Terrell Williams Time and volume beyond what is possible prior to market release.

Benjamin Richard Wang MD Not a better solution
Designed solely by engineers
Can't be manufactured
Poor Regulatory strategy
No clinical study data
Poor Quality Record
No demand by the medical community

Bruce Youngman Rather broad based question if you consider the variety of devices and their ranges of complexity.

Greg Smith I agree with Benjamin (Richard Wang), but start-ups in particular also need to identify / recognise their shortcomings in terms of managerial skills at different stages of the development. This is especially true if a device company is spun-out from an academic institute and is run by people inexperienced in getting sufficient and timely finance in place, people with inadequate technical and project management skills and people with no understanding of marketing and product launch. Projects can fail at any of these stages. Of course you can (very selectively and very carefully) outsource many of these aspects - especially technical development and project management. It's easier to identify how things can fail than it is to say how they will succeed. If it was that easy we'd all be doing it!!

Why do medical devices fail? In many cases, by design, same as any other device. That is not to say that the designers INTENDED the device to fail, only that the devices were not designed well enough for the eventual conditions. We do the best we can with FMEA, but we have to recognize that we can't foresee EVERY circumstance.
We also must remember that the installers, the MD's and PhD's, are also human, and that medical errors DO happen. Again, we can try to ensure that the device MUST be installed correctly. But humans make errors. Errors can lead to failures.
Thirdly, we must remember that everything that was put together, sooner or later falls apart. Medical devices are subject to the same universal law. We wish life were otherwise, but there it is.

Joseph Mandell I have to agree with Horace, Human error is likely the leading cause of failure; either the operator failed to use the device correctly, or the manufacture failed to realize that the operator was going to use the device incorrectly.

A good example is of this is took place in the early 1990s. Several manufactures came out with pulse oximeters. Unfortunately, more than one manufacture used a round 9-pin patient probe connector. However, the manufacture's used different wire configurations. This resulted in several patients receiving burns on their finger at the site of the red LED. The human failure occurred when the manufacture failed to use a unique locking / attachment system for the patient probe AND human failure occurred when medical personnel switched the probes between units manufactured by different companies.

Chris Newmarker I'm actually kind of thinking more along the lines of those areas of expertise that are ignored. I've heard warnings, for example, that early stage companies don't have a firm grasp on materials science or molding issues. There's that kind of stuff. I've also heard the complaint that as OEMs increasingly outsource their supply networks, they're having trouble maintaining quality standards. You hear about a medical device with problems in a wide variety of areas. (Qmed has had a few recall stories like that.) It seems to suggest systemic problems in the supply network, which does not bode well. ... So what am I missing here?

Greg Smith There's a difference between ignoring what's needed and not having the experience to know what questions should be asked of suppliers. There's also potential issues around undertaking proper Risk Analysis throughout the product's life cycle. FMEA just doesn't cut it these days. I don't believe that there are systemic problems within the supply chain, but if you're not following good practice, not adhering to your in-house QMS procedures, or just plainly don't know what questions to ask your suppliers then you're in the wrong game. A simple analogy is plastic moulding; you can use moulders that make "plant pots", or you can use moulders that have (& adhere to) all the correct accreditations for Medical Device manufacture. Anyone can short-cut the system, but the risks to patients are immense - not to mention the morality of doing that.

Chris Newmarker So then it sounds as though taking shortcuts, whether to save money or rush to market or both, can be a big reason for failure? ... On another note, I had an expert comment on Twitter that this GE recall was probably the result of not applying QS and QA procedures enough: http://www.qmed.com/news/ge-recalls-infant-resuscitators-over-serious-assembly-error

Terrell Williams Most big companies know from experience at least, that device failures like this cost more than quality control. These failures are so visible, and that's why GE stopped making pacemakers years ago. Big companies tend to be extremely risk adverse. (The GE lead to pacemaker connector failed.)

Chris Newmarker Some really insightful comments here. Thanks everyone! So some of the reasons for failure I'm hearing include not taking human error into account when it comes to how people use the device, not looking out for human error among manufacturers in the supply chain, taking shortcuts, and not employing human factor research. Anything else that would be good to add to the list?

Eugene Kaganovich RN,BSN,CDE Chris, if you are interesting more about human factor research, please let me know and I will email you more articles.

Eugene Kaganovich RN,BSN,CDE Hi Chris,
Thank you for your question
I was trying to analyze this for the longest time, why we have so many recalls. I’m RN, CDE working in ICU; we experience many problems with IV pumps recalls and malfunction ICU equipment.
Personally, I used almost all insulin pumps available on the market. Recently, I have switched to t: slim insulin pump from Tandem Diabetes care. They developed insulin pump based on intensive research studies by using human factor research.
I described in more details here http://www.advice4diabetes.com/tslim.html
I’m using this pump for over 7 months and no problems so far. I was involved in focus group and surveys with development of new version of medical devices (insulin pumps). I'm seeing ONLY change of the name Minimed 522 to Minimed 530G, but the device has the same body, mechanism and safety check
Every company, which is planning to develop new medical devices, MUST switch to human factor studies.

I couldn't agree more with Eugene. There is also a group in Toronto that have done a lot of good work in healthcare human factors http://humanfactors.ca. If you need specific contact person, please let me know.

Daryl Mullins PMP, CSM, MBA, BSEE Chris, some history from the FDA's early 1980's 16-page document entitled "Device Recalls: a Study of Quality Problems" [HHS Publication FDA 90-4235 (January 1990)] will help your organizing of the contributions from Group readers.

The work was published by the FDA's CDRH and covers Medical Device recalls from October 1983 to September 1988. 3,144 recalls were done in the 5-year period and 1,664 Quality problems were identified. The root causes of these 1,664 problems were grouped into 4 categories - 1 was GMP (780 instances) and another was Pre-production (Design, Components, Software, Package, Label) a total of 724 instances + some fewer cause types to total 1,664.

It would be very interesting to organize your contributors' inputs in a similar way for comparison purposes.

There should be some new root cause categories given the additional technologies in use since 1990; the FDA's work could be a nice place to start your continuation of "why do medical devices fail?"

I think that this list can be applied to a lot of designs other than medical devices. Is anyone wanting to take up this study? I am interested

Yea. They don't even check out properly devices that can kill or make patients very sick...like these Plastic Transvaginal Mesh Devices that have adversely affected hundreds of thousands of women, and men, worldwide.

Albert de Richemond, M.S., P.E. From my experience in investigating such failures, the majority of medical device "failures" is caused by human error, i.e., someone did something wrong and the device takes the blame.
Medical devices are very complicated things. Manufacturer's take great care in endeavoring to minimize the risk of failure. And it is a rarity for some manufacturing defect to cause a device failure. However, defect do occur.
However, medical device failures are really system failures. The surgeon didn't read the instructions for use. The checklist was not followed. The device was improperly maintained. The device was improperly stored. The nurse input the wrong settings. The cleaner didn't properly clean the room and left infectious material behind. The disposable device used was out of date. The switchover at a shift change didn't communicate the vital information. A clinician installed the tubing incorrectly. The clinician did not fully comprehend the physiology of the procedure. And someone got hurt. In each one of these cases, the "failure was the last event in a series of events that led up to the final "failure."
It is encouraging that the medical community is starting to make checklists standard procedures that are always followed, to use root cause analysis find the process and procedural errors that led to "failures," to say, "I am sorry" and own up to the errors that culminated in the final failure.
There is a wealth of information in closed claims, closed cases, and clinicians' heads that should be mined to help educate clinicians and everyone else about medical device "failures." Until individuals and corporate entities are willing to make this information available, what we know about why medical devices fail will remain anecdotal and obscure.

Aron D. Haber Lack of reimbursement is a huge factor in product failures or under performing products. A company will not develop a product unless they are sure that an insurance company or the government is going to pay for it.

Surface modification can be a factor from the design perspective. A company representative from a testing company said that the reason in-vivo thrombogenicity tests fail is due to lack of surface preparation on the device. In many of the failures, the surface looks like the Rocky Mountains under SEM.

Eugene Kaganovich RN,BSN,CDE For over 14 years working in critical care, I saw many medical devices which were not user friendly. Before we start using any medical devices we have mandatory classes. Malpractice is nothing to do with poor design of medical devices.
This is only my opinion that all companies must implement human factor in development of medical devices
This is FDA link - Medical device use-safety: Incorporating human factor engineering into risk management.
http://www.fda.gov/medicaldevices/deviceregulationandguidance/guidancedocuments/ucm094460.htm
“There have been numerous examples of infusion pump recalls due to poor designs. For example, in August 2004, the N’Vision Clinical Programmer was recalled by the Food and Drug Administration (FDA) because users were able to enter infusion times with incorrect units. Minutes were entered into the hours field, which caused patient overdose. The Colleague IV infusion pump could shut down while delivering critical medication to patients. The reason for this was that the “on/off” key was so close to the “start” key that nurses would often inadvertently turn the pump off when they intended to start delivery. Over 206,000 Colleague infusion pumps, used mostly in hospitals, were recalled by the FDA” ( Schaeffer, 2012)

Reference:
Schaeffer, N. E. (2012). The Role of Human Factors in the Design and Development of an Insulin pump. Journal of Diabetes Science and Technology, 261.

Terrell Williams It's well known that tens of thousands of US patients die every year at the hands of doctors who put the pacing electrode (lead) in the apex of the right ventricle. In today's society, it is apparently OK for the pacemaker to cause heart failure, as long as the device didn't fail. But to my mind, the patients are just as dead. Today, it's just as quick to position the lead so that it paces the cardiac conduction system according to two abstracts at HRS last summer. The location of choice is called the His bundle, the ventricle's nerve like structure which choreographs the ventricular contraction normally. The way I see it, pacemakers fail to sustain life much too often (barbaric).

Background. There are several studies that are well known to EPs: CTOP, a Danish Study, DAVID, MOST, and a sub-study of MADIT II, etc. They all lead to a similar conclusion.

For example, MOST, a 2,010 patient study is summarized in Table 2:
For cumulative ventricular pacing of <10%, the heart failure hospitalization was 2% For cumulative ventricular pacing of >90%, the heart failure hospitalization was 12% Suggesting 10% heart failure hospitalization due to high burden ventricular apical pacing.
(Adverse Effect of Ventricular Pacing on Heart Failure and Atrial Fibrillation Among Patients With Normal Baseline QRS Duration in a Clinical Trial of Pacemaker Therapy for Sinus Node Dysfunction; Michael O. Sweeney, Anne S. Hellkamp, Kenneth A. Ellenbogen, Arnold J. Greenspon, Roger A. Freedman, Kerry L. Lee and Gervasio A. Lamas; Circulation 2003;107;2932-2937)

Chris Newmarker So here's an interesting exercise to provide more food for thought... I recently wrote the following article about 7 medical device failures recently catching the FDA's eye:
http://www.qmed.com/news/7-recent-medical-device-failures-catching-fdas-eye

Two of the cases apparently involved poorly written instructions.

* FDA is warning that a Greatbatch Medical Orthopedic surgical tool could cause serious injury or death because it came with inadequate sterilization recommendations.

* is is updating labeling and training materials for the HeartMate II LVAS Pocket System Controller because some patients and caregivers experienced difficulties with the process of changing from a primary system controller to their backup system controller. Four patient deaths have resulted. (Some patients apparently tried to change the controller on their LVAS while they were alone. Sounds pretty frightening.)

Two seem to involve GMP.

* Medical Systems has issued a recall over Custom Procedural Trays/Kits containing Hospira-produced 1% lidocaine hydrochloride injection. Hospira itself is recalling one lot of 1% lidocaine hydrochloride injection contained in its Custom Procedural Trays/Kits after it confirmed that a customer found particles of oxidized stainless steel in a container of the cardiac drug.

* Healthcare is recalling a host of infant resuscitators from the past five years over a serious error during the assembly process. The FDA announced Feb. 28 that it has designated the recall as Class I. FDA officials think the resuscitators could potentially seriously injure or kill infants because oxygen and air wall inlet fittings on their back panels were reversed during assembly.

Two could be GMP or a supplier issue, or maybe both.

* 600 Philips Respironics Trilogy Ventilators could cause serious injury or death because of a power management board issue that could cause the ventilators to not actually deliver mechanical breaths, the FDA recently reported. Philips Respironics says it became aware of a problem during production testing, discovering that the Trilogy ventilators contain a potentially defective ferrite component on the power management board of the device.

* FDA last month designated a Medline Industries guidewires recall as Class I. The guidewires in question have the potential for the coating to flake off of the wire, according to the FDA.

And my guess would be this is poor design. (Though could it be a supplier issue, too?)
* Park Triangle, NC–based Teleflex Medical, part of Limerick, PA–based Teleflex Inc., has a tracheal tube that can apparently kink during patient use.

I especially think the instructions issue is interesting. The device itself might be safe, but if health practitioners or patients don't understand the risks, etc..

Burrell (Bo) Clawson I want to be realistic after 4 decades of designing medical products, so I will state categorically that their are an infinite number of ways medical (& any other products) can fail.


At the most basic level, materials have limited lifespans (environment & wear), people use products wrong or for the wrong job, use conditions can be outside the design envelope, handling, storage & reuse can damage items and properties of materials have variations which may cause shorter lifetimes, equipment with software needs to always "fail-safe" (but doesn't) and that leads to a corollary, in that all devices have useful lifetimes after which they must/should be discarded.

Once you start looking at a specific product that is complex in numbers of pieces &/or software, it becomes obvious that the ways of failing can be nearly infinite. Hence, a designer tries to eliminate the obvious modes of failure by design, over-design, and fail-safe choices.

A lot of things which are critical are deemed to be single-use items, just because of the problems with failure that can occur if you reuse them. That causes a lot of discussion as reuse is constrained by FDA regulations, but some hospitals do it anyway, against manufacturer's stated use conditions. In other words, the users can beat any great product until it "dies", whether it is reusable or not.

In the end you need a very experienced engineering team to get close to a "failure-proof" product, but you will never achieve it. There is value in being a "grey beard" in complex design choices.

Nick Yang My name is Nick.I work at ICN in one of the biggest hospital in california. I would like all medical instrutment companies to know you guys think the device itself might be safe, but if health practitioners or patients don't understand the risks, etc.
How do you make health practitioners or patients to undersatnd the risks?
If it is high risks and complicated, your device will be fail in market.
In your user manual, teaching class, do you use a lot of enginnering terminology?
After you sell the device, how often you go to visit the facility to undersand user problem?
If you don't consider user advise, complaint and improve your device, your device will fail in the market pretty soon.

Burrell (Bo) Clawson Yang brings up a good specific point on complex systems.

I designed the pneumatics @ Bourns in the early 70s for what became the Bear Ventilator. It was a monoblock system which mean there was almost (famous last word) no way to misconnect the "block". If it malfunctioned, the whole block could be switched in minutes without a spagetti mess of tubes.

If all AC power failed, it was still designed to "fail open" with pneumatic valves that let a patient breath on his own if he could, and the opposite if a pneumatic failure occurred the system automatically dumped any overpressure gas.

I didn't design the final package, which occurred after a management revolution. The pre-production version performed flawlessly for about 6 months.

So what was the bad failure mode for the final production ventilator? An alarm sounded, a nurse came in to check the patient and ventilator and the first thing they did was cancel the alarm, then take care of the situation. Nurses would then leave, but the alarm was still cancelled. You can guess what happened next. That was a software choice that should not have been made to allow an alarm to be cancelled permanently.

That brings up the next situation when alarms are set to go off with minor changes and as a result, you walk into a cardiac recovery room and 4 alarms are going off simultaneously.

Good design in such things is very complex.

Miguel Angel Torres Tello Hi Chris,

In LA countries, labels must follow certain local norms. These are not all the times complied or correctly assessed in customs. Then, we ended up with labels that are difficult to understand for the final user -becoming a possible cause of misuse.

My suggestion would be to bolster the risk assessments not only for the country of origin but for all the other markets the manufacturer is going to sell. In other words, if manufacturers updated their risk assessments considering the new markets they get into, less human errors worldwide would be the cause of failures.

In case an assembly, packaging or design part turned out to be the main drawback, manufacturers could reinforce the security through updated risk assessments.

In conclusion, we might say that a good, updated risk management systems for local and foreign markets might be a good strategy.

Terrell Williams I wrote earlier today about patient deaths associated with pacemakers. Some may not understand the root cause. If interested see <http://PacemakerPatientAdvocacy.com>

Does FDA care? I doubt it. FDA is about controlling the market, not the marketplace.

Chris Newmarker That's a great point, too, Miguel, that people raised in different cultures might react to similar situations or read the same instructions in different ways. Just because something is understandable in the U.S., doesn't mean it's going to work in Mexico or China or India. Any good examples of this?

Miguel Angel Torres Tello Hi Chris,

Thanks for your comments.
About some examples.
I have went through many MRI manuals that are quite difficult to understand in Spanish. Therefore, service engineers must struggle with the English versions (if they are not Enlgish readers) or with the Spanish rough translations.
In this case, engineers receive training but the real fact is that manuals are not a real guidance -for the service engineer or the operator.
For instance, at hospitals, I saw several times that operators (radiology technicians) fails on using the system properly -being the cause of more than one quench.
Based on this, a good risk assessment would consider:
-Work on better translations for the service engineers.
-Evaluate who is the main operator and give easy guidance in the local language for them.
-Record any feedback from service engineers and evaluate it in the QA departments.

I have other examples but this should be the most representative for the discussion.

Training

Ali Hamadeh Well they fall for too many reasons as every one seems to agree but in my opinion the number one reason would be:
"lack of test coverage for exceptions use cases and user errors".
Let me elaborate, Medical device manufacturers must perform verification and
validation tests to objectively demonstrate that the device meets its engineering requirement and its intended use. However, these tests are in most cases focusing on what the device do and how it will function in nominal conditions by end users. To oversimplify, most tests would state: "Do A, Press B and make sure C happens". Such tests are the bare minimum to show compliance but what is missing is what the device do and how it functions when I press B before doing A or when I do A and press B at the same time ..etc
Such tests are referred to as "exceptions handling" as they cover un-intended scenarios. The medical device under such exceptional conditions shall continue to function or at least fail safely meaning without causing harm to the end user or the operator.

Some of the medical device makers that I've worked with seem to focus more on meeting the regulations than in designing and manufacturing a reliable, effective product. This mechanism leads to other causes that have been covered by others on this thread.

Kurt Yockey A part of this controversy involves companies that know their devices are being used off label, and allow patients to be test subjects without knowing they are test subjects. Novel uses can provide new therapies, but such uses can create "failures" as well. Often these situations are identified by manufacturers as "user error" yet they have not done anything of substance to avoid off label use.

Some "user error" results from the strength (or lack thereof) of the training materials and program to which the physician is exposed prior to use. Warnings such as "don't over-tighten" can be useless to physicians if there is no information on how to appreciate that one is getting close to the "over-tighten" condition users are warned to avoid.

Lastly, the complaint that no one can anticipate every possible "failure mode" may be allowing themselves a justification for not taking the time, or use of experienced people to review a product and it's training program prior to launch.

All good stuff. Publio hits on an important point - when we implement compliance to a regulation or standard, it almost always results in lowering the bar. Standards and reg's are written intentionally to give the minimum acceptable requirements, and too often this is far less than we would have been doing on our own initiative, had the standard not existed. It is difficult to justify to the bean counters why we need to go over and above this or that minimum requirement, but someone has to stand firm and say, either we spend some money now to do it right, or we spend much more later to do it over.

From a regulatory point of view, many standards, when giving the minimum requirements, will also require that you perform Risk Analysis to determine whether the minimum is acceptable for your product, for your uses, in your use environment. Are we all as diligent as we need to be on this?

Another important point, touched on by several good comments, is that the designers need to take into account the end-use environment, especially the user. This is a key concept in reliability engineering, often given lip service and seldom the attention it requires. It's part of the accepted definition of reliability, which most of us could quote verbatim, and it's a huge wild-card when calculating reliability predictions. But do we really get to know the user and use environment sufficiently, or are we using erroneous assumptions?

We've all heard of attempts to make something "idiot proof", and attempts are made with human factors engineering and various analysis methods, but based on my experience, I would posit that one cannot begin to imagine the breadth and depth of human behavior as it relates to our products. When you actually put it into a user's hands, and repeatedly have them, the users, operate it, you begin to see an astonishing variety of failures, many of which would have been called impossible by the designers. Doing this pre-launch will reduce field failures and returns.

How is "failure" defined? Financial, material, mechanical, unintended consequences, didn't work as intended, didn't work period, user error, efficacy, sales, marketing, logistics? The med device industry has as many pitfalls as devices.

Mark Bly I worked for a biomedical company that made a device for wound care. They had no data to substantiate the claim their product healed wounds. Therefore, they had no CMS reimbursement code and were relegated to selling to the VA system only. They moved from wound care to pain management and had no data to substantiate the claim the product worked for pain......still relegated to the VA system because they had no CMS coding. This company enjoyed a steady cash flow because the VA pays for services used and products purchased within seven (7) days of the date it was used.

The upshot, this product was a pulsed radio frequency device that had no randomized control data to prove efficacy. The VA system in a small portion of the country, has embraced this product even with no data to prove it works. However, a great portion of VA's have stopped doing business with the company who makes the device and the company, when reps are fired or they leave, will no longer fill the positions due to underperformance and underutilization of their product.

Summary points:

1) The studies the company has launched ended with poor results and each of the studies were ended early.

2) If a company is going succeed in the medical device industry, they need to have conclusive evidence from a large population of patients to prove the money spent on the product is well spent and the larger number of patients are benefitting from the product.

3) Our VA system needs to choose products more wisely. The product referenced is an under performer and our veterans are being sent home with a product that in 33% of patients does a marginal job of healing if it works at all and in another 33% the product fails miserably. The remaining 33%, the product has proven effective.

4) the FDA needs to recognize companies and products with marginal track records and eliminate the market.

5) When the CMS approval is denied multiple times and/or the company cannot produce excellent data in a multiple well designed trial, the product should be pulled from the market.

6) is a 33% rate of healing enough to justify allowing a product to be used on a large population of patients who are seeking relief from either pain or wounds. Why should our veterans and our our private patient population be subjected to shoddy treatment practices?

In summary:

The cost of marginal products in our system today is a burden that drives poor patient care, extends the treatment cycle and delays good healing practices. In some cases, can actually increase complication rates and cause more patient discomfort. The goal for companies such as the one referenced is to provide a return to private investors and venture capitalist. The secondary goal is high quality patient care.
This serves as a round about lesson to the question of why products fail. The fact that many companies enter into the medical device market with the sole intent of the founders becoming rich at the expense of the system is a primary consideration.
We as tax payers pay for inadequate treatment prescribed by systems such as the VA.
Our veterans and our private pay population of patients deserve care from products with a much higher rate of success.
Our FDA and our clinicians should heed the warning signs of ineffective products that burden our system.

Ian Nemerov Great comments. I'd add one comment: culture.

When Toyota had a rash of recalls a few years ago, one theory was that, despite its legacy as a leader in quality culture, the company had in fact morphed to a culture where bad news was getting ignored. In the financial system meltdown a few years ago, one problem that was found in many companies was a quick dismissal or minimization of risk (or even identifying the risk-creator as a hero). Similarly, many device failures can be traced to causes that were known or knowable, when they could have been addressed much more simply economically, but the manufacturing culture did not value paying attention to that risk.

A contrary story, where embracing risk issues led to company success, occurred at Ford when Alan Mulally joined as CEO. Mulally instituted - and participated in - weekly quality reviews. Typically, the quality reviews were filled with green indicators, as was routine before Mulally joined the company. However, one week, after a recall, one indicator was red. When the red indicator was shown on the screen, the audience thought the presenter was going to be reamed. Instead, Mulally congratulated the presenter for raising the red indicator and said something to the effect of "That's the way to do it." The next week, a rainbow of indicators was presented. We all know the positive changes in quality and business results that have happened at Ford under Mulally's leadership. I would argue that one of the most important things he did to cause the turnaround was to change the company's culture by encouraging open discussions about risk and quality.

This issue is being discussed at leading hospitals, such as Johns Hopkins. They recognize that risks that lead to massive impact, such as a problem with a commercial airplane, get addressed immediately, but that risks that may only impact one or two patients get minimized, even if the aggregate of the risks that individually could only impact a few patients may collectively result in hundreds or thousands of patient issues each year.

I had the honour to manage 2 start up of dental medical devices even if in both cases they failed.
Reasons for failure:
Business model not attractive
other breakthrough technologies
Lack of vision of the Board of Directors
product not designed for the customer target
Manufacturer cost too demanding
Regulatory or clinical issues
R&D/Clinical/Marketing/Sales competition and different vision
Lack of competitive environment understanding
Focus on short term result and not on the long term
No risk attitude from Senior Management

Gordon Millar Having been inundated with approaches from IT specialists, all of whom claim to have developed the latest innovative "app" for patient self diagnosis/monitoring/treatment, none seem to have included a clinician or medical devices expert in the development. That's a sure formula for failure and demonstrates inadequacy in product development which any potential investor would question - that's before any submission to the regulatory bodies.

Gordon Millar One other comment I would like to make is that on occasion, new and innovative devices are sometimes evaluated by people with little actual understanding of their potential. Investment company personnel and here in the UK, Gov departments, are responsible for quite large budgets but without an in depth appreciation of the market, application or requirement, they do frequently miss an opportunity or demand a ridiculous ROI timescale. Sadly, designers and inventors are dependent on such people as they cannot fund the ever increasing development costs without proof of concept or seed capital and as a result, I am sure that many potentially viable products fall at the first hurdle. Cost of development and patenting is high, cost to validate can be phenomenal and the cost of "failure" is so prohibitive that I am sure it discourages many.

Andrew Clay Does any product launch without defects? What is "good enough"? In my experience we sometimes weren't able to build a sufficient quantity to do testing. You find defects the more you build, the more time manufacturing has to wring out process, the better. Then the customers do all kinds of crazy things that you didn't think about, even if you did great market research. I think good risk assessment, good beta testing, and staying close to your market are more important than trying to get to zero defects.

Gunalan Dass poor schedule service and user mistake one of common problem and recall of product never been notice to user or lack of knowledge on details need to be up to date on medical device.

updates and age and other uunproper usage are the causes of failarity.

I am a physicist working at the confluence of academia, industry, and the clinical community and have worked in start ups, universities and large industrial research labs. One common trend I have seen a number of times in the early stages of medical device innovation is the difference in view points of clinicians and that of the market. My first foray into medical device development came about from an approach from a clinical research scientist who was very well known in his field, had a formidable publications record and the need for a new form of GI diagnostic tool that could be used in kids. At our first meeting he assured me that the device would be a game changer and would be embraced by the global clinical community and result in 10,000’s if not 1,000,000’s of sales per year if we could fabricate a device with 12 sensors spaced 1 cm apart and meet a cost of $500 per use. This was enough to apply for grants and kick off our development project and within a year we had a device ready for testing. Technically the device was ok, but in every other aspect we were way off. The acceptable cost per use was governed by reimbursement codes for similar procedures and was closer to $50 than $500, the actual procedure being targeted did not even have a reimbursement code at the time, 12 sensors rapidly morphed into 36 sensors as a minimum, and the annual sales for the nearest equivalent device turned out to be around 800 units per year.
So, in this case at least, we failed at a very early stage by not doing our homework and by taking the word of a single clinical expert without validating his obvious excitement and passion for his particular area of research by going to the wider clinical community for advice.
I guess that the moral of the tale is that before you put any money and effort into developing a new medical device make sure that the problem you are solving, no matter how important it is for the patients involved, and how exciting it is for the researchers, is actually one that can also support a viable business plan into the future. Ultimately ‘Cash is King’ in start up land and if you can’t bring in more $$$ than you are spending, your device will fail.

Hi! Very interesting question, and great discussion here, thanks everyone.

My two cents... As someone involved in product management and launches and who's seen a number if different types of devices ( simple to more complex, with software, single use to multiple use, etc....). The below is to add to, develop some previous comments by other members and by no means structured or extensive. It is intended as food for thought from my limited experience...

- I have observed that extensive and repeated training are indispensable for some products. Having the capability to provide such training can become a strategic advantage to a company and may determine which markets one should consider launching in.

- For products that are challenging an existing one on the market: user habits is something to watch. Sometimes, an objectively neutral functionality becomes a source of error because clinicians have learned doing things a certain way with the first to market product. Even an attempt at making things more user friendly could go wrong if it goes against the habit formed by the previous product. I've learned that "user friendliness" and "intuitive design" are very subjective notions...

- The level of experience of the users matter. Device companies usually do pilot testing with a limited number of very experienced key opinion leaders and once launched, the product goes to everyone...

- Finally, to elaborate on another similar comment made before: is a device ever fully ready for market at launch? Unlike pharmaceutical products that give relatively predictable outcomes, device outcomes depend entirely on the human interaction... And I have observed that as soon as a product gets launched, even after extensive testing, we start finding potential ehancements and improvements to the design... Medical devices are almost like living organisms, in a sense... They are supposed to evolve and change after their first launch. The key for success, for a company, is listening and reacting appropriately...

Looking forward to reading further comments...

Patrick Mize, PhD Chris:

Fine question and many fine answers.....

There is a reason the FDA and the EU push QMS, ISO 13485, Design Control, Design Specifications. ISO 14971 Risk Analysis. Planning.

It is the thought that is put in ahead of time and then doing a really good job of risk analysis that will make or break a product launch. Principle members of the company not really understanding how to manufacturer a quality product that is sufficiently tested. Developers that never really talk to their actual customer, be it a professional or a lay person.

A 30 yo developer looks at the world a great deal differently than a 70 yo ex trucker or non technical person. But if that is your customer you better understand their needs and abilities fully before starting.

I learned vastly more in 4 hours about use of POC IVDs being a patient with chest pains in the ER than being in the lab for 20 years.

The best products can´t be used incorrectly, they don´t need or have a minimal need of instructions. They are designed work correctly. The last and worse place to fix a design problem is to put an explanation or warning into the instructions or package insert.

Products fail because the design process was ignored.

Want to see a great product, look to Dyson and the vacuum cleaner. There can´t be a more prosaic item on the market. Thousands of designs, 100 years of history. Dyson actual builds a product that is hard to misuse, fun and intuitive to use and amazing picks up dirt, it actually works!

I bet they have a great design team, detailed and meaningful risk analysis, and plenty of consumer contact and feedback.

It really is not rocket science that makes a good product, just good science and attention to detail.

Patrick

Chris Newmarker This has been an excellent discussion. Thanks everyone! I just posted a blog post summarizing some of your views: http://www.qmed.com/mpmn/medtechpulse/9-ways-medical-devices-fail

Ali Fouani Good question Chris. Let's try by asking "Why 5 times".

New on the blog: U.S. Food and Drug Administration Provides Medical Device Development Tools, Improves Evaluation Process. The tools are a great resource for those looking for materials pertaining to FDA approval requirements.

http://www.smckyems.com/fda-provides-medical-device-development-tools-improves-evaluation-process/