounts to little more than trial and error. In the extreme cases, the part might undergo dozens of design iterations before it is ready for manufacture.

On balance, CMMs only seem inexpensive. In the best case, their full cost is deferred for a while.

The Accurate and Comprehensive 21st Century Solution

When paired with advanced computer aided inspection (CAI) methods, 21st-century inspection technologies—especially non-contact white-light scanning—provide far more “bang for the buck” than traditional CMMs and optical overlay comparators. Let’s walk through a few of the reasons for the superior value proposition.

White light scan data are at least as accurate as CMM data, and far more accurate than laser scan data. In the shoulder dome study illustrated earlier, we submitted measurements from the WLS data at each of the 186 points from the corresponding CMM readout. When the customer compared the WLS measurements to those taken by the CMM, perfect correlation was found at all 186 points out to the sixth significant digit! WLS accuracy is expressed in the millionths of an inch when the technological process is properly implemented.

Equally striking is the comprehensiveness of WLS data. The 200,000 3D data points scanned from the shoulder dome were trivially few by the WLS technological standards, when compared to the millions of points that can be recorded in a few seconds at the speed of light. But feast your eyes the difference that just three orders of magnitude can make!

Engineers love data, and thanks to the level of detail now available, process engineers can have valuable insights that ten years ago they only dreamed of. The very moment that the quality engineering team saw the color plot, they were able to diagnose the cause of the ring-shaped surface anomaly, pinpoint the process problem and devise a solution within hours. This would have been impossible even if the CMM had dropped a few points into the depression—which it had not.

Good design depends on good geometric understanding. Engineering misjudgments of mere microns can ruin the outcome of a joint replacement and necessitate an expensive recall. Paired with the fantastic accuracy and comprehensiveness of WLS, modern computer aided inspection and analysis software can instantly ascertain geometric surface information, discern trends and catch new problems as they arise—before they begin to cost time and money.

21st Century Speed

The speed of a CMM inspection is limited by the speed of the mechanical touch probe. White light scanning takes place, as the name implies, at the speed of light. (See Exhibit 4.) A traditional commercial WLS configuration, devised in the early 2000s, can inspect a geometrically-simple part like a shoulder dome or an acetabular shell in less than ten minutes, yielding tens of millions of 3D data points that reveal the minutest surface detail.

Exhibit 4: White light raster lines create millions of data points, yielding a three-dimensional point cloud in a few seconds.

Very recently, self-contained robotic WLS configurations such as the Smart Inspection Station™ (SiS™) have entered the market. With a single technician-driven part setup, the SiS learns a part’s geometry and optimizes the scanning process to minimize the number of rotations and angled shots. The SiS generates a reusable setup file that is applicable not only to the part being scanned, but to all part sizes in the family, based on similar geometry. This contrasts starkly with the hours or days that might go into a new PC-DMIS program.

The SiS uses the setup file to capture and process individual scans in just a few seconds. By removing human interference, the SiS can create part scan files in as little as three minutes. That’s process-quality speed! Even if 100% inspection is required because of regulatory considerations, robotic part handling makes it practical. Quite soon, those costly inspection cells filled with dozens of inspectors armed with fixed gauges, micrometers and optical comparators will be a memory. So will the impact of human inspectors’ proneness to error.

21st Century Savings: Time is Money

The aerospace industry learned early in the last decade that modern inspection technologies can cut overall product development time dramatically. Early successes have demonstrated a similar result in orthopaedics. Recently, an orthopaedic design team approached us in frustration. A product revision had gone through 18 separate process iterations, aided and abetted by an inspection team using optical comparators, CMMs and fixed gauges. Employing WLS-based CAI, we were able to finish the product revision in only two iterations. If you’re doing the math, that’s better than an 85% reduction in cycles, and weeks. It’s time to relegate wasteful trial and error iterations to the ash heap of history.

Accurate and comprehensive geometric data allow engineers to diagnose and solve problems that might have vexed them for months – or even years, as we have found with some clients. (See Exhibit 5.) Less dramatic but at least as important is the opportunity to incrementally streamline production based on real-time insights, which can then be driven back into the process to create a cost-effective production measurement plan that actually makes sense.

Exhibit 5: Accurate Point Cloud Data in a Feature-Rich Colorplot Gives Greater Insight

 By broadening the focus to tooling, prototyping and other upstream activities, manufacturers can tighten control, catch emerging trends before they grow into costly problems and reduce the likelihood of costly overhauls. By trending insights, medical device manufacturers can improve their understanding of tool and machine wear and plan maintenance outages more intelligently.

Lastly and importantly, tolerances can be loosened and made more practical as manufacturers learn more about the realistic limits of their processes. We believe that extravagant tolerances need no longer keep good products off the market. By weaning themselves (and FDA) from slavish adherence to irrational tolerances, manufacturers can now cut scrap rates and time-to-market to a fraction of what it used to be.

Conclusion

In light of recent developments in fast, accurate and comprehensive inspection, it’s time to recalibrate our understanding of the Triple Constraint model of project management. Compared to the options that orthopaedic manufacturers had even as recently as ten years ago, it is now possible to balance quality, speed and cost-effectiveness as never before—as suggested in Exhibit 6.

Exhibit 6: Who says you can’t have it all?

Among the somber (but penetratingly accurate) diagnoses that we overheard at OMTEC was the fact that foreign manufacturers are rapidly gaining on the U.S. industry. Orthopaedic and spinal implants are two industries in which the U.S. still leads.

For the U.S. to retain its edge, four goals must be met:

1. We must continue to make medical devices to a higher quality standard than foreign competitors.
2. Devices need to become easier to use.
3. Costs must come down.
4. Innovations need to come onto the market sooner.

Healthcare reform and market headwinds have rendered the “First Mover Advantage” more significant than in the past. However, with modern 3D scanning and computer-aided inspection, U.S. manufacturers can still meet all four goals. Manufacturers that have the strategic vision to best adopt and use advantageous new technologies will lead the industry into the future.

ALL SYSTEMS GO.

With over a decade of industry leadership in advanced metrology, product design, investment casting and CNC machining, Level 3 Inspection LLC offers the world’s most accurate and comprehensive dimensional inspection of precision manufactured parts. L3I’s clients gain confidence and competitive advantage while cutting time and costs through broad production process optimization.

Bill Greene is a co-founding Director of Level 3 Inspection LLC and is leading the Strategy Development efforts while serving as CEO and VP Business Development for the company. Based on his Metallurgical Engineering and Economics educational foundation, he has focused his career on introducing and delivering business-favoring technology solutions to major industry clientele. Having engineered and automated most manufacturing processes, Mr. Greene is well versed in the needs, opportunities and ROI requirements of modern manufacturing enterprises.

Save Supplier