MD+DI Online is part of the Informa Markets Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.


Articles from 1996 In September

Specialty Compounds for Medical Applications: An Introduction

Medical Plastics and Biomaterials Magazine | MPB Article Index

Originally published September 1996



Most current medical plastics applications feature materials that have been compounded in some fashion to optimize their performance. For example, polyvinyl chloride (PVC) typically used in medical fabrication often contains 10 or more ingredients, among them stabilizers, antioxidants, and other additives. Most engineering resins are compounded with UV stabilizers, processing aids, and reinforcing agents

such as glass fibers. Thermoplastic elastomers, alloys, and blends are normally compounded in order to make up the standard grades. The compounds that result--so familiar as to be sometimes thought of as off-the-shelf resins--are what device manufacturers and processors use to satisfy the requirements of most medical applications.

Despite the availability of thousands of grades derived from dozens of plastic resins, medical product designers continue to develop new, application-specific specialty compounds. This article offers a general overview of the compounding process and describes some of the most commonly employed fillers and additives, in the hope that a basic understanding of compound materials will help facilitate timely, cost-effective product development. Although the information presented should be helpful in the production of any polymer-based device, the discussion is particularly geared toward compounds for catheters and other medical tubing products.

Depending on the precise application, a typical starting resin might be PVC, thermoplastic polyurethane, polyether block amide elastomer (PEBA), flexible nylon, polyethylene, styrenic block copolymer, or fluoropolymer. It is important to select a base resin or family that has most of the attributes required in the finished compound. For instance, if high tensile strength is called for in a semiflexible compound, PEBA resin may be selected. Many times, the various grades within a resin family can be melt compounded to provide intermediate properties more specific to the application. Reactive compounding, grafting, and compatibilizing of blends can also be achieved. Though very exciting, the creation of new compounds through reactive compounding is also an extremely involved process, given the virtually unlimited number of options. This topic alone could be the focus of another article.

The basic premise of effective compounding is first to incorporate an end product's functional requirements--for example, radiopacity, color, or lubricity--and then to fine-tune the formulation and compounding process to obtain optimum results. Of course, any discussion involving specific materials for medical applications is based on the understanding that the device manufacturer is ultimately responsible for all testing--including material testing--to satisfy regulatory requirements.


In order for a plastic to be visible under fluoroscopy, it must contain a certain amount of radiopaque filler. The type and amount selected depend on the base resin and on the size (thickness), surface smoothness, color, and desired properties of the finished device. Although numerous metals and other materials have been evaluated as radiopaque fillers, 90% of all compounds use one or more of the five substances discussed below to promote radiopacity (see Table I).

Barium Sulfate. Low in cost and very stable, barium sulfate is the most commonly used radiopaque filler. The preferred form of barium sulfate is a white powder, with particle size ranging from 0.5 to 2 µm. Some grades may appear light gray or brown, but these often require high-temperature predrying to drive out residual volatiles left over from the manufacturing process, and are best avoided if possible.

Barium sulfate can be incorporated into thermoplastic elastomers up to 60% by weight without a significant decline in properties; semicrystalline resins such as nylon 12 and polyethylene can tolerate a 40% loading. Amorphous resins--for example, polycarbonate--cannot be filled as high. Although white barium sulfate can be used as a pigment, its poor tinting strength means that high loading is required to make the plastic opaque in thin sections. The same poor tinting strength can be an advantage, however, when coloring compounds, making it possible to obtain a wide color range, including dark colors and black, when barium is used as the radiopaque medium. All grades of barium sulfate used in medical devices should meet the relevant requirements of the United States Pharmacopeia (USP).

Bismuth Subcarbonate. After barium sulfate, the most popular radiopaque filler is bismuth subcarbonate. Because of its instability at temperatures around 400°F, bismuth is a bit more challenging to compound than barium sulfate. It also may not be compatible with all resins: for example, some polyurethanes may depolymerize during compounding when bismuth is added. Though more stable grades are currently being developed, bismuth fillers are still not often used with many polyurethanes, especially aromatic polyether urethanes.

Bismuth is denser than barium sulfate, and is recommended when more radiopacity is required. Higher density means that an equal weight percentage of barium and bismuth translates to a lower volume percentage of bismuth. Bismuth can thus be used at relatively high loadings--30 to 50% by weight--with little effect on polymer physical properties.

For the better grades of bismuth fillers, particle size ranges from 1 to 2 µm. The powder is white and fluffy and, depending on the grade, can be difficult to feed into an extruder because of low powder-bulk densities--a problem that can be overcome with proper compounding techniques. Bismuth is a strong white pigment, difficult to color match. It also has a tendency to turn yellow (reducing to bismuth trioxide) when compounded, which again makes it hard to maintain consistent color from lot to lot. Dark colors are not possible with bismuth subcarbonate. In general, good bismuth compounds require careful control over the entire compounding process, including precise temperature control of the extruder. Once again, USP grades are required for all medical applications.

Bismuth Trioxide. Like bismuth subcarbonate, bismuth trioxide is very dense and makes an excellent radiopaque filler. Its major drawback is its yellow color. If processed too hot, bismuth trioxide compounds turn brown, which has little if any effect on compound properties but is often undesirable for aesthetic reasons. Trioxide has a very high bulk density and is easy to feed. Gritty surfaces are occasionally a problem, but can sometimes be corrected through melt filtering. Bismuth trioxide is generally recommended if a high level of filler is required--for example, 60% by weight--or if bismuth subcarbonate cannot be compounded satisfactorily in a particular resin.

Bismuth Oxychloride. A white powder with a soft, silky feel, bismuth oxychloride is more temperature-stable than bismuth subcarbonate and is compatible with a wide range of resins. Particle size is typically from 2 to 12 µm; particles have a platelet structure that aligns during processing to form a smooth, shiny surface. Bismuth oxychloride is susceptible to UV degradation, and therefore requires addition of a UV stabilizer to combat this effect. Coloring properties are similar to those achieved with bismuth subcarbonate, with some grades producing a pearlescent appearance.

Tungsten. A very heavy metal powder compatible with virtually all resins, tungsten provides high radiopacity. Loadings up to 90% by weight are typical. The best grades have particles measuring between 1 and 2 µm, with low impurities. Tungsten compounds often show a characteristic matte finish at high loadings, and are dark gray in color. Because tungsten is so abrasive, it can wear out mixing elements in just a few weeks. Pelletizer bed knives and rotors will also wear rapidly, and extruder screws running tungsten compounds should be inspected frequently.


Colored compounds require formulation with pigments or dyes. It is generally advisable to employ the least amount of colorants in the formulation as possible; a good rule of thumb is to use less than 1% by weight. Although color concentrates have been used in many applications with success, it is more difficult to formulate using concentrates because they may contain unknown ingredients; in several cases, problems with a compound have been attributed to the improper use of concentrates. If possible, it is best to avoid concentrates if one is developing a new catheter compound.

In color matching, it is important to understand the limitations of the base resin and any other ingredients and to realize that a perfect match may not be possible. The Pantone color-chart system is commonly used, but the colored plastic will almost never exactly match the colored paper sample. A computerized color-matching system with a database comprising medically approved pigments and compounds can significantly speed up the matching process. As is the case with other additives or fillers, there are thousands of colors to choose from, and color qualification for medical applications is a time-consuming, expensive process. It is always best to use a pigment that has performed well in similar applications and with similar resins. As always, thorough testing of the finished device is essential when validating any new compound.

Titanium Dioxide. The most versatile pigment is titanium dioxide, a strong, white powder pigment that gives a uniform, opaque white color at 1% loading by weight and is also used in combination with other pigments to obtain most colors. Though compatible with all radiopaque fillers, titanium dioxide will not, however, overcome the yellowing of bismuth compounds. The rutile type is most often selected, and grades with very low impurities are required for medical compounds; special grades may be employed to reduce brittleness when added to polycarbonate and other amorphous engineering resins.

Organic Pigments. Phthalocyanine blues and greens in various shades are among the more common organic pigments. These colorants are particularly strong, and a loading by weight of 0.1% is usually sufficient to produce a mid-tone color. Phthalocyanine pigments are temperature stable up to about 450°F; the reducing nature of nylons will often turn a green to blue if the compound is processed at higher temperatures. Quinaridone red and violet--also very strong organic pigments that are compatible with most resins--are stable up to about 500°F.

Produced through a natural gas process, channel blacks are a carbonaceous soot that acts as a strong black pigment with excellent heat stability when used in weight concentrations up to 0.5%. Carbon black is also a carbonaceous soot, but is manufactured by a furnace process. A low-impurity, low-sulfur grade is required for medical applications. Both carbon and channel blacks help improve the UV stability of a compound.

Dyes are organic compounds that are soluble in many plastics. These colorants are suitable for some transparent compounds and can be compounded in at very low concentrations of less than 0.1%. Because dyes are soluble, migration is potentially a concern, and dyes should always be evaluated with the specific resin they are intended to color.

FD&C colors--also called lake pigments or food colorants--are composed of a dye chelated with an inorganic substrate, usually aluminum hydrate. Lake pigments typically contain only 20 to 40% pure dye and therefore must be added in sufficient quantity to achieve a satisfactory depth of color. Although lake pigments, being nontoxic, are excellent for food packaging applications, they are soluble in water and have a tendency to bleed out of some resins. FD&C red and yellow have been used successfully at about 0.5% weight loading. These pigments also contain between 20 and 30% moisture, and must be dried before use. Lot-to-lot variation in dye content requires the compounder to color match each new lot for critical colors; manufacturers specifying these pigments should always design in a generous tolerance to allow for this fluctuation.

Inorganic Pigments. Useful inorganic pigments include ultramarines and iron oxides. Available in several different shades of blue and violet, ultramarines offer better temperature resistance than do organics. However, they are relatively weak, and may also discolor in acidic conditions.

Iron oxides are essentially a form of rust that comes in black, brown, and red. They have good heat stability, and the clean, cosmetic grades have been used successfully in many resins. Tinting strength is moderate, but much lower than that of carbon black. Small amounts used in combination with other pigments help to achieve a wide range of colors. Higher loadings of iron oxides on dark colors tend to produce a matte finish in some resins.

As with all additives, device manufacturers should verify that pigment suppliers follow GMPs so as to ensure the good quality and low level of impurities critical for acceptance in medical compounds.


Depending on the base polymer family and the level of additives already contained in the starting resin, a compound may require additional ingredients such as heat and light stabilizers, antioxidants, processing aids, or lubricants. Other common additives include plasticizers, cross-linking agents, coupling agents, reinforcements, nucleating agents, and conductive additives.1

Heat Stabilizers. Typically required for PVC formulations, the addition of heat stabilizers is generally not necessary for other resins.

Antioxidants. Additional antioxidants can be important for resins that are very susceptible to degradation during processing or that lose properties rapidly upon aging. Urethanes, for example, often benefit from additional antioxidants. Careful antioxidant selection is required to prevent blooming to the surface due to incompatibility. For medical compounds, the toxicological properties of the additive are of primary concern: the lower the level of antioxidant used, the better, and as little as 0.3% or less of some antioxidants has been employed successfully.

Ultraviolet (UV) Stabilizers. It is sometimes necessary to give a compound longer-term protection against degradation in the presence of light. Elastomers containing polyether segments, for example, are susceptible to UV degradation and may require a UV additive. The selection process is similar to that for antioxidants.

Processing Aids. Processing aids are designed to impart internal or external lubrication to a compound. Types of processing aids include fatty-acid esters, such as glycerol monostearate; fatty-acid amides, such as bisstearamides; waxes; oxidized polyethylene; and others. A small percentage of fluoropolymer or silicone is also sometimes added as a processing aid. These ingredients can reduce process degradation, enhance mold-release action, aid in the dispersion of minor ingredients, and help produce a slicker surface on an extruded or molded part. Migration of these additives can be a concern, however, along with adverse effects on secondary processes such as bonding. Several nontoxic and FDA-compliant grades have been used successfully in medical applications.

Lubricants. Improvements in the surface lubricity and wear resistance of polymer compounds can be accomplished with lubricants such as PTFE or silicone. PTFE powder has a very low coefficient of friction, and its uniform incorporation into a base plastic enables the PTFE particles at the surface to provide lubrication. PTFE is typically added at between 2 and 20% by weight. Particle size is important and should be matched to the application: the surface of extrusions containing PTFE is sometimes rough, depending on the particle size used.

Silicone is a lubricious fluid, added at loadings of between 0.25 and 2% by weight, that migrates to the surface of a plastic and provides lubrication. Care must be taken to use the proper amount, as too much silicone can make a compound hard to feed into an extruder and may also give the component an oily feel.

Silicone and PTFE work well together, and some of the best frictional properties are obtained from this combination. Graphite and molybdenum powders have also been used (at between 0.2 and 0.5%) as lubricants, and have the effect of coloring a compound medium to dark gray. Molybdenum is added primarily to nylon, since it also nucleates (crystallizes) the resin surface, improving lubricity.


For some standard applications, incorporating fillers, colors, and other additives into a resin can be as easy as preblending all the ingredients together and loading them into a single-screw extruder hopper for strand pelletizing. As requirements become more stringent--the case for many medical projects--the compounding becomes more interesting, involved, and specialized.

Many types of compounding machines and mixers are available, and satisfactory compounds can be produced on most of them. While individual operator preference plays a major role, certain machines are more flexible and do a better job than others on specific compounds.

Melt Compounding. Procedures for melt compounding encompass two kinds of mixing, distributive and dispersive. For example, when fibers are incorporated into a resin, good distribution is required without breaking down the fiber segments: this is characteristic of distributive mixing, a relatively gentle homogenization of the material. Dispersive mixing is associated with the breaking up of agglomerates and generally correlates to high strain and shear rates.

Machines capable of both distributive and dispersive mixing include Banbury-type batch mixers, Farrel continuous mixers, and twin-screw extruders. The twin-screw extruder offers the most flexibility, since the barrel length and screw configurations are adjustable. Downstream oil injection and powder or fiber feeding is also possible.

Pelletizing. For materials with good melt strength that are not too soft and tacky, strand pelletizing is preferred. With this method, the pellets are cylinder-shaped and have a diameter of approximately 0.1 in. and a length of 0.125 in. Strand-pelletized compounds feed well on virtually all extruders. Soft and sticky compounds are usually underwater pelletized: the pellets are cut at the die face in a water chamber, conveyed in a pipe with the water, and then separated in a centrifugal dryer. Pellets are typically football shaped or round, with a diameter of 0.1 in. Micropellets with diameters as small as 0.02 in. can be produced by underwater pelletization, but are generally not recommended for medical compounds because the micropellet dies tend to restrict flow and feed rate, causing degradation in sensitive compounds. Small pellets of about 0.06-in. diam can be produced with a special die that prevents such degradation; these pellets will extrude on the smallest of machines.

Compound Development. The stages of a typical compound-development project will follow a flowchart similar to that shown in Figure 1. Once the need for a new compound is determined, a development team is assembled and the compound requirements are defined. A project plan should include enough time to research existing designs and databases and to determine starting formulations. Acquiring raw materials can necessitate a lead time as long as 2 to 4 weeks. If possible, it is advantageous to schedule compounding time in advance so that materials can be compounded soon after they arrive.

Processing and physical testing come next: several iterations may be required to fine-tune both the formula and the compounding process. When the ingredients are firmly established, biocompatibility testing can begin. Several lots of compound should be produced in order to isolate the effect of control variables such as melt flow and to determine the repeatability and capability of the compounding process.

Once it is verified that the compound meets specified requirements, a production prototype run should be made and final product fabricated. Final validation should include complete biocompatibility testing, physical testing, and field testing. If the validation is successful, a formal specification is written for the compound and a product code assigned.


Specialty plastic compounds are used in many medical devices for which a standard grade of material cannot meet all the specifications of an application. When developing a medical compound, it is always best to select ingredients and processes that have functioned well in the past in order to save time and money. The successful development of new specialty plastic compounds that satisfy the requirements of today's innovative medical designs represents a continual challenge for both device manufacturers and compounders.


1. Gachter R, and Muller H, Plastics Additives, 3rd ed, Munich, Germany, Carl Hanser Verlag, 1990. This source contains detailed descriptions of a wide range of additives.

Larry Acquarulo is president of Foster Corp. (Dayville, CT), a supplier of custom thermoplastics, elastomers, and blends. He holds a BS in plastics engineering from the University of Massachusetts Lowell and an MS in materials science from the University of Connecticut.

Predicting Cryogenic Impact Performance of Medical Containers from Resin Properties

Medical Plastics and Biomaterials Magazine | MPB Article Index

Originally published September 1996


In many cases, containers used in the medical industry are subjected simultaneously to subambient temperatures and mechanical stress. For example, flexible plastic containers used to harvest platelets from human whole blood must undergo centrifugation at 4°C and encounter high gravity forces. Blood plasma and many of the precompounded drugs are stored at ­30°C. Some biological products require packing and shipping in dry ice (at ­79°C) to ensure arrival in a biologically active state.1­4 The ultimate challenge is posed by cryogenic storage, handling, and retrieval in the presence of liquid nitrogen, which is employed at a temperature of ­196°C. For these cryogenic product applications, the impact properties of the containers are an extremely critical factor.

This study examined the cryogenic impact performance of four flexible PVC containers. The containers and their constituent films and resins were characterized by different methods and instruments. From a product development standpoint, it is highly desirable to be able to test raw materials or components in the laboratory and then accurately predict how the finished product will perform in an actual application. Accordingly, the objective of this study was to test the hypothesis that the cryogenic impact performance of medical containers can be correlated and predicted from the test results of the resins and corresponding films. The procedure followed was to compare and correlate the data generated by three different test methods: dynamic mechanical analysis of the PVC resins, instrumented impact testing of the films, and performance testing of the products.


Four plasticized PVC containers, designated S-1 to S-4, were studied. The S-1 and S-2 containers were formulated with an identical plasticizer and had very similar glass-transition temperatures (Tg). The other two containers were produced from two different, lower-Tg, resins. The material for container S-2 was sourced differently from the other three resins. Flexible films were obtained from the resins by extrusion, prior to production of the containers through a heat sealing process.

The glass-transition temperatures of the PVC resins were characterized at 10°C/min by a differential scanning calorimeter (a DuPont 910 DSC). The resins' storage modulus (E') and loss modulus (E") were studied at 3°C/min by a dynamic mechanical analyzer (a Seiko DMS-100 DMA) and multifrequency data were collected.

Impact energy of the films was determined by a computer-controlled instrumented impact tester (Dynatup Model 8200) at a film thickness of 0.035 cm and at temperatures between ­45° and 35°C. The samples were mounted on a 15-cm-diam aluminum testing frame and subjected to impact at 3.3 m/sec by a 2-cm semispherical tup.

Finally, containers filled with 300 ml of water were frozen at ­10°, ­20°, ­30°, ­40°, and ­50°C for 24 hours prior to drop testing. The drop-testing procedure entailed a free drop of the frozen containers from a height of 5 ft to the hard floor, whereupon the containers' cryogenic impact failure rates were recorded.


Resin Characteristics by DMA. The glass-transition temperatures of the four PVC resins--­17°, ­20°, ­30°, and ­40°C, respectively--are shown in Table I. The storage modulus (E') of the four resins is shown in Figure 1, where resins S-1and S-2 can be seen to have higher brittle-ductile (B/D) transition temperatures than resins S-3 and S-4. All of the resins' B/D transition temperatures were noted to parallel the order of Tg--except for resin S-2, whose modulus was slightly higher than that of S-1 at B/D range. (Tables and figures are not yet available on-line.)

The resin Tg values in Table I correspond quite well to both the E' (rigid) onset and the E" (1-Hz) primary peak temperatures measured at 1 Hz. The minor difference in Tg can be related to testing rate and sample preparation. The E" (100-Hz) primary peak at 100 Hz demonstrated a shift to a higher temperature. Also, the E" spectra in Figure 2 show that all four films have a pronounced secondary relaxation peak--for example, resin S-3 at ­80°C and S-4 at ­95°C. Each secondary-relaxation-peak temperature was far lower than its corresponding primary relaxation peak.

Film Characteristics by Instrumented Impact Testing. Figure 3 shows the impact-energy analysis of the four films after they were subjected to impact at different temperatures. Results show two separate groups of films with a wide gap between them, even though no such obvious contrast appears in the storage-modulus data shown in Figure 1. This difference can again be related to the testing-rate difference between the dynamic mechanical testing and the instrumented impact testing.

If we examine the temperature range between ­50° and 0°C in Figure 3, it shows that the impact B/D temperatures of the films S-1 and S-2 occur at about ­15°C, while those of films S-3 and S-4 occur at about ­30°C and ­35°C, respectively. These B/D temperatures corresponded fairly closely to both the resin DSC Tg and the DMA E' (rigid) onset at 1 Hz and E" (1-Hz) primary-peak temperatures.

Container Characteristics by Drop Testing. The cryogenic impact failure rate of containers at different drop- testing temperatures is presented in Figure 4, which shows the temperatures at which the containers would have 100% and 0% survival rates. The data suggest that every container at the temperature range studied had two different failure-rate correlations with temperature. For example, for the S-4 container, the change in failure rate takes place in the neighborhood of ­25°C; the two correlations would meet and form an intersection point. Similar correlations were also found for the other three containers.

The drop-testing results indicate that a sharp increase in the container failure rate exists prior to the intersection point. After that point, the failure rate only increases slowly with a decrease in the testing temperature. The intersection temperatures of containers S-1 to S-4 were at ­7°, ­22°, ­14°, and ­25°C, respectively, as shown in Table II. Surprisingly, all of these intersection points--except that of S-2--match the E" (100-Hz) peak temperatures quite well. This suggests that the container failure rate is related to the loss modulus E" (100-Hz) of the films. The discrepancy with regard to S-2 is expected, since resin S-2 was sourced differently than were the other three resins.

Correlation of Resins, Films, and Containers. Figure 5 shows the container failure rate for the drop testing conducted at ­20°C as a function of the glass-transition temperature. The results indicate clearly that the primary variable for the cryogenic impact performance of the containers is the location of Tg. However, the data at a single temperature gave insufficient information for researchers to predict the entire temperature range over which the medical container could be expected to survive. Film impact data over a wide temperature range was therefore examined.

The impact energy of S-1 film in Figure 3 indicates a B/D transition at about ­15°C. This is noted as the first upturn in impact energy, and is essentially equivalent to the resin Tg. In most cases, the location of the B/D transition was also confirmed by visual and scanning electron microscopic examination of the fracture surfaces: at temperatures below the B/D transition, jagged, glass-brittle morphologies were generally observed, whereas above the B/D transition, ductile, high-elongation morphologies appeared. Similarly, the ultimate displacements that the samples sustained before failure could also serve as independent confirmations of this transition.

In addition, S-1 film in Figure 3 shows a significant change in impact energy at ­7°C, which was essentially equivalent to the intersection point of container S-1 in Figure 4. This suggests that the change in the film's impact energy is responsible for the variation in temperature dependency of the container failure rate. When the temperature was increased to about 8°C, a second upturn was observed in impact energy. In the vicinity of 8°C, Figure 4 shows that container S-1 demonstrated 100% survival. This multiple or stepwise increase in impact energy had been previously reported for multiphase materials.5 Recent data, however, indicate that this is a very common phenomenon for polymers with pronounced secondary transition. On the other hand, film S-1 at ­40°C essentially held no strength because its impact energy was nearly equal to zero (see Figure 3). This ­40°C temperature was found to correspond essentially to the 100% container failure seen in Figure 4.

Predicting Cryogenic Impact Performance. Based on the preceding analysis, the resin DSC Tg values were equivalent to the DMA E' (rigid) onset and E" (primary) peak temperatures at 1 Hz. These temperatures corresponded fairly closely to the B/D temperatures of films from the impact-energy analysis. In addition, the intersection point at which container failure rate changed corresponded surprisingly well to the E" (100-Hz) peak temperatures. Finally, the container S-1 cryogenic impact failure at 0% and 100% can be correlated to and predicted from the results of the lab testing of its corresponding S-1 resin and film.

Other Factors Affecting Product Performance. Apart from the effects of material properties, the functioning of any medical device can be influenced by product design and processing. Accordingly, it is critical to simultaneously consider material, design, and processing at every phase of product development in order to optimize medical container performance.


This study indicates that the data generated by the dynamic mechanical analysis of the PVC resins, instrumented impact testing of the films, and performance testing of the containers can be correlated. The results of the study support the hypothesis that the cryogenic impact performance of medical containers can be correlated to and predicted from the outcome of lab testing of the corresponding PVC resins and films.


1. Woo L, Westphal S, Shang SW, et al., "Relating Dynamic Mechanical Data to Flexible PVC Low-Temperature Performance," in Proceedings of the 24th North American Thermal Analysis Society, Baltimore, NATAS, p 171, 1995.

2. Woo L, and Ling TK, "Cryogenic Impact Properties of Medical Packaging Films," in Society of Plastics Engineers, Inc., Technical Papers, vol XXXVI (ANTEC 90), Brookfield, CT, Society of Plastics Engineers, p 1116, 1990.

3. Woo L, and Ling TK, "TPE and Subambient Behavior of Flexible PVC," Vinyl Tech, 12(4):198, 1990.

4. Woo L, Westphal S, and Ling TK, "Dynamic Mechanical Analysis and Its Relationship to Impact Transitions," Polymer Eng Sci, 34(5):420, 1994.

5. Bucknall CB, Toughened Plastics, London, Applied Science, p 298, 1977.

Sherwin Shang, PhD, is program manager at the Fenwal Division, Biotech Group of Baxter Healthcare (Round Lake, IL). Michael T.K. Ling is senior engineering specialist at Baxter's Medical Materials Technical Center (MMTC), where Stanley P. Westphal is a research scientist and Lecon Woo, PhD, is the Baxter Distinguished Scientist. Author specializations range from biomedical product/process development to mechanical/physical analysis, polymer morphology/rheology, and polymer processing.

IP Is Hip: What Makes Patents So Hot?

Medical Device & Diagnostic Industry Magazine | MDDI Article Index

Originally published September 1996

Reading through a raft of quarterly earnings reports from medical device companies recently, I noticed a curious trend: seemingly every report mentioned patent litigation expenses. Some even cited patent suits as the key events of the year, as in this somewhat dire statement by one company president: "Although there can be no assurances, we anticipate a successful conclusion of the patent litigation before the end of the year. To the extent that the goal of this litigation was to intimidate competition by imposing enormous costs, it has only been partially successful. We have indeed incurred heavy expense in the course of defending this action. We have not, however, been intimidated."

Now, while MD&DI has certainly covered patent issues frequently in the past, I must confess that they have not been prominent on my editorial radar screen. So, spurred on by this evidence, I resorted to my favorite research tool, the Internet. There I discovered that patents--and the larger area of intellectual property (IP to the hip)--are indeed a hot topic.

As a July 1 article in the Washington Post put it, "the boom in such technological fields as biotech, computers, and medicine, and the rise of ideas and inventions as major assets in the global economy, have made intellectual property one of today's hottest legal specialties." With companies spending billions on innovative products, the article adds, they don't hesitate to spend large amounts defending patents.

For the medical device industry, patent law is becoming as important as regulatory affairs. But because this industry is defined largely by its innovations, many of which build on previous innovations, patents can be both a boon and a burden. While companies profit from patent protection on their own products, they can also be threatened by other companies that may or may not have a good case for patent infringement. Frankly, I'm not sure whether this situation is good or bad. Opposing arguments that current patent law protects competition or hurts it can both sound convincing.

Naturally, no trend this hot or complicated could escape the attention of Capitol Hill. At least five major pieces of legislation have come and gone before the 104th Congress. The bills involved would have effected a variety of significant changes to patent law, from forcing patent applications to be published 18 months after filing (currently they are not made public until the patent is issued), to changing from the current first-to-file system to a first-to-invent system. The House reportedly agreed on July 22 that no action would be taken on any of these bills in 1996. But the issue will certainly resurface early in the next session.

The complexities of the questions involved prohibit much discussion here. But their importance to the medical device industry is underlined by the testimony before Congress last fall of Raymond Damadian, president and chairman of Fonar, an MRI pioneer: "This web of patent legislation is a carefully orchestrated design to render the U.S. patent useless, and destroy whatever potential it may have left to initiate competitive new business enterprises and new employment for the people of America."

Many in the device industry may well disagree with Damadian's opposition to the patent legislation. But many more, I suspect, may not even be paying the issue much attention. My own few hours of research have made it clear to me, however, that intellectual property law is becoming a critical area for the industry. If you're a bit hazy on the issue, I strongly recommend a quick refresher course, either on the Internet or in a library. Whether because of litigation or changes in legislation, an understanding of patent law can indeed mean the difference between success and failure for a company and all its employees.

John Bethune

Government Labs Offer R&D Resources

Medical Device & Diagnostic Industry Magazine | MDDI Article Index

Originally published September 1996

Technology Transfer

At a time when companies are looking for any help they can get to cut expenses, the National Institutes of Health (NIH) is making an offer few can--or at least should--refuse. Through its Office of Technology Transfer (OTT), NIH is awarding licenses to private firms to develop and sell technologies invented by scientists and engineers working in laboratories affiliated with the more than 20 institutes on the Bethesda, MD, campus, as well as at FDA. Each of the institutes also offers cooperative research and development agreements (CRADAs) that enable private and government researchers to work together on specific projects that might produce commercial products.

The idea behind the NIH technology transfer program is to promote public health by making available to private companies both ideas that have commercial potential and the expertise to bring good ideas to fruition. "The government cannot be a manufacturer of many types of inventions, particularly ones that require regulatory approval," notes Jack Spiegel, director of the division of technology development and transfer at OTT (Rockville, MD). Consequently, without some technology transfer mechanism, ideas arising from research done on the NIH campus or at FDA--ideas that might evolve into profitable and beneficial health products--would be wasted.

In other NIH programs, technologies developed with public funding but at extramural laboratories--for example, medical centers and universities--are controlled by those institutions. "The extramural program is by far the largest part of the NIH budget, but federal law calls for the institutions that receive those grants to have the first option to commercialize their inventions," says Steve Ferguson, a senior specialist at OTT.

By contrast, the technology transfer program awards licenses only on technologies invented at federal laboratories. There is no typical company that takes advantage of these licensing opportunities. Rather, a company obtains a license to develop a specific invention on the basis of its ability to turn it into a commercial product. "We have licensees representing all sizes of companies," Ferguson says. "We do not begin with a presumption that only big companies or only small companies can develop a technology. We look more at the specifics of a particular technology--what is needed and whether the company has the resources and interest to take it all the way through product development and commercialization."

Participating companies discover the OTT program through many routes. Individual companies may learn about inventions developed at NIH or FDA laboratories from the scientists who came up with them, for example, or from presentations at scientific conferences. "And NIH is a great facility for training people, so a lot of people come in and work on technologies and then leave to work in private industry or start their own companies--so word gets out about various developments that way," Ferguson says. OTT also disseminates information about inventions available for licensing through its web site ( as well as through various on-line and specialized publications, such as Knowledge Express, the National Technology Transfer Center, and the Biotechnology Information Institute.

The hundreds of new inventions that come out of NIH and FDA each year run the gamut from diagnostic reagents to ventilators, from nuclear medicine scanners to hand protectors. Among those now available for licensing is a ventilator that not only encourages patient-initiated breathing but decreases the chances of sickness and death while increasing the level of patient comfort. The device features an intratracheal pulmonary ventilation device and a low- resistance, flexible endotracheal tube, both developed by scientists at the National Heart, Lung, and Blood Institute. Other devices available for licensing include a handguard to protect against accidental needle sticks and a method for creating a barrier on latex products such as gloves and condoms that is impenetrable by even the smallest viral particle.

Inventions developed in government laboratories are first evaluated by OTT for patentability and commercial potential. The office submits many, but not all, commercially viable technologies for patenting. "We seek patent protection for technologies that require it as an incentive for the private sector to adopt them," Spiegel says. "But there are other inventions and technologies developed here for which private sector partners don't necessarily require intellectual property protection in order to motivate them to use or to develop the technology." When patents are sought, the office usually tries to obtain international coverage, because many companies desire worldwide patent protection before they enter foreign markets.

OTT can award any in a range of licenses, including those that allow the company to evaluate a technology for its commercial potential, to use an invention only for internal applications, or to commercialize a technology on a nonexclusive or exclusive basis. "There is an inherent preference in governmental licensing for nonexclusive agreements that promote competition and encourage the widest possible use of a particular technology," Ferguson says. "There is also a preference for field-of-use licensing, where two companies license the same basic technology but for different applications. That allows us to grant each company a separate license for the field of use it's interested in."

Once the licenses are granted, NIH wants to see the company progress toward developing the invention. Licensing agreements typically call for very specific actions by the company. Nonuse or failure to develop the invention are each grounds for revoking the license. Also, because the early R&D was accomplished with taxpayers' money, licensing agreements usually require that the products resulting from the invention be manufactured in the United States.

But licensing is only one way companies can leverage the expertise of researchers working for the federal government. Another involves joint research projects developed under the Federal Technology Transfer Act of 1986, which created CRADAs. "Sometimes a company's staff is working on an idea and it needs some expertise or resources that it doesn't have, but we do; sometimes it's the other way around," explains Ferguson. "Basically, CRADAs work because there is mutual interest and mutual benefit."

These R&D agreements make available to private companies not only government research staff but laboratory facilities, materials, equipment, and supplies. They also provide the private company in advance with an option to negotiate exclusive licenses on inventions made under the research agreement.

Regardless of the path chosen--either CRADAs or licensing of an existing invention--private companies have much to gain by taking advantage of research done at government laboratories. "To the extent that companies can use our basic discoveries to match up with their developmental programs, they can exert a considerable amount of leverage in terms of new product development," Ferguson says.--Greg Freiherr

FDA Launches Pilot Program for Third-Party Device Review

Medical Device & Diagnostic Industry Magazine | MDDI Article Index

Originally published September 1996

Product Approvals

Seven companies have been selected to participate in FDA's pilot program for the third-party review of selected 510(k) premarket notifications. First proposed by FDA's Center for Devices and Radiological Health in April 1995, the program took nearly a year and a half to design and implement. Operation of the two-year program began on August 1.

The third-party review organizations appointed by FDA (see list on page 22) range from large independent testing labs to European Union (EU) notified bodies to small firms that were once primarily consulting firms. According to John Stigi, director of FDA's Division of Small Manufacturers Assistance (DSMA), the agency expects that the seven companies will be able to handle reviews for all the devices selected for inclusion in the program. Despite industry requests to broaden the range of eligible devices, FDA elected to include all Class I devices not exempt from 510(k) requirements, but only 29 Class II devices. After conducting their reviews the third parties will submit their recommendations to FDA, which has 30 days to render a final decision.

Speaking at the Biomedical Focus Conference and Exposition in Minneapolis in July, Stigi said that FDA received a total of 37 applications for the third-party review program, 28 of which were ultimately determined to be reviewable. In addition to the seven companies appointed, he said that another 15 or 16 probably would have been capable of managing reviews, but were weaker in some respects than those selected. "Many entities that submitted applications did not put their best foot forward, but our hands were tied by the information they actually submitted. We couldn't consider what we intuitively knew about the capabilities of an organization."

The primary determining factors for recognition, Stigi explained, were the qualifications of the third-party review personnel and their freedom from conflict of interest. The agency signaled early on that it considered conflicts of interest to be a major obstacle. The April 3, 1996, Federal Register notice calling for applicants described the most common conditions that would indicate a conflict of interest as:

* The third party being owned, operated, or controlled by a medical device manufacturer or distributor.

* The third party or any of its review staff having any ownership or other financial interest in a medical device manufacturer or distributor.

* The third party being involved in the design, manufacture, or distribution of any medical device.

* The third party consulting for any medical device manufacturer or distributor regarding any medical device.

* The third party participating in the preparation of any 510(k).

* A fee charged or accepted by the third party being contingent on the type of recommendation made by the third party.

In order to participate in the program, some of the recognized third parties had to modify their operations significantly. For instance, Nancy Sauer, president of RDD Consultants, Inc. (Louisville, CO), saw the decision to become a third-party reviewer as an either/or decision. "Either we continued as a consulting firm or we made the transition," she says. RDD Consultants has ceased all consulting and now focuses entirely on its duties as a third-party reviewer.

TÜV Product Service (New Brighton, MN) also terminated its medical device consulting activities. Gene Panger, director of sales and marketing for North America, explains that TÜV has also changed its policy on employee-owned stock to accommodate FDA. Employees whose job function is to review medical devices are no longer permitted to own stock in any medical device companies.

According to Peter Walker, scheme manager, Medical Devices Group, British Standards Institution Product Services (BSI; Milton Keynes, Buckinghamshire, UK) did not previously undertake consultancy, so there was no conflict of interest. However, the company will follow the guidelines described in Standards of Ethical Conduct for Employees of the Executive Branch, a final regulation issued by the U.S. Office of Government Ethics, in its device review operations. This document was provided by FDA as part of the information packet distributed to all potential third parties.

Another important element in selecting third-party reviewers was the type of expertise applicants demonstrated. According to Alfred Bracey, associate director of DSMA, FDA eliminated some applicants because they didn't have the expertise needed to review the devices they listed on their applications. Other companies were eliminated simply because their applications showed they could review only a limited number of devices. The agency also rejected any firm that could not communicate with it electronically by both modem and fax. During the program, FDA will not review applications from additional companies wishing to become third-party reviewers.

In order to offer the expertise needed to review 510(k)s, some of the third parties have forged partnerships with other firms. For example, CITECH (Plymouth Meeting, PA) shares a facility with ECRI, an organization that has performed testing and technology assessment for users of medical devices for over 20 years. "While CITECH is the FDA-recognized third-party reviewer," president Robert Mosenkis explains, "most of the detailed review activities will be done by ECRI." CITECH will act as the liaison between manufacturers and FDA.

Similarly, BSI has formed a partnership with the National Institute for Biological Standards and Control (Potters Bar, UK), which will handle the review of biological materials and diagnostic reagents.

It is unclear whether or not experience as a third-party reviewer in other coun- tries weighed in an applicant's favor during the selection process. Two of the FDA-recognized third parties act as third parties in other countries. BSI is an EU notified body under most European Council directives, including the Medical Devices Directive and the Active Implantable Medical Devices Directive. TÜV is also registered as an EU notified body for these same directives and recognized as an in-country caretaker in Japan. Sales and marketing director Panger sees the firm's new role as a third-party reviewer as complementary to its existing international roles. "TÜV can now offer market access services for Europe, Japan, and the United States," he explains. "Our overseas clients will view us as an alternate route to getting FDA clearance."

Another third-party reviewer that has had experience in reviewing medical devices is the California Department of Health Services (Sacramento), the only public agency recognized as a third-party reviewer. Prior to enactment of the Medical Device Amendments of 1976, the department had a device approval program under state legislation passed in 1963. After FDA took over the regulation of devices, the department focused on inspecting and licensing medical device manufacturing facilities in California and on its role in protecting public health. According to James Barquest, acting chief, Medical Device Safety Section, the department still retained statutory authority to review and approve medical devices, but did not continue such activities because it became necessary for manufacturers to obtain federal approval to market their devices. The department will now accept 510(k) submissions only from California companies.

When the two-year pilot program is completed, FDA will measure its success a number of ways. The Federal Register notice described a successful program as one offering device manufacturers an "alternative review process that can yield more rapid 510(k) decisions." DSMA's Bracey adds that FDA will also look at whether the program has "significantly reduced the workload of FDA reviewers." In the end, Bracey says, continuation of the third-party reviewer program will depend heavily on whether "manufacturers make use of it."

At the Biomedical Focus conference, Stigi was asked whether he believed the device industry would ultimately find the program useful. In response, he noted that FDA review times for 510(k) notifications have been falling throughout the past year, and are currently running at just over 100 days. Some third parties have said they will complete their reviews in 45 days or less, but they are not required to meet any deadlines, Stigi added. On the other hand, FDA's goal is to render decisions on completed applications within 30 days. "That would give manufacturers a total of 75 days, compared to the 100 days FDA is currently taking for the review," he said. "Considering that manufacturers will have to pay for a third-party review, it remains to be seen whether they will consider the gain of 25 days worth the added expense."

Leonard Frier, president of MET Laboratories (Baltimore), believes that the program can benefit both industry and FDA. "FDA can spend more of its time reviewing PMAs [premarket approvals], IDEs [investigational device exemptions], and Class III devices, and thereby reduce its backlog," he says. "In the past, it has been extremely expensive for manufacturers to develop a product and then have to wait because of this backlog. Now there is a solution."

And, argues RDD's Sauer, some third parties do guarantee speed. Her firm, for instance, plans to guarantee complete reviews of in vitro diagnostic products in 35 days or less. Frier anticipates that most reviews will take about 30 days, unless there are complications with the submissions.

Whether the pilot program is a success or failure, the process for reviewing medical devices will certainly be different in two years. If the program fails to interest device manufacturers, FDA could argue that its way of conducting reviews is the only viable method for the United States. If the program succeeds in attracting some manufactur- ers, FDA could characterize the payment of fees to third parties as justification for FDA user fees. And, if dozens of manufacturers choose to send their 510(k) notifications to third parties, we may see the beginning of a new third-party review system in the United States.--Daphne Allen

Bar Code Standards: For Medical Products, More Work Is Needed

Medical Device & Diagnostic Industry Magazine | MDDI Article Index

Originally published September 1996

Walter W. Mosher

President, Precision Dynamics Corp., San Fernando, CA, and

Chairman, Health Industry Business Communications Council, Phoenix, AZ

For manufacturers of medical devices, the pressure to reduce the costs relating to their products has never been greater than it is today. Companies are actively exploring every place in the product life cycle where economies can be found, including areas related to the distribution of their products.

Bar coding systems have long been touted as a significant means for gaining such economies. But despite more than a decade of work by a number of organizations in the United States and abroad, the medical manufacturing industry has so far been unable to implement an acceptable universal standard for bar code labeling of medical products. Greater progress in this field is desperately needed, but manufacturers should first understand what has come before and what obstacles will need to be overcome.


In the early 1970s, the Food Marketing Institute and the Grocery Manufacturers Association formed the Uniform Code Coun-cil (UCC) to develop and promote the use of bar code labels in the food distribution industry. The labeling standard that UCC developed was both a data and a bar code structure known as the Universal Product Code (UPC) symbol. This symbol is now used throughout the United States to identify the items sold in most major retail outlets.

It was not until the early 1980s that the health-care industry began to focus on automating data acquisition activities. The initial efforts were led by several distributors, including American Hospital Supply Corp. These distributors proposed that the UPC symbol be the standard for labeling medical products, but this recommendation did not take into account how using the data structure would affect medical product manufacturers. Significant opposition to labeling medical products with the UPC symbol developed quickly.

In an attempt to resolve the debate on medical product bar code structure, in 1983 the American Hospital Association hosted a meeting of interested distributors, manufacturers, and end-users. The meeting led to creation of an ad hoc committee consisting of manufacturers, distributors, providers, and technical experts charged with assessing the labeling needs of the medical manufacturing industry and its distribution partners. The committee, called the Health Industry Bar Code Council (HIBCC) and later renamed the Health Industry Business Communications Council, was sponsored by the Health Industry Manufacturers Association (HIMA), the Health Industry Distributors Association (HIDA), the Pharmaceutical Manufacturers Association, the National Wholesale Druggists' Association, and the American Hospital Association.

In 1984, HIBCC produced the first health industry supplier labeling standard and, subsequently, the first health industry provider labeling standard. These standards enabled manufacturers to include the National Health Related Industry Code (NHRIC) and the National Drug Code (NDC) numbers assigned by FDA and used by the pharmaceutical industry. They also recognized the need of the medical manufacturing industry to encode product identification using alphanumeric symbols and product numbers of greater length than the UPC symbol, which can accommodate a product identification segment of only five digits. Use of the HIBCC standards was expanded to foreign countries in 1988, when the European HIBCC was formed and began promulgating the standards throughout Europe.

Meanwhile, the UPC data structure was becoming more widely used in retail distribution, and other industries began to see the benefit of automated data capture techniques based on bar codes. In 1988, UCC and the International Article Numbering Association (EAN) began to develop a global standard for product identification in the retail, pharmaceutical, and medical product industries. Unfortunately, the initial attempts focused on using the UPC 12-digit manufacturing product number and progress was slow. In 1991, UCC established an Industrial Commercial Advisory Committee to study ways that the UPC standard could be used to meet the needs of additional industries. Today, that committee has significant influence over the development of coding standards and applications for many industries.

It is unfortunate that when HIBCC was formed in 1983, UCC had not yet expanded its charter beyond the grocery and general merchandise markets. As a result, the HIBCC standards were developed along wholly independent lines, without UCC's assistance or input. Inevitably, the HIBCC supplier standard was not compatible with that developed by UCC and EAN, leaving manufacturers with a difficult choice to make in selecting an appropriate coding structure.


In an attempt to end the confusion over the two conflicting standards, HIBCC and UCC joined forces in 1994 to revise the HIBCC supplier standard so that it could make use of the application identifiers (AIs) of the UCC/EAN standard for specifying manufacturer product number and secondary information. AIs had been introduced by UCC in 1988 to identify product attributes such as lot, batch, serial number, weight, and size. The AI(01) was developed to allow the UPC data structure to be encoded in symbologies other than the UCC/EAN format. The revised HIBCC standard called for concatenating the AI(01) with the AI(240).

As part of the 1994 revision, HIBCC suggested that the AI(01) could be filled with zeros and the entire product specification could be located in the AI(240) data structure. Other modifications to the standard included deleting the alternate method of identifying products using embedded NHRIC or NDC code numbers in an HIBCC-format label. UCC provided a means in the AI(01) for directly representing this information. Another change to both the HIBCC and UCC standards was to make the packaging-level indicator portion of the data structures identical in both formats.

While UCC and HIBCC were finding ways to work toward harmonization, however, other forces were exerting influences in the opposite direction. Lack of cooperation between EAN and the European HIBCC, for instance, had reduced the motivation of European medical manufacturers to bar code their products. UCC intervened in 1995 and was helpful in creating an agreement between the two organizations on how to mark products for the European market. This pact endorses the UCC/EAN standard as the primary product-marking methodology, but also allows for use of the HIBCC standard. It encourages migration to the UCC/EAN approach when possible.

In 1995, the U.S. Department of Defense (DOD) adopted a Universal Product Numbering (UPN) standard for all products sold to the government. Although HIBCC proposed the use of AI(240) with zero-filled AI(01) as a structure for the UPN database, DOD decided that this data structure was too long for manual data entry during the government's transition from manual to automated material management. The scheme finally approved by DOD calls for the use of either the UCC/EAN or HIBCC method of product identification and requires that all products sold to DOD after July 1996 be marked with the UPN.

So far, neither HIMA nor HIBCC has endorsed migration to the UCC/EAN standard. UCC suggests that migration might be appropriate in certain instances, but supports both the UCC/EAN and HIBCC standards. HIDA endorses the use of either the HIBCC or UCC/EAN format and migration to the UCC/EAN standard when possible. Thus, at the present time, it is correct to use either the UCC/EAN or HIBCC standard. These standards will be accepted worldwide based on the agreement between the UCC/EAN and the European HIBCC, and will also be satisfactory for the DOD and many other government agencies, such as California's Medi-Cal.

A complicating factor for bar code choice, however, is that some European governments may be planning to require the use of the numeric UCC/EAN format, even though this requirement is not in conformity with the agreement between UCC/EAN and the European HIBCC or with the stated positions of any U.S. organizations. Therefore, the decision whether to change a product's labeling to the UCC/EAN standard must take into account where the product will be marketed.


The most important problem with migrating to the UCC/EAN data structure is the substantial cost to manufacturers of changing their product identification numbers to a five-digit format. This change would require modifying all product literature, packaging, and catalogs as well as both the manufacturer and customer databases. Many companies assign important information to all the digits of their product identification numbers.

Another problem is that medical practitioners may be accustomed to asking for particular products by product numbers, and changing these could put patients at risk. Similarly, considerable scientific research has been conducted and reported using product identification numbers. Changing the numbers could make such previously published studies useless.

One proposal for making universal standardization less difficult is to use the zero-filled AI(01), as suggested by HIBCC in the 1994 revision of its standard. In 1995, however, UCC evaluated the impact of this approach on the UCC/EAN standard and concluded that the method was not consistent with the published definition of AI(01). UCC therefore requested that it be deleted from the HIBCC standard. Fortunately, an HIBCC survey of manufacturers found only two respondents that were considering using the AI(240) with zero-filled AI(01), and none that had yet begun to do so.

Another idea is to assign an alphanumeric AI for primary data. This AI would consist of 20 characters and would include a 4-digit manufacturer ID identical to the HIBCC Labeler Identification Code, a 13-character product ID, and space for other information that is contained in the HIBCC primary label. This plan would simply incorporate the entire HIBCC primary label, and would not require any changes in product identification or the creation of cross-reference tables.

UCC recently proposed that manufacturers use cross-reference tables to relate each manufacturer product number to a five-digit number in the AI(01) and then concatenate this with the AI(240), which would contain the existing alphanumeric catalog numbers. This would require that both manufacturers and customers modify their databases, and would undoubtedly be both costly and confusing. For instance, one major manufacturer of medical products surveyed its organization and discovered that among its 20,000 products bar coded using the HIBCC data structure, more than 10,000 have product codes longer than five digits, and more than 8000 have alphanumeric product codes. Also, many of the company's divisions include important information in their product numbers.

Other approaches to UCC/EAN migration problems may be forthcoming, as HIBCC, UCC, and HIMA continue to work toward a solution that will accommodate all segments of the health-care industry.


Another area of concern for medical product labeling is marking small products, such as unit-dose packages that have inadequate space even for a bar code as short as the UPC symbol. HIBCC proposed use of a two-dimensional symbology, but the method was not adequately supported by the manufacturers of bar code readers. UCC has begun to work toward adoption of 2-D symbologies by all industries, and may issue recommendations for this symbology soon.

In the interim, HIBCC has proposed that the UCC/EAN 14-digit number be truncated to the 10-digit NHRIC and NDC numbers. This plan could create confusion, because there would be no way to indicate that the 10 digits represent NHRIC and NDC numbers. Therefore, imported products could have the same values for these 10 digits. HIBCC intends to modify its standards when an appropriate 2-D symbology is chosen and adequately supported by the bar code reader industry.


With the worldwide acceptance of both the HIBCC and the UCC/EAN standards, there is no reason to delay bar code labeling of medical products. New products, or products that do not have meaningful information in their product identification numbers, should be coded using the UCC/EAN data structure as soon as possible.

For other products, however, there is no immediate need to migrate from the HIBCC to the UCC/EAN format--and there are some reasons not to do so. The cost associated with forcing all medical product manufacturers to conform to a five-digit product code is excessive and unnecessary, and not in keeping with the present societal emphasis on reducing the cost of health care. It would also create confusion.

Actually, a longer alphanumeric symbol for product identification could be easily accommodated. If UCC and EAN would issue an AI of 20 alphanumeric characters, the HIBCC data structure could be embedded in it at minimal cost. The result would be a single worldwide standard for health-care product identification. If a manufacturer now using the UCC/EAN structure did not wish to modify its database, it could simply ignore any extra information contained in the new AI.

Most grocery and retail databases in the United States are already obsolete and do not recognize all digits of the UCC/EAN data structures, a situation that creates the potential for misidentification of products. As international trade increases, misidentification will become an even more significant problem. These systems must be redesigned anyway, and when they are, a 20-character alphanumeric primary data structure could be provided at little extra expense.

Unless UCC modifies its restrictive 14-digit numeric format, wholesale migration to its proposed standard would not benefit the medical manufacturing industry and would cost it a great deal. More work is needed to ensure that the medical product manufacturing and distributing communities can have a coding standard that really meets their needs, wherever in the world it might be used.

Biomaterials Research in the 1990s

Medical Device & Diagnostic Industry Magazine | MDDI Article Index

Originally published September 1996

An Interview with Jack Lemons

Professor of Biomaterials and Surgery, University of Alabama, Birmingham, and Chairman, American Society for Testing and Materials Committee F-4

In the field of biomaterials, attaining adequate support for research and development has rarely been easy. With its experimental beginnings some 50 years ago, the biomaterials industry went from being relatively low-profile to being anything but that in the 1980s, when the industry benefited from an influx of R&D dollars from high-tech companies.

But in the 1990s, public controversies such as those related to the use of silicone-gel breast implants have caused many suppliers to withdraw their products from the medical market rather than risk financial liability and class action litigation. Removal of these materials has led to the increased need for some sort of legislative relief so that biomaterials suppliers will go back to producing and improving materials for medical applications. One recent attempt at such relief fell short last May, when President Clinton vetoed the Common Sense Product Liability Legal Reform Act, of which biomaterials reform was a part.

To address this and other issues surrounding the biomaterials crisis, MD&DI recently spoke with Jack Lemons, chairman of the American Society for Testing and Materials (ASTM) Committee F-4 on medical and surgical materials and devices, at the World Biomaterials Congress in Toronto. While his experience in the medical industry has taken him from university classrooms to professional organizations and standards-writing committees, Lemons has kept his focus on the biocompatibility of synthetic materials for tissue replacements in surgery. In doing so, he has closely followed the growth--and current crisis--in biomaterials research and development in the United States.

What's your impression of the so-called biomaterials crisis, in regard both to the commercialization of devices and to the effect of the crisis on research?

I've followed biomaterials and the evolution of biomaterials and biomechanics from probably every aspect. What's key is that during this evolution, the United States really took a leadership position throughout the world. Unfortunately, that's changing rapidly. The U.S. medical device industry is moving offshore, to Asia and Europe principally, and many factors have caused this to happen. The biomaterials crisis is real. One of the more critical factors that have led to this crisis is the liability issue. Industry in the United States is often put in a no-win situation with regard to investment and return on investment. While these companies need to be safe and efficacious in everything they do, they also need to be able to earn a profit.

Is the biomaterials crisis having an effect on academic research and research programs in general?

Yes, and it's a continuing subject of both professional presentations and private discussions among investigators. It's certainly of grave concern to students who are going through formal academic programs and looking toward a future in this field.

I recently conducted a summer seminar for graduate students in which I reviewed what I've seen happen within past decades, and I tried to make projections about the future. This is the first decade in which it's been very difficult to make projections, simply because everything is changing. So the crisis is real, and though it only comes into focus during certain discussions, it is a real part of what's going on in U.S. universities. Worldwide, research continues, but U.S. funding has decreased significantly, and everyone can feel the effects of change.

How has the relationship between the U.S. device industry and its academic counterparts changed over the past decade?

For many years I've been involved in the development of sources for graduate student and program funding. I've spent a significant portion of the last 20 years interacting with the major U.S. industries, attempting to move forward from where we were in former years, when materials that were intended for other applications were adopted as biomaterials. Such materials were successfully used during those years, but they were originally formulated for other purposes, so we redirected, reconstituted, or surface-modified them.

During the 1980s, companies like DuPont, Dow Corning, Bristol Meyers, and 3M--major sources of high technology--became involved, and researchers were finally constituting, testing, and evaluating new substances as synthetic replacements for tissue. But because of liability considerations and the potential for class action litigation, larger companies have changed and they are no longer becoming involved. Instead, they've backed out of major investment, and unfortunately it's costing everyone significantly.

Certainly there are programs that were started more than five years ago that still exist today, and most people in industry would agree that such programs are important in terms of ensuring the quality of research that can be conducted in the United States. But due in part to the current scarcity of these types of programs, there's no question that U.S. device manufacturers now view moving parts of their companies overseas as a necessary part of doing business. We're going to see the financial and structural effects of that throughout the rest of the 1990s. So while the impact of these moves is not a big factor yet, it certainly will be down the line.

In the future, the need for biomaterials is going to be greater, because we're now talking about combining synthetic materials as matrices for drug delivery and containment, and there's going to be a need for millions of joints, heart valves, tooth-root replacements, and so on. And with an aging population, this need for biomaterials is going to become more urgent. But the high-technology resources that could have been applied to improving what we have today are being compromised. Industry needs to rethink the equation somehow, and President Clinton's veto of the Common Sense Product Liability Legal Reform Act puts this in focus. Companies will no longer be able to make these investments unless they're given incentives to do so.

What's the mission of ASTM Committee F-4 on materials and medical devices?

The primary mission is to develop standards that gain a consensus from all parties involved in using biomaterials--including general-interest groups, industry, patients, government, and regulatory bodies. We convene twice a year, with the goal of determining what standards need to be written. That then leads to the development of accepted standards through a controlled and balanced consensus process that is truly representative. Ultimately, those standards are published as American National Standards, which give those in the discipline a way to ensure a material's properties, characteristics, sizes, dimensions, and tolerances.

Many ASTM Committee F-4 members also participate in technical advisory groups (TAGs) that, in representing their countries' interests, contribute to the globalization of industry. The ultimate goal of the committee is to help businesses become more international, because medical care should not have boundaries.

How do you decide when a standard is necessary in a particular area, and how do you decide when to stop writing a standard for a material or device?

It's very important that many constituencies be represented in the process. That's accomplished by allowing anyone in the organizations concerned with surgical implants to bring a request to one of the committees. For example, we recently had a request from the Association of Operating Room Nurses, Inc., regarding cotton products. Association representatives came to the society and said that national standards were needed in five areas, and we immediately started the procedures necessary to initiate the drafting of standards.

Each draft standard then proceeds through a predetermined voting process, and all concerned parties contribute to the development of the final standard. This leads to the published standard. It is a relatively slow process, taking a year or more to produce a standard that is published by ASTM. Thus, anticipating when a standard will be published is quite difficult.

One problem is that general-interest participants--those who come without any other support except maybe their academic programs--are finding that their resources are being depleted. This makes it difficult for the committee to maintain the organizational balance that can provide a consensus.

We're trying to establish ways in which we can normalize participation and anticipate when a standard may need to be developed. Many ASTM members attend meetings of organizations related to the biomaterials field, listen to the papers, bring back recent results, and contribute these to the working groups. The reverse of that is that we often go to FDA or the National Institutes of Health (NIH) to find out what would be of interest to them in terms of national standards. They then give us ideas that are brought back before our executive committee, and we identify individuals to work in those areas.

What is the relationship between the work of the ASTM standards committee and FDA's regulation of medical devices?

Each of our meetings is attended by a number of people from all levels of government agencies, including the National Institute of Standards and Technology, FDA, and people employed directly by NIH. So there is multidisciplinary participation.

Do you think the committee has adequate representation from device manufacturers? Is there any information they might not be aware of with regard to the standards process?

I firmly believe that the participation of device manufacturers has been critical to the evolution of the process. They participate as actively as possible, but one of the difficulties is that some companies have many international subsidiaries. So as we look toward international harmonization, it's very difficult for them to support groups from the United States exclusively because they also have to deal with agencies in the other countries where they do business.

While obtaining balanced U.S. participation has been difficult, we try to ensure that industry has a proper role that is appropriate within the voting interests. For example, if 10 members are from one industry, ASTM officially recognizes a voting representation that is balanced among membership categories. I've been impressed by the industrial participation, and we try to promote involvement in every way we can. Device manufacturers are key, often having the data that are needed for these standards.

Balanced representation has been a central issue with some of the standards, because segments of the community do not completely agree with one another. And when they do not fully agree, a consensus vote is delayed because of the "negative vote" process. This process gives each member a right to resolution of the negative vote issue prior to the publishing of a standard, thereby ensuring that the interest of any member is not overridden. It's done as fairly as possible. But again, it's incorrect to have a small percentage of the voting population dominating the process.

Do you see the work of the committee helping to contribute in some way to the resolution of the materials crisis?

The standards in and of themselves can contribute from the standpoint of providing baseline information to which anything that's different or new can be compared. Although this committee is not intended to be involved in research, it does clearly recognize the importance of all information related to safety and efficacy. Therefore, when a standard is written, it always includes a rationale to justify writing a standard on that material, device, application, or test method. In addition, we include a precision and bias statement that must be based on statistical significance. This is extremely useful--it contains key information that can further understanding.

Do you have any advice for device manufacturers confronting the potential shortage of materials?

Clearly, all need to participate. I try to encourage companies to continue the process they have been involved in previously--that is, constituting, fabricating, developing, and promoting new and unique biomaterials, as well as improving those that exist. Confronting the potential shortage is extremely important, especially with an aging population, issues of quality of life, and needs in developed countries.

Interest on the part of device manufacturers is there, but how involved they become still depends on executive-level decisions regarding benefit and risk. One aspect of risk is certainly financial. If there is a potential for major litigation, an executive committee must decide how to balance the cost/benefit equation.

Can you see any way around the current dilemma, short of introducing another supplier indemnification bill?

In time, given the globalization of industry, there's certainly the potential for reaching out to a worldwide community for biomaterials or devices. But right now the developed countries of the world have the greatest opportunity to take strides toward significant improvement. I believe that U.S. industry should be in a leadership position, but I'm not sure how this should be accomplished within the current environment.

One way would be through national legislation to properly protect suppliers while still protecting patient and consumer rights. We need to balance the equation somehow and carefully develop meaningful national legislation that can protect all parties while still maintaining the interests of those who have truly been subjected to disservice or malpractice. This will be very difficult to achieve.

There have been suggestions, from the Health Industry Manufacturers Association and other organizations, that the federal government should be more directly involved in basic biomaterials research, and perhaps function as a sort of national supplier. What is your impression of these ideas?

Government has been providing financial support for the special area of orphan materials, for which there is clinical need but not really a significant commercial benefit. Maxillofacial reconstruction is an example where the government, through a grant, provided opportunities to make high-quality materials available for facial reconstruction. Industry could not afford to do this because it would cost too much on a relative basis.

As a part of my role within a medical center, I'm interested in the quality aspects of total patient health. Our intention is not to compromise any aspect of patient health care through a surgical implant or reconstruction--in fact, it's precisely the opposite. So we need to understand both basic and applied phenomena in order to move forward toward understanding all aspects of clinical outcomes research. The Society for Biomaterials and a number of other professional organizations have joined together in an effort to get government more involved in supporting basic research that would in turn help support this entire process. There's a great need for a better understanding of biomaterial/tissue interactions, and how they relate to host interactions.

As a group, we're focusing on more fully understanding clinical outcomes through cause-and-effect relationships, going back to the device, and then going beyond the device to the biomaterials used for its construction. But it's a very complicated series of steps. We're asking NIH and other organizations in government to subsidize this effort through grants.

The Society for Biomaterials has also developed a plan to better understand the roles of analyses of explanted implants, registry tracking, and the analyses of data developed. It is critical to more fully understand the success ratios for devices. In this country, we do not have the denominator; we do not have a full record of device implantations and explantations. Some programs emphasize studies of devices that have not served as intended. However, we need to know what's truly representative, in that we need to know what has happened in total. This process is going on now, and it should be supported.

What are the most interesting developments now occurring in the field of biomaterials?

What's happening is really exciting. For a very long time, biomaterials researchers have been primarily physical scientists--engineers, physicists, and chemists--whose research focused on developing synthetic materials for tissue replacements. Most of us realized, as time went by, that we had to be equally involved with clinicians, biologists, cell biologists, and microbiologists--people who understand events at the molecular-atomic level.

We're moving into a transition phase in which biomaterials will be part of treatment as one component with biologically derived materials of new tissue-engineered products. These will use macromolecular forms, cells, products, and solutions in combination with biomaterials to actually regenerate natural tissues, as opposed to only placing something synthetic. I believe in the next decade we're going to see the opportunity to regrow tissues or organs as analogs of the original and to implant them in a functional form. It's immensely exciting because it's at the cutting edge of fundamental science--understanding not only the physical properties but also the biological properties of the contiguous interface between these different systems.

Can you foresee a "golden age" of biomaterials if we can overcome current legal and procedural difficulties?

Yes, but clearly we have to recognize that there is a very large, significant patient population that needs to be served with standard synthetic materials that function as devices--heart valves, vessels, tooth roots, joint replacements, bone plates, rods, and so on. The need for these is not going to go away, so we have to improve them as much as we can given the limits of our understanding today. We can make quantum-level improvements over what we do now in that sector--dealing with quality-of-life issues in an aging population and with life-and-death issues for some members of that population. So as a group we need to focus on these issues on a continuing basis, while still placing a percentage of our efforts into what's new and what's better.

New-Age FDA Brings Many Reforms

Medical Device & Diagnostic Industry Magazine | MDDI Article Index

Originally published September 1996

James G. Dickinson

Here's a list of dreams long dreamt by members of the medical device industry:

* FDA inspections scheduled a week in advance by day and by subject to be covered.

* Faster establishment inspection reports (EIRs).

* No-surprise FDA-483 notices of investigators' observations.

* Cross-training of industry and FDA employees at the same technical workshops.

* Collaborative drafting of FDA compliance policy guides that instruct agency investigators.

These dreams are but a few of the FDA reform realities being networked at FDA-industry grassroots meetings around the country. Combined with pilot testing of external product reviews by FDA's Center for Devices and Radiological Health, the earnest changes under discussion in the grassroots meetings add up to a new age for FDA and its relationship with industry.

"When I joined the agency in 1960," FDA Chicago district director Raymond Mlecko told over 100 device industry representatives at a grassroots meeting on June 28, "we wore white hats, and you guys wore black hats. Now it's you who are wearing the white hats!" He said it jokingly, but his words underscored those of his boss, FDA Midwest regional director Burton Love. At a Chicago grassroots meeting with drug industry representatives a few days earlier, Love had said, "It's not just talk. There is going to be less government." FDA's concern, according to Love, is that with an anticipated 30% cut in resources for its field operations in 1997, the agency won't be able to protect the public without optimum cooperation from industry. The agency's traditional investigator culture is consequently moving away from the "up against the wall and spread 'em" mentality of the past to a kinder, gentler problem-solving approach.

At the June grassroots meetings in Chicago, Mlecko had a dozen of his investigators contribute their perspectives to the mix of industry ideas being worked on in various breakout sessions. The resulting reform suggestions were then to be compiled within weeks by Chicago district program analyst Marvin Mortensen into a report that would be mailed to all participants and to associate commissioner for regulatory affairs Ronald Chesemore.

With FDA's national pilot program of medical device inspection reforms (preannounced inspections, annotated FDA-483 notices, and formal closeout letters) already launched, the grassroots meetings now focus on initial feedback from the pilot program and a host of new reforms. Feedback in Chicago was enthusiastic. Overwhelmingly, representatives of companies that had received preannounced inspections applauded not only the way they were done, but the demeanor of the reoriented FDA investigators conducting them.

The main benefit of preannouncement, they said, is that companies are able to gather the needed records in advance so that time is not wasted looking for documents after the investigator requests them. Some Chicago district investigators even schedule specific records for specific days, a week in advance. And if any of the scheduled days is inconvenient to the firm, the investigator tries to schedule an alternative. Another important benefit is that the company is able to make sure that the appropriate staff are on hand for the inspection.

Taken together, these two benefits add up to an inspection that may be completed in half the time it would have taken under the old practice of unannounced inspections. Of course, if significant problems are encountered, the inspection may drag on, and all bets are off if a possible prosecution enters the scenario. On that score, Mlecko's investigators repeatedly tried to reassure the grassroots meeting that they want to avoid prosecution. "We don't like going to court any more than you do," one commented.

FDA's traditional concern that giving companies advance notice of an inspection, and of the documents to be examined, could lead to falsification of records seems to have disappeared in the new-age agency, along with the basic operating assumption that there are malefactors in every company. I asked five of Mlecko's investigators about this, and each said they have other means of detecting false documents. It is no longer necessary--if it ever was--to ambush companies to find document falsification.

To me, that says a lot about the new spirit controlling the agency's enforcement arm. The basic assumption of industry crookedness has changed--or, to be more cautious, is changing--to an assumption that industry has the same goal as FDA, namely, patient safety.

The openness and apparent candor of the exchanges between FDA and industry personnel at both grassroots meetings was surprising both to me and to Mlecko, who asked me later how it compared to others I had attended. He was obviously wondering whether it was unique to his district. I told him it did seem unprecedented to me, although I had not been to every grassroots meeting in the nation.

Assuming that there was candor and openness in Chicago, it was revealing that no industry attendee expressed the view that FDA should have fewer inspections, or curtail the intrusiveness of inspections. They did not even complain about the public availability of FDA-483 notices or EIRs. In short, they supported FDA's statutory obligations and appeared to recognize that compliance with them could strengthen their companies' standing in the domestic and international marketplaces.

The point of departure in this collegiality was efficiency. Doesn't FDA realize, several industry attendees demanded, how upsetting it is to have an inspection interrupted, sometimes for a month or more, because the investigator suddenly has to rush off to deal with a higher-priority matter? And does it make any sense to require companies to file a Freedom of Information request for their own EIR? And how can FDA pretend that the EIR isn't ready when it has written a warning letter that must depend on the existence of a finished EIR?

The FDAers present admitted to the logic of such criticisms and promised to try to make their bureaucracy behave more efficiently in the future. Whether or not the bureaucracy will respond appropriately remains to be seen. Some industry cynics believe that nothing truly meaningful will happen without a legislative hammer to pound on that bureaucracy. In their minds, the only reason the agency is today giving the appearance of paying attention is that Congress has been diligently working on serious and tough legislation. If the Democrats take back Capitol Hill in November, these cynics contend, the old FDA will rise again.

I doubt it. The radical changes earnestly being discussed at grassroots meetings and on the 14th floor of FDA's headquarters in Rockville owe more to the Clinton administration's "Reinventing Government" initiatives, which began before the Republicans took over Congress, than they do to the extra catalyst that the Republican majority later provided. In fact, some of the reform ideas now in the spotlight actually began inside FDA before Vice President Al Gore's Reinventing Government drafting team asked the agency for its contributions.

Changes this fundamental do not come overnight, or in response to any single initiative or crisis. They evolve over many years, and involve the talents of many people and many constituencies. The fact is, most of this work has already been done now. What we are witnessing in the FDA reform movement, both on Capitol Hill and in the grassroots meetings, is the coming together of the various ideas for change, some well considered, some not. Out of it all, and independent of any bill on Capitol Hill, a new-age FDA is emerging.

James G. Dickinson is a veteran reporter on regulatory affairs in the medical device industry.


Bill Evans

For technology companies in the highly regulated medical environment, making the transition from product concept to production as fast as possible is critical. The more planning that goes into the early stages of product development, the greater the chances of lessening the time, errors, and costs related to a project. This article, written from the perspective of an engineer trying to decide which rapid prototyping technique is most appropriate for a given problem, should serve as a guide for engineers at the beginning of a project. It will enable them to intelligently plan the use of the numerous, and quickly evolving, rapid prototyping techniques and direct data transfer methods--from making early breadboards through to beta production.

Before committing a design to expensive and time-consuming physical models, design engineers may want to consider some of the recently developed "virtual product" techniques.

Multimedia Interactive Product Demos. A multimedia interactive product demo is a relatively new technique that allows users to test drive products before committing their hardware and software development dollars. These demos usually consist of a computer screen mock-up of the control panel and possibly an animation of what the operator may see during the procedure. The user then navigates through the demo using a mouse. Animations, changes to display information, and audio feedback help give the user a more-realistic product experience. With this technique, even subtle and sophisticated operating scenarios can be simulated to help clients make the conceptual leap from a written product description to the feel of a virtual product demonstration.

These kinds of demos are especially useful for products, such as surgical and diagnostic equipment, with significant user interface issues. They work well in the tightly regulated medical marketplace because they allow changes to be easily made in the early stages of development. They can also become highly refined as the design develops, and may sometimes be used as a supplement to medical equipment training videos. These demos can also be used as part of a comprehensive user-testing program if specific safety issues or user-error tolerance issues need to be addressed.

The relatively low cost and time requirements of multimedia interactive demos are big pluses for this technique. Creating a demo can take from as little as one-tenth to one-half the time of doing a traditional hardware and software mock-up, with a similar reduction in cost. And compared to more-conventional prototypes, a demo is highly portable (when a color laptop computer is used). Additionally, it forces designers to create a kind of story board early in the process, which requires using another level of critical thinking.

The multimedia demo technique gives the designer, engineer, or potential customer a more visceral feel for how a product will behave. However, because the demo is not the real thing, conducting reality checks with certain aspects of the design is recommended--especially when it comes to human ergonomic issues such as a display's readability at low light levels. Similarly, the tactile feel of a product cannot be reproduced on a computer. If an important part of the design involves rotary controls and push buttons, it may be necessary to find examples of real controls and base product specifications on them.

Another advantage of a multimedia interactive product demo is that it can help catch mistakes by being put into the hands of many people, including potential users, salespeople, and the like. It encourages viewers to analyze and constructively criticize the interface design rather than simply consider it a finished product just because it looks that way on the computer screen. An added benefit of using these demos to refine user interfacing is that the final interactive model becomes a very tight specification for software and electrical engineers--much better than a traditional written description.

When to use this technique: Multimedia should be used when there is complex user interaction with a product and when there is a need to quickly prototype and test the ideas with real users. (For example, this may be the case with surgical lasers, surgical instrument electronic controllers, and any complex diagnostic equipment.)

A few design firms include this specialty in their skills. Check around and ask for demo examples, but beware of designers who use these tools for creating software interfaces. Instead, try to find a design firm with product, rather than software, experience.

Computer-Aided Industrial Design (CAID) Rendering. The CAID rendering process involves using software to create a product as a 3-D model and then planting it in an imaginary setting using a variety of photorealistic rendering packages (well-known packages include Alias and CDRS, among others). The quality of such images is common knowledge, since they are used extensively in visual effects for Hollywood films.

The main advantage of CAID is that engineers can obtain good criticism of a product before committing to a physical model of the design. And since CAID rendering minimizes the conceptual leap that must be made with traditional product sketches, it can be a powerful tool for engineers and designers to communicate ideas to upper management, and useful with early-stage product focus groups.

The speed--weeks less than that of traditional methods of making and photographing models--and relatively low cost associated with this technique are additional advantages of using CAID rendering for product walk-throughs. Costs will vary, but a good rule of thumb is that CAID costs for smaller objects will be similar to those of models and photography, whereas CAID cost and time requirements for larger products will likely amount to only a fraction of those related to model-made equivalents.

However, the very strength of CAID rendering can also become its weakness. The polished look of CAID images can inadvertently discourage criticism of a product. Engineers may be lulled into a false sense of reality by eye-catching CAID images. Similarly, upper management could be led to believe that the product is further advanced than it actually is and thus hold back full criticism.

CAID rendering also has drawbacks when it comes to ergonomic issues. For instance, it is unwise to conduct early studies of intimate handheld products using this technique--a quick handmade foam model may be a better choice. On the other hand, CAID may be good for larger products, such as racked instruments or relatively static product housings that have to look good but aren't physically handled every day.

Another issue to be aware of is that some CAID packages do not accurately handle underlying numerical details, such as internal component sizes. So if accuracy of underlying data is really important for a particular project, be sure the CAID program is good at this particular function.

When to use this technique: CAID rendering can be used as a powerful presentation tool for product concepts that do not involve intimate ergonomic contact (for example, large diagnostic equipment or any product that is predominantly a "box"). If computer-aided design (CAD) compatibility is important, your CAD vendor will likely know firms that can offer this service to suit your CAD.

There is now a vast array of rapid prototyping techniques available for physical prototypes. The most common are explained here. If some are unfamiliar or appear useful, request samples from vendors or keep an eye out for them at trade show exhibits and presentations.

Stereolithography (SL). This prototype technique most closely resembles magic in the product design world. With this method, a floppy disk or E-mail data file is sent to a vendor. Based on that information, layers are built up in a resin bath to create a faithful reproduction of the specified product.

Two advantages of this technique are that data do not have to be handled by a machinist operator who could potentially make errors of interpretation, and that with some careful shopping around, one can end up with a prototype at one-third the cost and in about one-third the time of traditional models or machining. (Many vendors can turn STL 3-D CAD files into parts in about one week.) The result is more chances to get it right by doing more model iterations and ultimately ending up with a more mature design, even if it doesn't save money overall.

There are, however, a number of drawbacks to this technique. First, resin is a brittle and thermally unstable material, which means that a model can warp in the trunk of a car on a warm day. In addition, the model is built up in shallow steps--typically 3 to 5 mil--so small details tend to wash out. The hand-finishing can sometimes be poor, so care must be taken in specifying to vendors. And since working from direct data doesn't leave the traditional paper trail of drawings, data must be checked very carefully.

One way to minimize this problem is by taking full advantage of the power of CAD to run interference checks and by driving through the design with engineers before sending out the model. It also helps to freeze a version of the database as a reference. And while SL is low-cost and fast, remember that direct data machining companies are trying to stay competitive on pricing and can produce models in more-durable materials, so it may pay to shop around and weigh the differences for each situation.

When to use this technique: SL should be used on smaller complex parts, usually those intended ultimately for plastic or metal molding manufacturing techniques. Any plastic injection molding, metal die casting, or other part intended to be made in an allied process that will fit in your vendor's machine is a candidate. (SL machines are limited by size; for example, the part must fit in a 6- or 12-in. cube.)

SL vendors such as Scicon Technology (Valencia, CA) can be found in the prototyping services listing in the phone book or by contacting a CAD vendor for recommendations. Manufacturers such as 3-D Systems, Inc. (Valencia, CA), will also make referrals to local service bureaus that have bought their machines.

Solid Ground Curing. Solid ground curing is a relatively new variation on SL. It currently offers a price point and delivery similar to SL, but with a slightly stronger and more thermally stable material. Otherwise, it has advantages and disadvantages similar to SL. The process still builds the part up layer by layer, using a photopolymer that is fixed with UV light and then backfilled with wax. Each layer is then milled to the correct thickness before the next one is applied.

Currently, the solid ground curing method used by Cubital America (Troy, MI) has a maximum step resolution (meaning the thickness of each layer as it is built) of about 6 mil, compared to SL, which can go down to 3 mil. On a final note, the quality and strength of prototypes produced by SL and solid ground curing methods change frequently as vendors strive to improve their competitive positions. Consequently, it is important to keep reappraising these techniques for strength and accuracy. Also, remember that it would be difficult to sterilize SL or solid ground curing parts using any process that elevated temperature above 100°F.

When to use this technique: Solid ground curing has constraints similar to those of stereolithography, except it is more appropriate for parts that need to be a little stronger and more resistant to heat and handling. A vendor of solid ground curing equipment should be able to refer you to local service bureaus that have bought its machines.

Selective Laser Sintering (SLS). The SLS technique also builds up a part layer by layer. As the name implies, a 3-D object is created from layers of powder sintered by a powerful CO2 laser. Because it can create parts in thermoplastics strong and thermally stable enough for evaluation in more-realistic conditions, SLS is exciting for medical applications. Both nylon and polycarbonate can be sintered, and the final part will have about 75% of the properties of the normal polymer. The layers are 5 mil thick and, overall, the part can be as accurate as one formed by SL (± 0.005 in.).

These characteristics make the process ideal for complex plastic prototypes that need to be highly functional and made more quickly than by machining. Appearance may be a concern, however, because the sintering process makes the parts look slightly granular. This can be rectified with skilled finishing, but such finishing tends to slow the prototyping process and make it more expensive. The cost and scheduling requirements are similar to those of SL, but as with any of these rapid techniques, one must comparison shop between processes and vendors, because quotes vary significantly.

Casting wax and even metals can also be laser sintered. SLS of a metal mixture of 60% steel and 40% copper is in its infancy; currently, metal parts are used for creating tooling cavities and have less-smooth surfaces that make for poor-looking parts. But keep an eye on this process, because it is likely to improve and add more thermoplastic materials to its capabilities.

When to use this technique: SLS should be used on complex plastic parts that must have functionality close to those made by the intended production technique, which is usually plastic injection molding. Part aesthetics are poor, and engineers should keep in mind that machines are limited in size to 12 in. diam by 15 in. tall. SLS equipment vendors such as DTM Corp. (Austin, TX) may be able to refer you to local service bureaus that have bought their machines.

Other Rapid Prototyping Methods. There are several other processes worth considering that are not yet as commonly used as some of the previously mentioned techniques. These techniques include fused deposition modeling (FDM) and laminated object manufacturing (LOM).

FDM creates a part one layer at a time from a filament of material precisely manipulated in the x and y axes, with each layer being melted to the preceding layer. Possible materials include ABS, including a specially formulated medical grade called MABS, as well as elastomers and investment casting wax. MABS parts can be gamma sterilized. According to Stratasys, Inc. (Eden Prairie, MN), a manufacturer whose systems use the FDM process, parts must fit within a 10 X 10 X 10-in. working envelope and the process will typically provide ±0.005-in. accuracy.

LOM takes low-cost sheet materials like plain or specially impregnated papers and builds them into a solid part suitable for form, fit, and function evaluation or as patterns for casting. This technique is based on the simple idea of a laser cutting each layer, then laminating it to the preceding layer. Large parts can be made to 22 X 32 X 20 in., with accuracy of ±0.010 in.

As is the case with many rapid prototyping techniques, manufacturers of FDM and LOM systems are continually improving both the resolution of their processes and the materials they can process. So keep an eye on these and other techniques, and take notice of samples and demonstrations at industry trade shows.

Direct Data Machining. Frequently used today, direct data machining is accomplished by designing a part in CAD and then sending the CAD data file, rather than traditional drawings, to a vendor for machining. If a company can afford the time and money, it may find that it is better to direct data machine plastic parts rather than using SL or solid ground curing, since the parts are aesthetically cleaner, more accurate and thermally stable, and made from stronger materials. Machining this way can take as little time as SL (about one week), but much depends on part size and complexity as well as your relationship with the vendor.

Many vendors are highly versed in handling direct data with metals. However, there are a number of problems to be aware of when it comes to direct data machining of plastics, which are often needed when making copies of injection moldings or early prototypes of disposables. Many of these challenges revolve around the significant differences in the way plastics behave structurally when they are injection molded versus machined.

When plastics are molded, the polymer chains are well aligned and tend to optimize the strength of the part. But in the direct data machining process, strength and integrity can be lost when the chains are broken as the part gets sculpted from a solid block. Therefore, direct data machined plastics behave quite differently from molded plastics. Pigment and color can also affect the ease of machining--a white sample may machine in a radically different way than a black sample. Of particular importance to the medical market is that some plastics may be highly suitable for sterilizing when molded but, when machined for prototypes, their dimensions and features can change significantly during autoclave, gamma, or EtO sterilization. As a result, companies should test-sterilize small machined samples before committing to expensive prototypes.

When to use this technique: Direct data machining is effective when used to mimic molded parts in the intended production material or to more quickly machine anything that would in the past have been machined from fully dimensioned drawings by more traditional techniques.

Many CNC machining shops are familiar with direct data machining. Although such shops may be biased toward metals, simple fabrications in plastic can also be handled. To mimic more-complex plastic injection moldings, contact specialized model-making service bureaus, like Visual Engineering (Plano, TX), that may be more familiar with the issues. As with any manufacturing process, quality can vary dramatically, so test vendors with smaller simple parts and work up to the tricky challenges.

Combining Machined and SL Parts. To help control the costs of prototyping very large or intricate parts, it is sometimes advantageous to combine SL and machining. SL may be more cost-effective for some intricate details but may not have the mechanical properties needed for features such as slides, snaps, bearing surfaces, and so on. For example, direct data machined plastics can be used to handle the sliding parts, which are then dropped into the main part that has been made by SL.

There may occasionally be a need to combine real molded parts with SL or machined parts to get special part qualities. A good example is that of living hinges, which may be needed in intricate small mechanism parts for medical disposables. To select hinges with the proper dimensions from the myriad low-cost polypropylene containers available, take calipers along when going shopping. Then machine the hinges out to incorporate into the prototype.

Laser Cutting of Light Materials. If used appropriately, laser cutting can revolutionize rapid prototyping. It radically reduces the time and cost of producing innovative ideas, particularly in plastic. For example, the process can create acrylic parts in as little as 24 to 48 hours from 2-D CAD files at a cost of $3 to $4 per piece, whereas it may take one to two weeks at a cost of $50 to $100 for similar parts using traditional machining processes. This translates into the opportunity to create more design iterations and refinements, as well as low-cost multiple copies for more extensive testing.

Laser cutting essentially involves placing a flat sheet of material on an x-y axis bed and using a high-power laser to accurately cut 2-D blanks of the 2-D shaped file given to the vendor. The thickness of cut possible will depend on the material, which can include foam, wood, paper, cardboard, acrylic, and plastics, among many others. A thickness of 0.5 in. is possible in most of these materials, but check with the vendor before assuming that a specific plastic can be cut this thick. Incomplete cuts can score the surface for marking or decorative purposes. (Laser cutting of metal will be addressed later.) A bonus is that many plastics are effectively flame polished on the edges as they are cut. This can be particularly important for copying the surface qualities of molded parts (for example, hydrophobic or hydrophilic surface qualities) or maintaining optical clarity.

Laser cutting is also useful for mocking up dynamic mechanisms such as latches or hinges. This is done by creating a representative cross section through the mechanism and then making laser-cut profiles. The parts are then laid out and fitted together with pins for hinge points to observe how the mechanism will operate in reality.

In addition, laser cutting is useful for mimicking disposable plastic parts that carry fluids, and for making sheet-metal prototypes out of polycarbonate. Polycarbonate has a bending property similar to that of steel but is clear, thus allowing see-through sheet-metal mock-ups. If one decides to mimic sheet metal this way, it is essential to debug the laser scoring and bending process with the vendor. If polycarbonate is scored too deeply, it will crack, so establish appropriate bend allowances for the 2-D blanks prior to forming.

Creativity is the key to fully exploiting the laser cutting process. As an example, Bridge Design makes its prototype printed circuit boards by laser cutting a blank glass fiber­based PCB substrate (no copper cladding). It's cheap and fast. If one cuts the through-hole component patterns, components can be stuffed in to check clearances in complete product assemblies and to test switch and potentiometer elements.

Laser cutting can also be used to evaluate airflow or appearance issues of sheet-metal boxes. For example, to mock up a box in sheet metal with vents and various hole patterns, a kit of these different facets can be created out of laser-cut cardboard and then stuck together in the desired 3-D form. Actual fans can then be mounted into the assembly and airflow can be measured through good approximations of the intended sheet-metal enclosure.

There are a few important points to remember when adapting 3-D design within the 2-D limitations of laser cutting. If the parts cannot be created from a single 2-D blank, they may need to be layered. The 3-D structures can be built up by combining the first set of 2-D shapes with a second set. Rotate them at 90° with respect to the original 2-D shape for even more creative combinations, and expect to experiment quite a bit before succeeding. Obviously, there are many ways of bonding laser-cut assemblies. Double-sided tape and solvent cements are some of the better methods.

When to use this technique: Consider using laser cutting for any intricate plastic parts that can be represented as 2-D forms, or build up more-complex 3-D forms with multiple layers. The process can be used to mimic injection-molded disposables and fluidic components. It is also especially useful for creating quick breadboards of any dynamic mechanism that can be represented by clever combinations of 2-D shapes, such as gears, levers, hinges, pivots, slides, detents, and so on. Laser cutting equipment vendors like Lasercamm (Menlo Park, CA) or local model-making companies may be able to refer you to nearby service bureaus.

CNC Water Cutting of Foamed Plastic and Rubber. CNC water cutting is closely allied to laser cutting and is worth considering--especially for foam rubber materials. With water cutting, one can do complex fabrications in foam with much more precision than would be possible with manual cutting or even die-cutting. Neoprene up to 1 in. thick can be cut with great precision.

This technique is a great way of mocking up large volumes of experimental gaskets. Combining these gaskets with plastic or metal underlays produces complex prototypes that are highly cost-effective for a clinical trial or a beta test site. The costs and schedules of this process are similar to those of the laser cutting of plastic.

When to use this technique: CNC water cut any intricate 2-D shape in foamed rubber or plastic for gaskets, seals, or fabrication of disposables that need these materials. Using this technique avoids laying down expensive and time-delaying dies. Unfortunately, this is a difficult service to track down. Jet Stream Water Cutting (Hayward, CA) may be one place to start. Also try asking local foam converters if they know who has the equipment in your area.

Laser Cutting of Metals. Before deciding to laser cut metals, be aware of their limitations; metal is not as easy to laser cut as plastic or wood. Much more powerful equipment is needed, and there are longer lead times with most sheet-metal vendors. It is also important to pay attention to thermal issues. With close-packed vent patterns in thick sheet metal, considerable warping due to thermal buildup can occur. Some materials--such as copper beryllium, which is poisonous when it burns--are simply not amenable to laser cutting. However, in such instances substitute materials can be used.

To create sheet-metal models, one can make a CAD file and unfold it manually on a CAD system, adjusting for bend allowances or use. More-sophisticated CAD systems, such as the ProSheetmetal module of ProEngineer, allow unfolding of sheet-metal parts automatically and take bend allowance into account. However, for those who are not highly skilled in this area, it is advisable to create the CAD file of the desired parts in their finished 3-D state, and let the vendor accept the liability of unfolding the blanks.

As with any sheet-metal fabrication, lead times for laser cutting can vary from days to months, depending on the complexity and vendor relationship. However, expect to get flat blanks done more quickly than any parts that require bends.

When to use this technique: Consider using laser cutting on any sheet metal fabrication that would have been done using traditional CNC punching. The technique will be most competitive when the shapes required are complex and would otherwise need either complex manipulation and nibbling of the metal or custom punches (fan rings and custom connector cutouts, for example). Most larger CNC sheet-metal vendors are now using laser cutting.

Story continues with PILOT PRODUCTION


Return to the previous page

After proving the product design with a single model, multiple copies may be required for pilot manufacturing or beta site testing. Following are some possibilities to help make the transition to small batches of parts.

Casting of Urethane Plastics for Copying Injection Moldings. Once there is a proven version of a product, multiple copies will often be needed for marketing feedback, design evaluation, or clinical trials. When the number of copies rises above four or five, and one wants to quickly and inexpensively mimic production injection- molded plastic parts, then urethane casting is a viable option.

The urethane family of plastics has broad characteristics, ranging from rigid to rubberlike. These plastics are not as strong or thermally stable as typical thermoplastics, but they achieve usable properties for closely controlled beta tests. The procedure for casting multiple copies of a design for pilot production or alpha and beta site testing is as follows.

First, create a pattern (which, in many cases, is the original model used to prove the design). It is invested in a silicone rubber mold to create a cavity, which usually has a limited lifetime of up to about 25 units, depending on the quality of the original tool or on the intricacy of the molded parts. Once the silicone rubber cavity has been made, the appropriate grade of urethane is poured to cast the part. Remember when doing this that it is possible to get UL fire-retardant grades of resin, but be sure to meet the minimum material thickness to qualify for fire retardancy. (Casting of plastics is definitely an art, and as such is typically done best by model makers.) On simple parts it is also possible to skip the pattern stage and direct data machine a cavity into a block of plastic.

One can also combine materials. In mimicking areas of molding that need to be structurally strong, set in small sections of the real material, such as polycarbonate snaps or acetal runners. Another thing to consider is that one can cast the parts in self-color, a process in which pigment is put in with the resin. This is helpful if the fit between the parts is critical and paint layers won't work. Just as in the SL process, it is important to remember that with urethane casting, the material is highly heat-sensitive. Thin wall sections below 2 mm are especially prone to warpage.

Be aware of how lengthy a process casting can be. It takes up to a week to build and cure the silicone rubber mold. This tool can then produce a set of parts only every 24 to 48 hours. Be sure to factor in finishing or painting time after that. Add all this up, and 30 prototypes can take quite some time.

When to use this technique: Urethane casting is a good choice for short runs of any plastic parts that would otherwise have to be made with costly and time-consuming tooling. Check with the leading model-making service bureaus in your area to find out if they have extensive experience with this technique.

Soft Tooling of Plastic Moldings. When people speak of soft tooling, they mean many different things. In a way, silicone rubber molds of plastic parts could be considered the ultimate soft tooling since they actually are physically soft. But sometimes people refer to either soft steel tooling or soft aluminum tooling, which involve either less-hard P20 steel or aluminum and can produce a large volume of parts at modest piece-price increases over conventional injection molding.

Soft tooling can be the solution to some short-run needs. It proves even more meaningful for runs numbering in the hundreds or thousands. This process is typically in the range of 50 to 90% of the cost and schedule of a hard tool equivalent. (Generally, an aluminum tool might be guaranteed for 50,000 to 100,000 impressions, while the full-priced conventional hard tool is good for 500,000 to 1 million.)

When to use this technique: Consider soft tooling of any injection- molded plastic parts when total production volumes are low (less than 100,000) or the market window is very short and every week saved is worth considerable revenue.

Many injection molders now offer this option. Increasing power of solid modeling CAD has enabled a new generation of rapid prototyping bridge the gap between silicone rubber tools and traditional aluminum tooling, and to shorten the tooling cycle to as little as four weeks. To find companies familiar with this technique, check the trade press for toolers who advertise this service, or ask existing molders for referral. Companies such as Phillips Plastics Corp. (Hudson, WI), PTA Corp. (Longmont, CO), and Plynetics Corp. (San Leandro, CA) are just a few places to start. Some directories may also list this service under rapid or fast-turn tooling.

Creating Complex Parts from Multiple Use of a Single Small Part. Another technique in the area of soft tooling plastics is effective, for instance, when making a disposable medical device from a number of repeated small elements. This can perhaps best be explained through the analogy of a comb. To evaluate specific materials for their engineering properties, it may be appropriate to mold one tooth of the comb quickly in a prototype mold and have many copies of it made and then fabricated into a frame combining all the features. The key elements of the part can be evaluated in the correct materials and processes in quickly produced small tools (production in 5 to 6 weeks is possible for a small cavity, as opposed to 12 to 16 weeks for larger production tools). The remaining frame parts are made using a much lower-cost technique such as laser cutting or machining.

Conversion from Prototyping to RIM Production. If the intended pilot or production volume of a product is in the range of 20 to 50 units per year, some of the previously mentioned pilot processes are viable techniques. This is particularly true for the cosmetic enclosure components of many types of medical equipment. As mentioned earlier, urethane casting can work well for making this transition, especially if one plans to scale up the process to RIM (reaction injection molding of polyurethane foam into a modestly priced low-pressure tool).

For example, imagine a part the size of a typical 15-in. computer monitor bezel. In order to make a few copies in cast urethane as previously described, one might spend $10,000 on the pattern, $5000 on a silicone rubber tool, and from $200 to $300 per finished part. Now say one wants to scale that up to RIM. A RIM vendor may charge $10,000 to make another pattern, $15,000 to create the tool, and $150 for each part. Clearly, a large chunk of the tooling cost is a pattern that is being made twice. With a little forethought, the same pattern can be used for both the early handmade urethane castings and a RIM tool.

To get the most out of a pattern, one needs to understand shrinkage rates. Any casting or molding process involves a shrinkage from the size of the pattern or cavity to the size of the final molded part. With urethane casting and RIM, the shrink rate is very similar. Typical shrink factors are somewhere in the region of 0.001 to 0.003 in. per inch (or 0.1 to 0.3% shrink). However, every vendor works with slightly different numbers due to the resins used, so check the shrink factors for both the urethane casting vendor during the early prototype runs as well as for the intended RIM vendor. One note of caution is that intricate thin-wall stereolithographic models have a tendency to break when invested for silicone rubber tools. Make sure your urethane casting vendor knows if you need to keep the patterns intact for future use.

When to use this technique: The RIM technique is useful for any large or complex enclosure parts that would be much more expensive or impossible to machine or bend from sheet metal. The material is strong enough to replace large metal panels. RIM molders such as Design Octaves (Santa Cruz, CA) are generally listed under a seperate category in larger business directories.

Like many things in the business world, picking and nurturing rapid prototyping vendors is about building a relationship, including trust and understanding. Look for vendors who are honest in their appraisal of the job and realistic about their schedules.

Bill Evans is principal of Bridge Design (San Francisco), a product development consulting company.

Return to the previous page

The latest entry in the race for rapid prototyping is a ballistic particle manufacturing (BPM) technology first examined by MD&DI in November 1994. Although not commercially available then, a 3-D printer that makes use of this technology is now finally on the market.

Using a patented material-jetting system that shoots molten microparticles of thermoplastic into 3-D models, the printer's five-axis build process--referred to as digital microsynthesis--builds an object by combining thousands of identical dots of material. The nontoxic plastic is shot from a piezoelectric jet head in a round molten state at frequencies as high as 12,000 microparticles per second. These particles flatten upon contact, slightly melting the surrounding plastic and establishing a gluelike bond between drops. Because the jet head can deliver the material in any direction, objects can be built with their true curvature without the stair-stepping effect that can occur from jetting the material straight down.

BPM Technology, Inc. (Greenville, SC), began shipping the $34,900 Personal Modeler system this month.

Return to the previous page

While rapid prototyping techniques offer product design engineers a variety of options for making the transition from product concept to production, verifying that the process was done correctly can be both costly and time-consuming. A natural complement to CAD/CAM and rapid prototyping systems, reverse engineering systems can help designers and production engineers to speed the inspection of both tooling and finished parts. A reverse engineering system can digitize a part's internal and external features using automatic, cross-sectional scanning, and then compare this data point set with the original CAD file to identify how far out of dimension any point of the part's surface is. Most commonly used in first-piece inspection, other applications for the system include quality assurance for tooling as well as true reverse engineering--taking an existing part for which there is no accurate CAD data (perhaps because the part is old or was modified after the mold was made), and essentially re-creating it.

In this reverse engineering system by CGI (Minneapolis), users are able to enter the net processing dimensions, horizontal and vertical sampling, and filtering criteria. After selections have been made, the part may be previewed before final processing.

Return to the previous page

Many CAD and rapid prototyping vendors talk about being able to transfer data easily and quickly. But all too often a carefully prepared CAD file that has been downloaded is not translatable or there are other compatibility issues, such as those caused by DAT tapes that are not properly formatted for a vendor's system, incorrectly set export parameters in your CAD system, or the incompatibility of file-naming conventions between the DOS, Windows NT, and UNIX worlds. Following are tips for preventing such glitches.

  • Before finalizing a design, use a trial part and test out the transfer process with the vendor.
  • Use a dedicated line and invest in the highest-speed modem available. For those with an FTP site on the Internet, this is a cost-effective way of transferring large files. When compressing files for transfer, make the files self-extracting to avoid compatibility issues on the receiving end.
  • Avoid some UNIX to DOS problems by keeping to an 8.3 naming convention when originally naming the files (for example, use DRAWING.DWG instead of COMPLICATED_DRAWING_NAME.DWG).
  • Consider carefully how data will eventually be used.
  • Don't believe claims of integration until you have tested them with real-world problems that your particular company needs to solve.
  • Remember that, with the proliferation of computer E-mail and courier services, it is possible to work with the best vendors even if they are located in another part of the country.