designThe Magazine for Medical Product Design & ManufacturingImage:Yevgen_Lyashko/iStockphoto.comSteve Augustyn, Team Consulting Ltd, Ickleton, Cambridge, UKTo effectively manage the entire lifecycle of a medical product, from creation to end-of-life processing, whilstachieving cost efficiencies, apply DFX principles from the outset.Design for X (DFX)reduces the opportunityfor error and improvesthe quality and cost efficiencyof the final product. DFX is acatch-all term for establishingthe performance of your designsolution against differentcriteria.1 The term can beused to describe design forexcellence, but whilst this termis good for adopting a generalmindset, it does little to helpwith focussing on the key areasof design performance. In thiscontext the X in DFX is aninterchangeable characteristicthat the design performance canbe measured against.1 X hasseemingly countless variants,including:00 Design for manufacturingand assembly (DFMA)related to managing production costs.00 Design for production, which matchesthe design against production costsand time-scales while maintaining therequired quality.200 Design for aesthetics governing thevisual aspects of the design; design forergonomics examining the performanceof the interaction between the user andthe product.00 Design for maintenance to review howthe product will be maintained across itsfunctional life.00 Design for inspection, design forcalibration, design for disassembly,design for end-of-life processing, designfor strength . . . the list literally goes onand on.Objectives and design forlifecycleThe typical cost of changecurve3 shown in Figure 1maps the increase in cost tocorrect a problem as the designand implementation costprogresses. It isn’t possible togive precise figures, as designprogrammes vary dramatically(fitting steam catapults to youraircraft carriers late in theday costs a staggering amountof money whereas making aplastic spoon handle slightlythicker so it doesn’t break is alot more manageable). A goodrule of thumb is that at eachstage of a development process,the cost of correcting an errorincreases tenfold. Therefore a£10 error in the concept stage(for example, correcting apoint on a specification) costs £10,000 tocorrect by the time you’re in production(spec rewrites, tooling corrections, repeatedtests, product recalls, process validationreruns and so forth). This financial impactprovides a compelling argument forapplying DFX principles from the outset.The cumulative approach of DFX can beidentified as design for lifecycle. 1 Figure 2Design for X, andBe Prepared for Anything
design(adapted from BS 8887) shows how someof the different DFX approaches map ontothe life of a product. This issue is too wideranging to give definitive answers for anyof the significant areas of impact in thespace of a magazine article, but some ofthe references at the foot of this articleshould prove useful. If we consider thethree main phases of a product’s life—creation, use and end-of-life processing—the implications of DFX become clearer.Creation and realisationDFMA is one of the better understood andapplied aspects of DFX. Whilst the originsof design for manufacture, particularlyfrom the context of standardised designs,can be traced back to Eli Whitney4 at theend of the 18th century, it is Hitachi’sAssembly Evaluation Method from the1970s and the work of Peter Dewhurst andGeoffrey Boothroyd at the University ofMassachusetts that really defined DFMAwith a deep, analytical methodology.Originally based on a series of chartsand lookup tables, the Boothroyd andDewhurst DFMA approach is nowavailable as a software package that canbe used to analyse designs at any stageof development to obtain an estimatedcost. This methodology will give you apredicted cost of different approaches,allowing the designer to choose the mostcost-effective option. However, this processis limited by the knowledge built into thesystem and it will only report one piece ofthe puzzle. More specifically, it will allowthe designer to compare two puzzle piecesthat he is holding.A component assembly variability riskanalysis review (CAVRA) can provide anumerical score to compare the relativeassembly risks in different designs.5Running a detailed analysis on a largepiece of equipment would dramaticallyeat into the time and budget available,but by using the guidelines as achecklist as the design progresses, alot of the benefit can be achieved for afraction of the cost. For high-volumemass-produced parts, spending moretime on the DFMAprocess can beeasily justified,but sometimes athorough reviewof the designwith experiencedproductionengineers cangenerate a lotof value (see thesection titledCaveats and goodpractices).One equallycrucial piece of thepuzzle is design forproduction. This isdistinct from DFMA, as it deals more withthe logistics of manufacturing; in the areaof medical device development, this canhave an even bigger impact. For example,you run your DFMA analysis to comparea steel spring against an elastomeric one.With the number of features required,the elastomer comes out as the clearwinner so you merrily proceed with theinformation neatly tucked away. Now youhit production planning and you find outthat the minimum order quantity for yourmedical-grade elastomer is three metrictons (enough for five years’ productionat full capacity) and the material is on a12-week lead time, three weeks after theproduct is supposed to be in clinical trial.Use and implementationThe use and implementation of theproduct crosses many boundaries andexpertise including mechanical design,industrial design and ergonomics. Theeffects of mechanical design can be feltmost keenly in areas such as design forreliability, design for service and designfor wear. For example, when I worked onthe design of photocopiers at Xerox, a lotof work was done to prevent paper jamsin the machines. Jams are inevitable—younever know what sort of rubbish someonewill feed into the machine. However,designing a paper path in the machine thatis easily accessible to the user allows thesefailures to be recovered simply and quickly.By reducing the importance of design forservice you can end up in the nightmareposition known to many car owners,where replacing something as simple as aheadlight bulb can necessitate a trip to thecar dealership. Whether the decisions thatled to this state of affairs were motivatedby aesthetic or pecuniary concerns, I’llleave to the reader to decide.Determining and measuring the benefitsof industrial design and ergonomicscan be more difficult. Poor design on aFigure 1: The cost of correcting a problem increases considerably at each stage of theproduct development process, as shown by the cost of change curve.CostofchangeSpecify Implement Verification ProductionBy reducing theimportance of designfor service you can endup in the nightmareposition known tomany car owners,where replacing aheadlight bulb cannecessitate a trip tothe car dealership.
designconsumer product or a fatal flaw in theinteraction could lead to an unappealingproduct or something that is inherentlydangerous. An appealing or easy-to-use product may compensate for somelacklustre engineering, but a poorlyconceived product will present a barrierthat no amount of clever engineering willbe able to balance out. What constitutesgood industrial design is beyond thescope of this article, but good guidance isavailable for the definition of interactiondesign. Two resources I have found usefulare the Principles of Universal Design6and ISO/IEC Guide 71.7 Both of theseguides are based on an inclusive designapproach, but the principles hold upwell for all types of product interaction.Design reviews, state space analysisand user trials all work well to matchthe performance of the product againstidentified performance characteristics.End-of-life processingDesign for end-of-life processing is oftenlow on a list of objectives for a designer.Working in the medical device space doescome with its own requirements related toproduct safety, and it can be necessary toconsider how to safely disable a single-useproduct. With issues of cross-contaminationor dangerous, poorly controlledreprocessing, finding ways to permanentlydisable products after they have beenused once can become very important.Single-use, auto-disable syringes such asthe BD SoloShot8 show how such featurescan be easily engineered into products. Itmay be enough to ensure that productscan be incinerated without producingharmful toxins or optimised to use theminimum amount of material. Beyond basicsafety issues and avoiding toxic materialswherever possible, manufacturers andimporters are coming under increasingpressure to consider the processing of theirproducts at the end of life.With rapidly rising demand in developingcountries, the pressure on limited resourcessuch as rare earth metals is going tobecome ever more acute, and recyclingand reprocessing will become ever moreimportant as European guidelines suchas WEEE and RoHS place responsibilityon the postprocessing and constructionof electronic products to minimiseenvironmental impact. This is an areawhere car companies are leading the way,with BMW (among others) adoptingthorough end-of-life strategies to recoverand reprocess as much material as possible.9Design Engineering magazine has an articlethat provides a very useful overview tothe whole issue of sustainable productdevelopment and manufacturing.10Caveats and good practicesIt costs time and money to analysedifferent designs against the various Xcriteria and to implement the resultingchanges. These development costs willhave to be annotated over the entireproduction run (or the projected paybackperiod). Unfortunately, numerous papersand academic guides to DFX act as if theengineers’ time costs nothing and that allactivity directly benefits the product.One of my favourite engineering quotesis attributed to Henry Ford: “An engineercan do for a nickel what any damn foolcan do for a dollar.” Whilst this is apretty clear rallying call for the benefitsof applying DFX methodologies, it isimportant to keep DFX in the context ofthe whole development process. The crucialpoint is to make these methodologies asefficient as possible and understand thatthe outcome of a review is limited by theknowledge, experience and prejudices ofthe review panel or process.Going back to the cost of change graph(Figure 1), applying these principles earlyin the design process will reap the largestrewards. Applying “three yards of DFMA”once the prototype has been completed andtested probably will lead to compromiseand frustration if the design is found to beDesign of parts andmanufacturing processesPiece partmanufactureAssemblyDesign of parts andmanufacturing processesDisassemblyPiece partreprocessingMaterial recovery ordisposalDesign for disassemblyDesign for serviceDesign for lifecycleDesign for manufactureDesign for manufacture and assembly Design for inspection Design for end-of-lifeprocessingUseFigure 2: How various DFX approaches map onto the life of a product. Diagram adapted from BS 8887.With rapidly rising demand in developing countries,the pressure on limited resources such as rareearth metals is going to become ever more acute,and recycling and reprocessing will become evermore important.