• Decrease font size
  • Return font size to normal
  • Increase font size
U.S. Department of Health and Human Services

Medical Devices

  • Print
  • Share
  • E-mail

Infusion Pump Workshop: Transcript for May 25th, 2010 (Part 1)

FTS-HHS FDA CDRH
Moderator: Melissa Eakle

May 25, 2010
6:00 am CT

Melissa Eakle: Okay. Good morning. My name is Melissa Eakle. I’m the Network Leader for the Infection Control, Dental, General Hospital and Infusion Pump Network at the Center for Devices and Radiological Health and also the Chairperson of the Infusion Pump Working Group.

I’m going to be doing Master of Ceremonies duties today and introducing our speakers that come to speak.

Our first speaker is going to be Mr. Tony Watson and I want to tell you a little bit about Mr. Watson. He’s the Director of the Division of Anesthesiology General Hospital at the Food and Drug Administration.

Prior to his present position, he was Chief of the General Hospital Devices Branch in the same division for five years and the Acting Deputy Division Director for nearly a year.

Prior to assuming his role as Branch Chief, he was an engineering reviewer for over ten years in another division. He graduated from the United States Naval Academy with a degree in general engineering and he served as a naval officer before being honorably discharged and joining the FDA in 1994.

He has a Master of Science Degree in Management Information Systems and a Masters of Business Administration from the University of Maryland, University College.

Please welcome Mr. Tony Watson. Thank you.

Anthony Watson: Good morning. Technical difficulties notwithstanding, pleased to see everybody here this morning. I believe that today is going to be an important day in infusion pump history.

As Melissa mentioned, I am the Division Director of the Division of Anesthesiology, General Hospital, Infection Control and Dental Devices. It’s quite a long name, but we have primary pre-market review responsibility for infusion pumps.

We need better infusion pumps. We need infusion pumps that are more reliable and have more intuitive interfaces.

I think reasonable people could probably say that this workshop is too long in coming even though there have been other workshops similar to this before. It’s the culmination of a series of signals, investigations, recalls and decisions that probably predates most people in this room, definitely predates me.

We have a story to tell about why we are here today. I ask that you listen to the story and each presentation so that we can learn from each other. Each group of people here has to be an active better for us to get better infusion pumps on the market; regulator, manufacturer, user, and healthcare administrator.

That is why I am thrilled to have AAMI, AdvaMed, our regulatory counterparts from other countries and members of the user community here today. You will hear from some of these groups today.

Please indulge me for a moment as I recount significant events that brought us to this moment today. We would also be interested in hearing your stories of epiphany regarding infusion pumps.

My involvement with infusion pumps goes back to 2004 and my first month as Branch Chief for the General Hospital Devices Branch. I was forewarned in my first few days as Branch Chief by then Division Director Chiu Lin - Dr. Chiu Lin that I would be inheriting infusion pumps; a device area with -- as he described it --a challenging history.

I think that could be an accurate statement; challenging history.

The first month I went through a major recall for an infusion pump event that resulted in a death that could likely have been detected by just a small focus group evaluation prior to marketing.

The case was so compelling that it planted the need for human factors testing in the minds of us at FDA. This pattern of missed opportunities repeated itself over and over again. In 2006, we began to consistently ask for human factors testing for every infusion pump submission.

It was not easy. This change met with resistance from some manufacturers who questioned our authority to make this change without guidance. There were other manufacturers who recognized the importance of the issue and simply conducted the testing.

Word spread and over time the new became the old, became the norm. Now human factors testing is routine.

Closely related to the human factors issue was a software issue. We also noticed that a large number of infusion pumps had problems that were related to software defects.

Part of the problem was that we had not consistently requested full documentation of the software as we had promulgated in our guidance document for software.

So we began to request this information in full. Again, met with some resistance. Some manufacturers had become accustomed to promissory notes promising to validate the software prior to going to market.

And again, other manufacturers simply complied and just provided what we had asked for. So, once again, the new became the norm. Promissory notes for software validation are a thing of the past with respect to infusion pumps and that’s a good thing.

In the midst of these changes, we were going through a recall situation that would change the landscape of infusion pump reviews permanently. This particular recall situation was so provocative and resource intensive, it was transformative.

In fact, in this recall situation the seeds of our present guidance document were sown.

By mid 2006, FDA management was ready for a fresh look at these devices and created an interoffice group to find a way to reduce the number of recalls. This group engaged in s systematic evaluation of the entire pre and post market review of infusion pumps.

This was not an easy nor unanimous process. There was significant debate about the proper courses of action. Believe me when I say every option was on the table.

In the end, we all agreed that we needed to just do what was right and ask for the information to improve these devices, at least from the total product perspective.

This information is different, very different than what we’ve asked for in the past and, in fact, it is pioneering in the fact that we produced a total product lifecycle guidance document whereby we ask for information that has typically been reserved for the post-market arena in the - into the pre-market stage.

The information we are requesting will challenge not only the manufacturers but the FDA staff, as well. We also agree that recalls were never - will never be completely eliminated but this effort will put us on the path to better infusion pumps.

Forty years of hard work brings us to this point. Let me pose some food for thought.

Many of you I know flew here today. I am certain that most of you do not think twice about whether you would make it here, but if you did, you quickly dismissed it because we take flying for granted.

Chances of a malfunction are rare and it is even more rare that you would experience a mishap that would result in injury or death. In fact, it’s so rare that it is newsworthy when a plane has to land uneventfully without so much as a scratch on any passenger.

If we had flying mishaps rates comparable to infusion pump MDR rates, many fewer people would fly. Badly it might become so common that it would not even be newsworthy.

Thankfully for us, airplanes are considered safety critical systems and designers and regulators take the time to put safety first.

Does it not make sense that we as designers and regulators take the same approach with another safety critical system, infusion pumps?

I ask you, can we ever see the day when infusion pump mishap is actually newsworthy?

I believe we’ve put together a program that puts FDA’s thinking and supporting data on the table for all of us to see. We hope this workshop will facilitate a frank exchange of thoughts and ideas on how to get a better infusion pump.

Along those lines, the tendency is naturally going to be to gravitate toward the infusion pump guidance as the centerpiece. Indeed we have accommodated this by allotting time to discuss the thinking behind it in the agenda.

Although we believe that the guidance is a pivotal cog in the machinery, we do not want this workshop to be all about the guidance. We want to know your successful best practices around infusion pumps, in use or design, we also want to hear your challenges.

We have to learn from best practices, as well as any challenges. I want to close by saying that this initiative is not about blame. It is about the infusion community circling our collective wagons and engaging a challenge.

I attended an AAMI meeting in March in Washington where I had the opportunity to meet members of the infusion pump manufacturing community.

Based on the commitment, energy and enthusiasm of the people in that meeting, I left completely convinced the critical mass for change has been reached.

So I applaud you for taking on this challenge, all of you. We are sure to have growing pains. I am optimistic, though. I probably have no right to be, but I am, in fact, optimistic.

Because when I look at the people I work with, when I see the commitment of the members of the group at AAMI and I work at the comments sent to me on this initiative from all corners, it seems that there is so much ingenuity, resourcefulness and perseverance. We really have no business failing.

Thank you.

Melissa Eakle: All right. I’m going to introduce our next speaker. This is Lieutenant Commander Mary Brooks. She’s been with the Food and Drug Administration since 2005. She’s currently in the post-market area as an analyst with the Center for Devices and Radiological Health.

She’s also been a pre-market reviewer with the Office of Device Evaluation and a collaborative reviewer between the compliance and evaluation offices.

She is also a member of the home healthcare committee since 2006 and prior to that, she served as a combat medic, EMT and ER Tech at Fort Sam Houston, Fort Belvoir and in Seoul, Korea.

She also an active duty officer of the Public Health Service and has provided emergency response for 9/11, the Pentagon Attack, Hurricane Katrina, Rita, Wilma, Gustav and Ike, and she’s been a trauma responder for President Ford’s funeral President Obama’s inauguration.

Please welcome Lieutenant Commander Mary Brooks.

Mary Brooks: IT issues. I want to make sure that my slides get up there.

Thank you so much and the audience looks very familiar from yesterday and I was telling everyone today that the workgroup yesterday with the homecare was fabulous. So we really appreciated all your comments and the feedback that we received to help work on a guidance document.

Making sure they got my slides.

Melissa Eakle: I wish I had some jokes to tell you because obviously Murphy’s Law is when something can go wrong, it does go wrong and at the worst time. So I apologize for the delay. We will be with you as soon as possible.

Mary could talk but she really has a lot more impact when you can see her slides when she talks.

Mary Brooks: There I am. Okay. There we go.

Wonderful, so everyone in the back can see, I hope. I see heads nodding. That’s good.

Well my name is Mary Brooks. I’m an analyst in the post-market surveillance group, (TEB) Number 2 and as Tony mentioned briefly, when we were looking at the guidance document, we incorporated post-market information into it.

We normally do it any way but in the guidance document, we provided the information as part of the background. So this is us looking at the total product lifecycle and trying to explain throughout the guidance document why we’re asking for the information that we’re asking.

And we were hoping that by providing post-market data and some of the adverse events that we’re seeing in device failures that it would be a little bit more clear as to why we’re asking some of the information in the guidance documents.

I do appreciate everyone being so kind about the technical difficulties today.

Next slide please.

So because I know today our audience is a lot of manufacturers, people on the webinar still need a general idea as to the whole background of MDR than post-market analysis, so bear with me.

I know this is going to be dry for some of the people in the audience that do this day in and day out, but there are several hundred people on our webinar that are clinicians who have no idea about MDRs.

So who has to submit MDRs to the FDA? We know that manufacturers must report MDRs adverse events to the FDA no later than 30 days from the day they receive or otherwise become aware of the information from any source that reasonably suggests that a device they market, one, may have caused or contributed to a death or serious injury, or two, has malfunction and this device or a similar device would be likely to cause or contribute to a death or serious injury if the malfunction were to reoccur.

Next slide, please.

We also know that user facilities must report to the FDA within ten working days of any death. We also know that if they are unsure of who is the manufacturer of their device, that they must file a report with the FDA if that adverse event was a serious injury.

Next slide please.

And we also know that voluntary reports are received through the FDA through the MedWatch Report Form. These reports are received from patients, family members, physicians, clinicians and also the public in general.

Next slide please.

So what can MDRs be used for? Well they help us provide a qualitative snapshot, a snapshot of adverse event reports for a specific type, a specific device or device type.

And these types can be malfunctions, they can provide a severity of the events, they can provide patient injuries as a result of the device failure and sometimes these injuries are not immediately following the device failure.

And it also provides a snapshot as to when the device failed, what additional therapies were needed.

Next slide, please.

So this is the latest data from 2005 to 2009 of all MDR reports received by the FDA. At the very top, I don’t have a little pointer but on the far right you’ll see 2009 data that we had a little over 360,000 summary reports; summary reports to the FDA that are very shortened reports that manufacturers can send in to us.

And we had a little over 200,000 of individual reports.

Next slide please.

So where do infusion pumps fall into the category? Today we’re talking about external infusion pumps. Within that category, we have large volumes better known as ambulatory infusion pumps by some. We’ll refer them today as large volume infusion pumps.

In all of the reporting between 2005 and 2006, large volume infusion pumps ranked Number 6.

When we continue to look at the data for that five-year period, we also noted that insulin pumps ranked Number 8. So as far as the categories, large volume ranked Number 6 and some pumps ranked Number 8 for all adverse events reported to the FDA.

Next slide please.

So when we looked at the 56,000 reports that is in the infusion pump guidance document, we broke those down. We looked at who’s reporting to the FDA in that snapshot of the five years and we noticed that it was a little over 54,000 were manufacturers.

We also noted that 850 were user facilities and a little over 1000 were voluntary reports.

Now what I will notice is that we have a very small amount of user facilities reporting to the FDA. We know that it’s a requirement for them to file adverse events with the FDA.

So I know that our room is full of a lot of manufacturers, but we do have clinicians on the phone, so you are required to report to the FDA for your adverse event reports on the 3500A and I will actually be talking pretty in depth to that tomorrow.

Next slide please.

So then we broke down those 56,000 reports into event types. We know that 62% of all those reports were malfunctions. We know that 34% of the reports were injury and we know that 1.26 or 710 of those reports were associated with a death.

Next slide please.

So then we broke out the data for that five years and then we said well how many of those 56,000 reports, how do they break out by the pump type itself?

So we know that we have large volume infusion pumps represented 47% of all the adverse event reports. We know that insulin pumps represented 45% of all the adverse event reports, enteral pumps were 3.9% and elastomeric with 2.8% and PCA, patient control analgesic pumps, were 1.6%.

So let me just talk in brief about these in categories of infusion pump. We have large volume infusion pumps, also ambulatory pumps. They’re the most common type of infusion pump used in clinical practice, and they can infuse up to large volumes of medication such as IV fluids, antibiotics and chemotherapy.

They’re used in hospitals, outpatient surgical suites, free-standing infusion centers and also in home health. Infusion pumps cross all levels of healthcare.

Our next category is insulin pumps. Pretty self-explanatory; it’s used to deliver insulin and they’re mainly used in the home health setting.

Next slide please.

We have patient controlled analgesics, PCA. They’re mainly used for pain control in post-op patients and they’re used in patients and surgical centers. These particular devices that I’m showing here today actually have a bolus selection button where a patient has breakthrough pain they can give themselves a bolus of medication.

Next slide please.

And we also have elastomeric pumps. These are widely used for the antibiotic and chemotherapy administration in the home health setting and also in nursing homes.

Next slide.

We also have enteral pumps. They’re commonly used for patients who need continual nutrition replacement therapy. They’re used in hospitals, nursing homes and home healthcare.

Next slide please.

So looking at the 56,000 reports that supported - that was in use in the background section of the guidance document, we looked at a 10,000 foot view of the adverse event reports by providing the - that manufacturers reported the top device problem code.

Out of all the reports, 96%, close to 97% of the reports were from you, the manufacturers. We know that 92% of those reports covered two types of pumps which were large volumes and insulin pumps.

So we got two large pumps in play here which are your large volume and your insulin pumps and we know that the majority, 97%, of our reports came from the manufacturers.

Now some of the challenges of the MDR reporting is that we really don’t get enough data. These forms as many of you heard yesterday are difficult to fill out and there’s a box for everything to be put in and not everyone is diligent.

We have manufacturers who are fabulous at report, absolutely fabulous. We have some manufacturers, not so good. So there’s always room for improvement in reporting. We all know that.

So looking at the top device problem codes, we - these are just the top listed ones that kept coming up over and over again when we did our analysis.

And I have to thank (Audrey Morrison) and (Katie Cummings) for helping with this analysis. They’ve worked diligently over the last six months pulling data and looking at the data in a zillion different ways. So I really want to give them credit for providing the information and the in-depth analysis that they’ve done.

So we’ve looked at the top device problem codes. They were device displays in error message unknown. Manufacturer or the reporter was unknown as to why the device failed.

We have broken components. The device needs to be replayed. We’ve had battery failures, alarm failures over and under infusion issues and also human factors and use error.

Next slide.

Then we looked at the top reported patient problems so when you have an adverse event, what were the top reported events to the patient? Well we know that the large majority that there were no consequences or impact to the patient. We’re happy to see that. We’re very happy. So the device failed but it didn’t hurt anyone.

But we also know that we had 710 deaths reported to the agency, and that some hospitalization were required. We also have unknown. The information was unknown to the manufacturer or to the reporter as to what adverse event happened to the patient.

We also know that there were no patient involvement in some of the adverse events. We know that this also can include out-of-the-box failures. That there was no patient physically involved.

And another top reported patient problem was hypoglycemia.

Next slide please.

So what we wanted to do was just take a closer look at the death report. So we looked at three-year data; 2006 to 2009 and these adverse event reports were actually - all the additional information, all the manufacturer responses, everything was looked at in order to come up with this analysis.

And again we provided a top picture. So when we looked at the large volume infusion pumps, we had 99 deaths for a three-year period. The three main side effects or the problems were identified as over-delivery of medication, the device displayed an incorrect display and also adverse events related to human factors.

Next slide.

And then we looked at some of the device problem codes, and we determined when - excuse me, for this particular area, I pulled out narratives directly from the adverse event reports and the narratives read: Battery failure calls, interruption and delayed of therapy.

If you’re infusing a critical care drug that’s keeping a patient alive and the pump fails, and that medication was used to in order to keep their blood pressure up, taking that away and having to find another pump when it fails, patient can die when that medication is stopped abruptly.

We also know that software error code displayed caused delay in therapy and there was a potential for programming error caused the over infusion. The nurse may have been confused when programming the two different channels.

The device passed the user facility testing, so we know that this was use error.

When we looked at insulin pumps, we had 310 deaths. We had issues with alarms, failure of battery, motor, priming, infusing, display and also blank screen.

We also know there were incorrect software programming calculations. So this was at the code level. Someone is sitting - I want to make sure that slides are there. So we’re on device problems. Back up please. One more please. One more. Wow, you really had a heavy finger there. Wrong way. Am I out of time? There I am. Okay, there we go.

So we had insulin infusion pumps, 310 deaths. We had issues with alarms battery, motor, priming, infusing, display errors and blank screens. We also had incorrect software programming calculations.

So when someone is writing code at the code level, when they said two plus two is four, they accidentally put in two plus two is three or two plus two is five.

And so when the clinician was programming, thinking that they were receiving five units of medication or whatever, they were actually receiving five, 15, 30, depending on whatever the programmer put in as the equals.

We also had self-activation or self-keen and that was when the pump was accidentally bumping, you could actually reprogram it by just slightly touching the surface.

Next slide, please.

So when we looked at some of the adverse events by descriptions in the narratives, the device experience, intermittent failure or completely ceased to operate when changing the program settings with the up and down button.

There was also unintentional rebooting of the pump.

Next slide.

Okay. When we looked at the death reports for the two thousand - for the three-year data, we received 26 reports. Top problems were unknown, incorrect software programming calculations - speak louder?

Okay, sorry. Can you turn my mike up a little bit? I want everyone to hear me. Is that better? All right. Should have told me sooner.

So for the 26 reports, we had unknown issues, incorrect software programming calculations and I just explained that at the code level. We had human factors and use errors. We had excess flow or over-infusion and we had difficult to program or to calculate pumps.

Next slide, please.

Some of the reported problems in the PCA narratives is that they reported over-infusion, was related to programming error, clinician made a change in the drug concentration in the middle of an infusion.

The device was found to be within specs.

Next slide, please. When we looked at elastomeric infusion pumps device problems there were four death reports. We know that there were issues with human factors and use errors. We’ve had reports of unknown, bacterial contamination of the device and failure to infuse.

Now interesting, when we looked at the elastomeric, a lot of these can be - next slide please.

We know that they can be look-alike pumps. Now we know when it comes to drugs we have look-alike, sound-alike drugs. Elastomeric pumps, we can have look-alike pumps.

So this particular death was related to the facility intended for the pump to infuse at the 150 ml per hour rate, but they selected the 100 per ml. This patient obviously died because we’re talking about death reports.

Next slide please.

And we have enteral infusion pumps. In that five-year period - excuse me, three year period we have five death reports. They were related to over-infusion, incorrect display and also human factors.

Next slide.

One of the concerns with this death report in the narrative was the far right digit on the liquid crystal display screen was not visible. The pump was set at a rate of 250 instead of the intended 25.

Next slide, please.

To wrap up, who has to report? Manufacturers, user facilities and we really do appreciate the voluntary reports from the public.

We know that we have an absence of the denominator data. We’ve estimated - one document that I read that’s out in the public estimated roughly 2 million infusion pumps on the market.

We know that it’s important to report to the FDA and also to the manufacturers. We also know that good data is very important. It helps us keep safer devices, helps you manufacturers take that post-market data back into your second generation pumps.

So preview of a coming attractions, I’m going to talk a little bit later about the collective voices and then tomorrow, I’m going to talk about MAUDE.

Melissa Eakle: Okay. And I’m going to have to bring this down because your next speaker is height challenged.

Valerie Flournoy has worked in the Office of Compliance at the CDRH for nine years. Since May of 2007, she’s been Chief of the General Hospital Devices Branch and prior to that she worked in the Dental, Ear and Nose and Throat and Ophthalmic devices branch.

She has also worked at the Center for Drugs Evaluation and Research for 23 years as a microbiologist. During her years with CDRH, she performed collaborative microbiological studies on antibiotic drug products that were established for the USP reference standard.

She obtained her Master’s Degree with an emphasis to healthcare administration from the University of Maryland, University College.

Please welcome Miss Valerie Flournoy.

Valerie Flournoy: Thank you Melissa and good morning. As Melissa stated, I am Valerie Flournoy, I’m the Chief for the General Hospital Devices Branch in the Office of Compliance in CDRH.

My presentation focuses on external infusion pumps, the recalls, the root causes of those recalls and adverse events and the CDRH initiative to improve the quality of infusion pumps.

You may ask yourself why is it that the Office of Compliance is concerned with the external infusion pump. Approximately five years ago the Office of Compliance started noticing that there was an increase in recalls that were related to the infusion pump and we became very concerned.

Why the concern? Because it was determined that external infusion pumps was the leader in the number of device recalls, there were 56,000 adverse events, as Mary Brooks just informed you. There were 710 deaths that were associated with these devices.

So we became very concerned and felt that there needed to be something done to improve the quality of infusion pumps and to protect the public health.

Next slide.

So our first thing is you may ask yourself is some of the information that I’m going to discuss today is familiar to the manufacturers that are here but may not be familiar to the clinicians is exactly what does the Office of Compliance consider a recall or FDA.

It is a firm’s voluntary removal or correction of a marketed product that is in violation of the Food and Drug and Cosmetics Act.

Secondly, it’s a voluntary action that’s taken by the manufacturer when it’s determined that the device here, we’re talking about the external infusion pumps, may be misbranded or adulterated.

Next slide.

It is also an inspective message to remove or correct an FDA regulated product from the market and it is an alternative to FDA, an initiated court action for removing violative products from the market, hence the import detention, also.

Next slide.

This slide represents that if you had where a manufacturer has come into your healthcare facilities and actually has removed the infusion pump and taken it to another location to make some type of correction, and here you see listed on this slide, the types of corrections that they may have done, this is called a removal.

However, if these actions take place during a regularly scheduled maintenance, it is not considered a medical device recall.

Next slide.

This slide, as you note, is a correction where it is done on site in the healthcare facilities, meaning the manufacturer may come to your facility and makes these types of corrections. As you note at the last bullet here, this is the only difference between the removal and the correction because it includes patient monitoring.

Next slide.

What a recall is not, it is not a market withdrawal. It is not a stock recovery and it is not a safety alert.

Next slide.

How did the Office of Compliance become aware of the external infusion pump recall situations? As you will see on this slide, this list is not inclusive but I would tell you the five top ways the Office of Compliance has been processing recalls and receiving information.

The first bullet is the firm appropriately will notify the district offices and we do a lot of collaborative work with the districts, especially the recall coordinator that the manufacturers deal with, where you all would submit your correction and remove reports or your 21 CFR 806 report.

Secondly, through FDA inspections, that is when the investigator comes and inspects the firm. The investigator may look at your corrective action and preventative action records. He may look at your complaint handling. He may look at your medical device report.

And through that he would determine that the manufacturers may have, they’ve either upgraded something to correct the malfunctions for the device.

And thirdly, through the adverse event reports, as Mary Brooks has told you, through the MedWatch report and through the MDRs and of course, a lot of the firms, and the majority of them do, they voluntarily recall their devices.

And the fifth way is through the medical recall alert with our foreign country agreements.

Next slide, please.

FDA’s responsibility in processing the recalls for any device, but especially with the external infusion pump, is to provide regulatory oversights of the recall process, to comment on your proposed strategy and to review the communication.

And please note the last three bullets that the district offices are the ones that do - that conduct the audit checks, witness the product destruction and approve your reconditioning plan and will notify the firm of their termination.

Next slide.

FDA also will look at the firm’s recall communication. And this recall communication I can’t stress how important it is because it notifies the users, or anyone who has received, purchased the device about the hazards that relates to the device’s malfunction and this may be through a press release or through detail instructions about the failure and the risk of the medical device.

Next slide.

Your firm - the firm’s recall communication should provide the following as you see bullets on the slide and the reason being is it needs to be identified to the users the product that’s being recalled, hence here the external infusion pump.

You need to explain to the users why the device is being recalled and what the hazard is that’s related to the device. This should also state whether or not that the device should stop being used or that you’re going to stop distribution of the device.

And lastly but least, your instructions under communication should tell the users what it is that they need to do with the device and how to mitigate the failure.

Next slide.

We also look at the firm’s recall strategy. And as you can see by the depth which is just to make sure that all the users have been notified, a public warning that states what exactly is the hazard that’s related to the recall and the effectiveness check in which the district offices make sure that all of the users have been notified regarding this recall.

Next slide.

And it also will - is a method to be used for the level of the effectiveness check that will be conducted.

Next slide.

Here on this slide you will see the recall classification, and you may ask exactly how does FDA determine the numerical designation and that is we assign that number to indicate the relative degree of risk.

And please note that normally that the recalls are Class I, II and III but unlike the classification of the devices, which are assigned by the Office of Device Evaluation it’s the opposite.

Class III devices are PMA approved devices and they have the most risk associated with them.

Next slide please.

So what exactly is a Class I recall? It’s the most serious recall that you could possibly have, and it’s a reason with probability that the use of or exposure to a violative product, hence we’re talking about the external infusions pumps, will cause serious adverse health consequences or death.

Also note in this slide you will see - oh, no. Oh, here we are.

You will also see at the bottom of the slide that it states that we process - the Office of Compliance process 14 Class I recalls for the external infusion pumps in Calendar year 2005 to 2009.

Well you might say well 14 is really not a lot. But just consider this, that if you or your loved one or even your children were on an infusion pump and it now malfunctioned and it either over-infused or under-infused, hence is the reason for the serious and the risk of a Class I recall.

And an example of this is that failure is a result in over-infusion or under-infusion of the same medications and whereas normally you would have saline in a pump, but it also could happen with vasopressins and so that’s what happens with the Class I recalls.

And some recalls when we receive them have multiple defects that are going on, so for instance, not only is the pump over-infused or under-infusing, there may be an issue with the door not closing properly.

Next slide.

On this slide is a Class II recall classification which is the - a violative product may cause temporary or medically reverse or adverse health consequences or where the probability of serious advents health consequence is remote.

And please note that on this slide we process 70 Class II recalls in the calendar year 2005 to 2009 for the external infusion pumps.

And normally the recalling firm notifies your customers and sometime asks them to notify their intended recipients regarding the recall. A press release would be issued if there was a specific way to do so, for example, if there were a health defect that could effect a large population.

An example of a Class II recall that was processed is a software code caused a logic error that can report old, inaccurate information resulting in the user being confused and the potential for over or under delivery.

Next slide.

A Class III recall is where the violative product is not likely to cause adverse health consequences. And normally in a Class III recall, the firm normally does - just notifies their customers, but a press release is not issued.

And an example of this type of recall is an instructions for use where an error in the educational booklet that was included with the insulin pump related to performance insulin sensitivity testing failed so that - just think that if you were a person that was on insulin and you were using the pump and this happened to you, just consider how critical that could be for your health.

Next slide.

During - some time during - when we’re processing the recalls, we conduct a health hazard evaluation. And this is a risk assessment that provides the Office of Compliance in classifying the recall and determining what action is required by the firm and what FDA would need to do to protect the public health.

But please note that during the conducting of the HHE that we assume that no corrective action has been taken on the device. Also we do have a committee doing the HHE of experts across our center and if need be, we do request other participants across centers from the Center of Drug, Biologics and Food.

And also in the HHH- excuse me, HHE process, we ask that you do not come to the conclusion that it’s use error only but that you will have conducted all analysis to rule out that there could possibly be any other issue with your pump or your devices.

Next slide.

Normally when we do the HHE is because we have a likely Class I or high Class II recall or we don’t have an appropriate precedence that’s available in the classification of the recall or that there’s unique scientific clinical or public health issues that are involved.

Next slide.

And as I said that during the HHE process that we’re defining what the risk is and as you see, the risk is the technical assessment of the defect that is the likelihood of that device failing. It is the clinical understanding of a product defect in a real life situation.

But please note, also, that the lack of a reported injury does not mean that your - that the HHE is a lower risk or that it’s going to be classified lower, meaning a Class II or a III recall.

Next slide, please.

On this slide you will see this pie chart and please note that the recalls that were processed that the most defects that we saw during the processing of recalls related to the external infusion pumps were related to the design controls and process control which were both 41%.

And when you see non-conformance in the top right, your left - no your right corner -- sorry -- that this is a failure in a purchasing patrol or acceptance testing or supply audit and PMA which is 6% is related to the pre-market approval application mainly Class III devices.

From this pie chart, it was determined that there was a need for a better designed pump and better manufacturing process.

So we recommend to all manufacturers that you should conduct trend analysis on the failures of your devices.

Next slide.

And the next following - next slide again, please.

You will see I will provide you with examples of some of the reasons that we - that the external infusion pumps had malfunctioned and here I have a software example where it was related to a key bounce and this defective software interprets the sale key stroke as two separate entities resulting in over-infusion. And this is where a nurse enters the infusion rate of 20 milliliters per hour but the device is programmed to 200 milliliters per hour.

Next slide.

Next we have a human factors example, where the user interface key placement, and this is where the power button is placed next to the infusion start button, resulting in inadvertent pump shutdown.

Just think that if you were in OR and undergoing a procedure and the interface key was punched and the pump stopped functioning, therefore a serious health consequence has just happened, could possibly happen to you on the table.

Next slide please.

The next example is of a hardware design flaw and this is the for the pump use a common Intel microprocessor and Intel’s data should specify the use of two different clock circuits depending on the frequency.

The pump designer didn’t follow explicit instructions in the data sheets and used wrong clock circuits causing the pump’s shutdown at random intervals.

Next slide.

Hence you have seen that I informed you of a story of why the Office of Compliance or CDRH felt that there was a need to do something about the infusion pumps to have a better quality pump on the market.

We’re asking that all clinicians, whether you think it’s a minor thing that if the pump or your device that you’re using, you keep punching the key, that you report that. It’s very important.

And to the manufacturers, we ask that if you have any questions regarding recall - the recall process, that you contact either your recall coordinators and you surely can call the Office of Compliance for assistance and the processing of recalls or if you have a question regarding the process.

And so before I sit down, I’d like to say that the Center developed this new Infusion Pump Improvement Initiative guidance document in an attempt to prevent the error discussed in my presentation and to eliminate failures that we have seen within the last several years.

Thank you.

Melissa Eakle: All right. We’re going to be - now I’m going to be introducing Mr. Al Taylor. Al joined the Food and Drug Administration in 1990 and he oversees a group of electrical software and system engineers who provide technical consultations in regulatory matters involving medical device design and production.

He divides his attention between engineering, analysis and laboratory investigation of specific medical design issues and broader topics such as quality and risk management.

He also oversees the development of custom instrumentation for the laboratory scientists at the FDA. He’s a principal author, along with Bill Midgette of the off-cited FDA publication “Design Control Guidance for Medical Device Manufacturers”.

Prior to joining us, Mr. Taylor gained extensive experience in the design of mission critical electronic systems including avionics and space-borne systems, nuclear power plant control systems and air traffic control telecommunications systems.

He has a BS in Electrical Engineering from Carnegie-Mellon. Please welcome Mr. Al Taylor.

Al Taylor: Thank you Melissa and good morning to everyone. Get that up a little bit.

So I’m here today representing my colleagues in the medical device labs at FDA where we engage in an active research program aimed at measuring the performance of medical devices, developing test methods to compare medical device performance to a predicate or to a gold standard.

And another significant focus of our research is understanding the mechanism of operation of new technologies.

One of the things that we really put a lot of emphasis on is developing tools that can be used by manufacturers to measure the performance of devices and improve that performance.

Then we also spend a lot of time consulting on design issues involving medical devices, and in recent years a disproportionate amount of that time has been spent dealing with infusion pump issues.

Next slide please.

So about three weeks ago now, FDA announced our infusion pump initiative and there was a flurry of media activity and followed immediately by a flurry of response from different stakeholders on the web.

And in rummaging around, I heard, saw a lot of comments from users saying thank you FDA and a few comments saying it’s about time.

But I also saw quite a few comments along the lines of this one, and this is what we at FDA call the denominator problem. We have 56,000 adverse events reports, but there are millions of these devices out and they’re used to perform tens of millions of infusions.

So when you look at it that way, you know, it’s -- next slide please -- you know, it’s fair to ask what’s the big deal here, the actual incidents rate of adverse events is very low for these devices.

So you know, that is a fair question and not all the commenters were nearly so polite in expressing that sentiment but, you know, could we put this in perspective, so that’s actually what we’re trying to do here today is gather a variety of perspectives on this issue.

And what I want to do in the next few minutes is five you an engineering perspective.

Next slide, please

So how serious is the problem? Again, from my perspective as an engineer in a lab, when we ask manufacturers what cause these events, we were often told that the root cause was user error.

They said the user didn’t clean the pump properly, the user must have pressed the wrong button, the user was supposed to check for damage before using the pump, user didn’t verify that the infusion rate was programmed correctly prior to starting the infusion, the user didn’t charge the battery fully before use.

And in each of these cases when we did our independent root cause investigation of the causes of these failures, we found that poor product design, poor engineering caused or contributed to the adverse event.

And moreover, most of these design deficiencies were foreseeable and they were preventable. From our perspective, even one preventable injury is one too many.

So, you know, if you call this the denominator problem, at that point the denominator goes to zero.

Okay, next slide, please.

So, you know, just to reiterate, we found that many of these infusion pump problems are due to poor engineering. We’re hearing from biomedical engineers and hospitals who are tearing their hair out trying to keep the things running.

We hear from nurses and doctors that they’re spending too much time dealing with sick pumps instead of sick patients. And we hear from diabetics who use insulin pumps on a daily basis and they’re expressing similar frustration.

And yes, we’ve also heard from people who are really happy with the particular pump they’re using and wouldn’t want another one, so we’re not trying to say that all pumps are bad. Some of them are.

So in the next few minutes, I want to drill a little bit deeper into the technology and give you just a few examples of what we’re talking about.

Next slide, please.

So before I do the first example, I just want to - since most of you are going to recognize the products or the circumstances that I’m talking about, we’re not going to pretend that you don’t, but it’s not really our intent to focus on the manufacturers or the specific product involved. Instead, we’d like to you to focus on the nature of the design errors and the consequences for you as the patients.

In all these cases, and quite a few others that we don’t have time to show you today, corrective actions have minimized the problem. So the problems I’m going to show you have largely been fixed.

Next slide, please.

So we’re up to Slide 7 now and this is a case that Valerie just alluded to a couple of minutes ago and I want to show you a little bit of the detail. So this is a large infusion pump, large volume pump, that was reported to be shutting down at random intervals without warning.

As Valerie said, an interruption in therapy can be -- and Mary said it, too -- can be life-threatening, depending on the medications being delivered.

The manufacturers investigation concluded that a defective component caused these malfunctions and when FDA did an independent investigation, we determined that the component wasn’t defective, it was working completely within the specifications but that particular component was an inappropriate choice for use in that circuit.

So let’s look at it and see. Next slide, please.

So, on the right you see a pump that’s probably very familiar to a lot of folks. It’s been used in healthcare settings since 1997. It’s got a great set of design features from a user’s point of view. This is, you know, a really nice pump to use.

On the left you see a close-up of the microprocessor chip. It’s an Intel 80C186 microprocessor which is actually a direct descendent of the microprocessor used in the original IBM PC.

Now those of you who are old enough to remember that back to 1981, the original IBM PC ran at 4.77 megahertz and we’re now up to 3.2 gigahertz, which is about 700 times faster. And back in 1981 when this game started every couple of years Intel would come out with a new version of the chip that was a little bit faster than the one before.

And the manufacturers wanted to take advantage of that additional speed so they can make their devices do more things. And so they were scrambling to use the latest chips as the technology evolved. So by the mid-1990s when this pump was actually developed, the 80C-186 was up from the original 4.77 megahertz up to 40 megahertz.

And at that time 40 megahertz was considered blindingly fast. Okay. Now right below the picture of the microprocessor you see a photo of a quartz crystal. This is a 40-cent part that sets the speed of the microprocessor. You can specify a crystal to operate at any frequency or speed that you want and the trick is that there are lots of different kinds of crystals.

So you not only have to specify the frequency of the crystal but also some of the other parameters. Now you recall that the issue was that the pump was shutting down without warning. And this was one of those pesky intermittent failures that everybody hates because by the time you remove this pump from service, get a replacement and take the thing down to the lab and set it up on the bench, you plug it in and it works.

There is no indication of what went wrong. So this problem, which as you’ll see could easily have been detected at the design stage, once you put it on the market it becomes really hard to figure out what’s going on. And most people in that circumstance when they see that the pump fails in use and then it works when you try to recreate the problem on the bench.

You immediately say well, maybe it was a software problem. But in this case it wasn’t a software problem. So it took several years. The evidence that we have is that by 2003 the manufacturer had figured out that the crystal was involved somehow. But even then it took two years to understand what the real root cause was and it was only after FDA got involved that the root cause was identified.

So next slide please. So when the issue came to our attention and that was as a result of a routine FDA inspection of the firm where it was noted that this was an unresolved problem, the first thing that my engineer did was to pull the data sheet for the microprocessor.

Now every integrated circuit has a data sheet that lists the performance limits and the precautions for using that chip. And most data sheets have a sample circuit that you can copy that if you hook it up this way it’s guaranteed to work. And every engineer is trained to read these data sheets and deviate from it at your peril.

This is just basic. So in this case the microprocessor is so complex that the data sheet was actually a book, that book right there. And Chapter 5 of that book was devoted to the topic of selecting the crystal, not just a few paragraphs or words but a chapter. And the figure you see here is an excerpt from that chapter and the words underneath it are a verbatim quote.

And what they basically say is that if you were running this microprocessor at say 20 megahertz, which was the speed that the older version ran at, you would use the circuit on the left, Circuit A, with one specific type of crystal. If you were running at greater than 32 megahertz and in this case remember the manufacturer wanted the processor to run at 40, the book said very explicitly you have to use the circuit on the right.

And you have to use this special kind of higher speed crystal. Well, the manufacturer, the person who developed this design somehow or another missed all of that and used the circuit on the left, which the data sheet said explicitly would not work reliably above 32 megahertz.

So it’s really to us a pretty basic error not to do what the data sheet said. And to put this in some context, when this chip came out in the mid-90s, this higher speed changing the circuit to accommodate the higher speed was a really big deal. And Intel had articles about this in all the engineering trade magazines talking about the differences between this chip and the predecessor.

It was pin for pin compatible with the predecessor except for these few minor details that you had to pay attention to. And somehow or another all of that got overlooked. And it wasn’t caught during verification testing even though the Intel book explicitly listed some things that you should do to check and make sure that your crystal circuit was working correctly. That apparently was not done.

And then it shouldn’t have taken seven years to find the root cause and it shouldn’t have required an FDA engineer to do it. So on the positive side we understood why the failures were occurring. It was obvious what needed to be done. The circuit needed to be consistent with the data book.

So unlike a lot of problems, once we understood what the problem was the fix was easy. Just change the circuit to be like the data book. And it was easy but expensive because there were 200,000 units that needed to be changed. So that was what was done.

So in this case, an easy solution. In many cases, not so easy solution. Next slide. So this is a recap. I think we have pretty much covered this. Next slide. So moving on now, the second case I want to cover is another large volume pump but from a different manufacturer and in this case we heard about the problem not from an FDA investigation or from the manufacturer but from frequent user reports of door breakage.

So again, when the door breaks for this pump design the door places pressure on the infusion set that keeps that infusion set in contact with the pumping mechanism. So if the door is loose you may have inaccurate flow rate and in the worst case you could have free flow because of a loss of pressure on the tubing.

So again, in this case the manufacturer attributed the problem to rough handling by users. And one biomedical engineer from a hospital that we talked to about this told me that we shouldn’t have to handle these pumps with kid gloves. They have to work in the environment and the clinical care environment is stressful for people and it’s stressful for devices.

So when we actually visited a hospital to see what was going on for ourselves and within ten minutes we found another possible explanation. So next slide. So here is a nice family photograph of the pump in question. And again, this pump has a lot of really cool features. In this case the unit in the sensor is the microprocessor, the brains of the unit.

And you can put up two modules on either side and one of the features is that the pump is designed so that if you need to add an infusion when you already have some infusions in process, you can plug in an additional module while the pump is running. Most medical devices, you’re supposed to turn off the power when you connect or disconnect something.

This one you’re allowed to connect or disconnect the module with the pump running. So and the pump itself, the family has a bunch of different kinds of modules. There is also a syringe pump module that we don’t show here and some other monitoring modules. But this is a common configuration at the patient’s bedside with three large volume pump modules and one post dosimeter module. Next slide please.

So here is this photograph of a cracked hinge. And again, the damage doesn’t look so bad in this picture but the consequences can be pretty devastating. Next slide. So when we looked at this it only took a few minutes actually to see what you see here is the tubing set has this blue piece of hard plastic on it.

The mechanical engineers call that a feral and it fits into a slot at the top of the pump that is supposed to keep the tubing properly positioned through the pumping mechanism and presumably it also if you drop the bag or pull on the tubing, that feral locks the tubing in place so that you can’t accidentally yank the tubing out of place while the machine is running.

So what you also see is a little bump on the inside of the door that is designed to fit between the two flanges on that blue plastic piece and hold it in place when the door closes. You can see that, right? So next slide please. What we noticed very quickly was that blue plastic piece is slightly out of position when you close the door.

Now the pump catches on the bottom flange and that keeps the door from closing. If you push hard on the door it turns out that the hinge is the weak link. Something has to give and the hinge breaks. So there may be other root causes for this problem of door breakage but this sure looked to us like one of them. Next slide please.

So as with the previous example there was a clear design deficiency that wasn’t identified in any stage of development. It was overlooked even after the reports of broken doors started piling up. And once FDA applied some pressure the manufacturer devised a corrective action that reduced the incidence of this problem.

However, in this case unlike the previous case where there was an easy fix, once this thing is on the market it’s really not so easy. There are some things that with 20/20 hindsight you’d like to be able to do over again. But it’s very hard to go back and start from scratch when you already have a product on the market.

And that’s really why all the engineering textbooks say you have to find these problems before the product gets on the market when you have the luxury of time and it’s not so expensive to fix the problem. Next slide please. So a third case and I hope you don’t think I’m picking on the same manufacturer but I picked these cases not to say anything about one manufacturer or one model pump versus another but just because they were good examples of the kinds of issues we’re dealing with.

And they have the advantage of not being too technical so in this case we had frequent user complaints of sparks and burnt connectors in the same large volume pump. And again, this led to an interruption of therapy but when you have sparks and smoke coming out of a machine, that raises the concern of starting a bigger fire.

In this case, probably very, very unlikely but it’s still something to be concerned about. In this case the manufacturer identified the root cause of these failures as improper cleaning of the pumps. And again, the FDA labs investigated and found another explanation that although we had great difficulty proving it conclusively certainly the circumstantial evidence leads you in a different direction.

So let’s look at that - next slide please. So I said something there on that previous slide about one of the issues for the users was churning of inventory. And when we visited a hospital they took us to a room where I’m sorry about the fuzzy photograph but the photographer is a klutz. It was a brand new digital camera and I hadn’t figured out how to use it.

But regardless of that you can see a cartful of pump modules that are defective and awaiting return to the manufacturer. And what you can’t see in this photo is every one of those boxes is filled with the same modules that have already been packed. So as you might suspect, this is a substantial fraction of the hospital’s inventory.

And they told us that this was one month’s worth of failures and these were the failures that they couldn’t fix in house and had to send back. There were a lot more failures that they were fixing in house and weren’t going through this process. So we are talking about some impact to user facilities.

Okay. So next slide please. So here is the review of the controller module on the left and on the right you see one of these pump modules. And you see the black connectors, the pump module’s designed actually to hook on to the controller module and then there is a clip at the bottom of the pump module that fastens to the bottom of the controller module and holds it in place.

But that connector as well as providing power to the pump module and control signals back and forth, that connector is also supporting the weight of the up to two outboard modules. Next slide please. So this is just a close up of the connector so that you can kind of see how they fit together. And the connector on the left you see has an upward facing hook.

And the mating connector fits over it and kind of pivots down into place. That’s how you put these things together. And again, that’s done with the power on. So what is the problem here? Next slide please. So here is a photograph of a burnt connector.

And you’ll see that the second pin on the left kind of appears to be missing in action. It’s actually there but it’s obscured by melted plastic and just discolored. So it’s there but you can’t see it. But it’s pretty clear that there was a lot of heat in that spot.

And the adjacent pins you can see that they are heat damaged but the appearance suggests that the heat is really coming from that one source. And that pin that is the ground zero for this incident happens to be the pin that delivers power to the outboard module. So immediately it leads to the question and also we found out that the ground connection for that power supply is on the adjacent pin and in some circumstances on the adjacent pin on the other side as well.

So now we have an issue. It’s kind of a cardinal rule in electronics that you don’t put power and ground pins next to each other on a connector if you can help it because it’s too easy to short out the power supply, which usually has bad consequences. So the immediate suspicion here is that we had a power supply shorted and that’s what caused all of this heat, sparks and damage.

Now can we go back a slide for a second? Never mind. Let’s go forward again. The reporter for this incident said that fire came out of the unit. As an engineer I think we can say categorically it wasn’t fire in the sense of open flame. From a lay point of view where there is smoke there is fire.

But that’s not always true. So what we suspect is we had sparks coming to of the unit, perhaps little bit of molten connector and maybe with quite a bit of force but it really wasn’t a fire. But it was definitely shutting down the pump and in some cases that can mean shutting down as many as three infusions in progress and needing to reprogram all of those channels once you get another machine.

A few users reported receiving minor burns because of course their hands were on the pump when this happened. Now again, for months the manufacturer had been telling us that the root cause was improper cleaning and you can see some green corrosion over on the right of the picture on one of the pins.

So that was evidence of is the cleaning procedure getting too much fluid and the theory was that the cleaning fluid was getting inside the connector and causing short circuits. But the biomedical engineer told me that they after they started having this problem they retrained everybody on the cleaning procedure.

He showed me the procedure they were using and there was really no question in anyone’s mind by the time that we arrived on the scene that cleaning wasn’t the issue because they were being very, very careful with their cleaning and they were still having this problem. Next slide.

So what’s really going on? Well, we took some of these connectors back to our lab and looked at them under the microscope. And the top pictures really just show you the heat damage on one of these connectors close up. But the bottom left photo shows that one of the connector housings is cracked.

Now if that connector housing is cracked that means that the mating connector when it goes in may go in slightly at an angle. And if it’s at enough of an angle you can imagine that the connector pin might land in the adjacent slot instead of the one where it’s supposed to.

And in the photo on the right, I don’t know if you can see it from a distance but there is definitely on each of the barriers between the pins there is a wear pattern that shows very conclusively that the mating connector was consistently going in at an angle and scraping on one side of the barrier as it was going in.

And I guess if you’re sitting in the back you’ll have to take the word of the folks that are sitting up front. Next slide please. So I’m sorry - there are actually a number of things going on here that are quite interesting and I want to make sure that I cover all of them. So we speculated again that the forces of mating the pump module to the main module could put quite a bit of sideways stress on this connector.

If it wasn’t perfectly lined up when you tried to mate the connectors, that was probably what was causing these connector housings to break. And again, our speculation was that the pins instead of going in in alignment, they were going in off by one and causing the short circuit. So we looked at the power supply circuitry and we discovered that the power supply had no means to limit the current if there was a short circuit.

There was no fuse that would blow in this circumstance so the full power of the power supply was available and we calculated that the power supply could actually supply several hundred watts of power into a short circuit that was created in this way. Again, we didn’t actually - we were looking at connectors, not complete assemblies.

So we didn’t actually recreate the failure mode in the lab but all of the evidence that we had is showing this as the probable cause of these things. And ideally if you are delivering power to both sides of the module in addition to blowing the fuse you would want to isolate it so a short circuit on one side of the pump would not interrupt infusions in process on the other side.

But that wasn’t there either. So next slide. All of you have computers, laptops. You have seen laptops that fit into docking stations, you have seen iPods that fit into docking stations. This problem of docking connectors has been around as long as there have been electronics.

Back in the ‘70s I learned a few design principles that when you have docking connectors you provide first features that provide for mechanical alignment. If you look at the bottom of your laptop you’ll see metal pins on the docking station that fit into sockets on the bottom of the computer that make sure that the assemblies are lined up before the connectors even mate with one another.

If you look at your iPod, the iPod fits into a slot, which is designed to line the iPod up before the connector mates. And if you also look even more closely, you’ll see that the docking connector in the iPod dock, it can wiggle a little bit so that if there is any remaining misalignment in the assemblies when those connectors come together, the movable connector in the dock can actually adjust itself to fit together with its mate.

So the whole design is done with the intent of minimizing the stress on those little tiny connector pins to ensure they don’t get bent and they do go to where they need to go. And these are just fundamental mechanical engineering principles. And again, this product really had none of those features.

So I’m sorry we’re still here. So this is just a list of kind of what we found. And again, we haven’t really proved a root cause but the one that the manufacturer identified is just not consistent with the evidence from users and all of these things that we found are clear deficiencies in the design.

The shortest thing that if you had to do over again you’d probably do different. So next slide. So this is kind of the world that we’re working with. You see design deficiencies. They’re not identified there in design reviews. They’re not caught during verification and validation testing.

And even after we have failures we’re not seeing these be identified. So again, the bottom line is that when FDA got involved the manufacturer never really agreed with our tentative conclusion about the root cause. But they did make some significant changes to the connectors that would minimize the incidence of this problem. So this is another problem that has been fixed. Next slide.

So I just have to say that what I’m showing you here is the tip of the iceberg. Many of the design issues we have seen with infusion pumps are much more complex, equally of concern and I do want to emphasize that there are a lot of medical devices out there including some infusion pumps that are well designed.

We don’t in my 17 years with FDA I have only - I can count with the fingers of one hand the number of cases where a manufacturer really didn’t care about quality and was set out to make a profit at any expense. I probably have three fingers left over.

In almost all the cases that I’m talking about everyone set out to do the right thing and somewhere along the line the engineering fell short. So you can see now I hope why we think that we need to do pre-clearance inspections to make sure the manufacturer is doing the kinds of verification and validation they need.

I think you can see why we’re going to be looking more closely at the engineering design in the pre-market review phase and why we’ll be expecting engineers to do clinical evaluations of these products and try them out in the real world while there is still time enough to fix the problem before they’re actually on the market.

So that’s what we’re going to be talking about for the rest of today. We’re going to be hearing other perspectives and thank you for listening to mine.

Melissa Eakle: All ready. Back again. Before we take our break I’m going to reintroduce Mary Brooks. I’m not going to go through the whole bio because you’re probably waiting for your break. I do want to remind people as you leave there will be someone at the door with cards. You can write down your questions if you want to submit them that way.

For people on the line please do not hang up. Please remain connected during our breaks. We’ll be back in approximately 15 minutes. Thank you and here is Lieutenant Commander Mary Brooks.

Mary Brooks: I’m coming. All right. Well, I’m back again. I will speak louder this time I promise and please, I know everyone is tired. It’s hard being the last speaker before a break so I just want to talk to you about collected voices.

Dr. (Shuran) has been a fabulous center director and allowing us to have a workshop is fabulous. But also having the infusion pump working group, working with my fellow colleagues from OC, Office of Compliance, Office of Engineering Laboratories and ODE sitting together and trying to figure out what we can do.

We kind of get on our soapboxes at work. It’s like what can we do? So this is just a collection of our voices and also a collection of voices that we have been hearing. So collected voices. Next slide please. Okay. So we did some outreach obviously. We wrote a guidance document and I think a lot of people here today are here to hear about that guidance document.

But we also provided a white paper. This information is available on the FDA CDRH Web site for infusion pump initiatives. We wrote a letter to manufacturers, we developed the Web site and we have been working with our international regulators from around the world.

We have people here from Health Canada, from Sweden, Germany and the United Kingdom because we know that infusion pumps cross the pond and they go up to Canada and they are around the world. And the problems that we have here in the United States are also problems around the world.

A lot of the same manufacturers are overseas. So they are here today to learn more about infusion pumps and our initiative and we are helping them share the information. We have also reached out to stakeholders with international calls and also national calls.

And we have also worked with our clinical community. We have worked a lot with the MedSun Hospital System. For those who are not aware of the MedSun Hospital System, it is a little bit over 350 hospitals that report directly to the FDA and we can get real time data. We can ask them a survey in the morning, we can get answers by the afternoon.

So if we have concerns about a particular device operating in a particular way we can reach out to nine different hospitals in the morning and get answers in the afternoon. So we have a very strong network with the MedSun Hospital and I thank Susan Gardener for making sure that initiative continues.

And we have also made sure that we have had focus groups. We have talked to clinicians on the ground who are using these devices. You’ll hear some of their comments a little bit later today also with some of the MedSun Hospitals. Next slide please. For our focus group we talked to infusion nurses in the northern DC, Virginia and Maryland area.

And we roughly had a little bit less than a dozen nurses for about two hours and we just asked them some general questions. And the take away that we got from them was that they wanted labeling to be readily available. They didn’t care exactly how and they didn’t care if it came from the manufacturer, they didn’t care if it came from the FDA. They just want it to be available. They wanted it to be available in one location and on a Web site if at all possible with availability to be downloaded.

They really wanted a standard user interface and they were a little confused that the FDA can’t mandate that and it was difficult to explain that we do not regulate how you design your devices and what you think is a great device. That is up to you the manufacturers. So if they want a standard user interface, manufacturers have to work together to get that.

They also wanted regulators and manufacturers to reach out to them directly. And we are doing that by having this workshop and making sure that we outreach to the clinicians, that we have provided a tremendous amount of education material on the infusion pump Web site and we’ll continue to do so with other devices that we deem appropriate and necessary. Next slide please.

So one of our stakeholder calls that we heard a message from, there was this fabulous gentleman. He was a farmer I predict in the Midwest and he was a long-term TPN patient. So he was receiving total parental nutrition and had done so for decades and he was very kind on the stakeholder call. And I know that some of you were on that call.

And he expressed wanting a pump like a ’57 Chevy. And he was a fabulous man. He wanted a ’57 Chevy pump. He wanted a pump that worked. He didn’t want a lot. He wanted it to work. He wanted it to be simple and easy to use and durable because he received his TPN therapy while he’s farming on his tractor.

So not everyone receives their TPN therapy at night. A lot of people are receiving their TPN therapy while they’re working during the day and they’re using backpacks to carry their devices around and they are getting off and on buses and in and out of cars and they really need these devices to be durable and this was his voice. Next slide please.

So I also found a great analogy on the Web site, at least I think it was a great analogy and I wanted to share it with you, was a 2.7 million mile car. And you ask yourself how can someone get a car, it was a 1966 Volvo P1800 if anyone has one. This particular gentleman, the article that I read, gave out some really great recommendations.

And I thought this applies to any device, any product that’s out there in commercial use. He said you have to read the owner’s manual. You have to read the owner’s manual. He said don’t cut corners on routine maintenance. He also said be diligent about your maintenance and that it can make a profound difference.

He said if you hear a tinker in the motor don’t turn the radio up, don’t ignore it. Yes. I see people smiling and laughing. I’m guilty. He also said make sure you choose your mechanic carefully. He said make sure that you give the mechanic the complete details about the noise. So how does this work with infusion pumps? What is the message to manufacturers and what is the message to clinicians?

For manufacturers, you need to - let me go back. Yes. We’re back on the car. We’re still on the car. I’m on my soapbox. So manufacturers, obviously you need to be writing a really good manual. Clinicians, you need to read the manual. You need to make sure that if the manufacturer provided labeling, provided a great CD for training that you actually use it.

Don’t cut corners on routine maintenance. Manufacturers are telling you how often you need to PN the devices. The hospitals’ biomechanics and their engineering departments need to make sure that they have got a system in place for routine maintenance and choose the mechanic carefully and make sure that you tell them what is going on with the pump.

So clinicians, you need to make sure that whatever was happening during that infusion that you give them all the information so that they can troubleshoot. I don’t know about you but I have taken my car into the shop because I’ve heard a noise and they drive it around the block two or three times and they give it back to me and they say there is nothing wrong.

I didn’t tell them that it happened when I was going 55 miles an hour up on 495. Maybe they would have taken my car around the block up to 495 going 55 miles an hour. So it’s important to tell the biomeds and the manufacturers what the device was doing, what it sounded like, how it was operating. Don’t just put fire on the side of the pump or broken with a sticky note.

Give more information. Next side please. So the FDA, we came up with some infusion pump risk strategies and we spoke to clinicians, pharmacists, clinical and biomechanical engineers and health information technology professionals, facility administrators and managers, home health nurses and patients using infusion pumps at home.

When we were in one of our work rooms the affinity diagrams that we had up linking all these messages was pretty impressive. We had paper after paper trying to make sure that we were getting the messages to the right people and some of our messages, these are all available on the Internet for download. So if you’re a clinician you can go to the FDA’s Web site and get the messages.

Also for pharmacists, all these messages are on the Internet. I’m just going to quickly talk about a few of the messages. I know that we’re getting close to our break and I do not want to hold you up because we have got more information to share with you.

So I’m just going to quickly talk about some of our risk reduction strategies that we have identified for a few of our people out there. Next slide please. We have asked clinicians - back up a little bit please. I think you skipped one. There we go. We asked clinicians to plan ahead.

These are not complete off the Internet. I had to shrink them down a little bit and take out extra words just to get the overall sense. But the complete messages are available on the Internet. We asked clinicians to have a back up plan in case their infusion pump failed and we have provided the details for that.

Next slide please. We also asked clinicians to label. Think about labeling with their devices. Make sure that you label the pump channels so there are less adverse events that can happen. Next slide. We also ask clinicians to check, to verify the infusion pump is programmed for the right dosage and the right rate and the volume to be infused.

We’ve asked clinicians to get an independent double check of the infusion pump that is by with a second clinician. We’ve asked them to monitor for over infusions and under infusion signs for high risk medications and we also need to make sure that they monitor their patient and the infusion per their facility’s protocol. Next slide.

We also asked clinicians about use to make sure that available resources to the bio engineers are available. We’ve asked them to use the drug libraries when at all possible and to make sure that they are using the slide right. I got the sign that I’m out of time so we’re just going to skip to the very end.

We also spoke with facility administrators and managers. So as we go to our break and I really appreciate everyone listening to the collected voices from the FDA and from around the country, (Jason) is going to start smiling at me now. But when we were talking about what do we name this and I said I think that it should be titled Collected Voices.

And he shared with me a fabulous story that said a friend in college said is it wrong if the voices in your head tell you to do good things? And I’ll leave you with that.

Melissa Eakle: All right. You are free for 15 minutes. The bathrooms are straight through the doors. Please pick up your cards on the way out and I’d like you after you write your questions down to drop them at the registration table.

(Karen Jackler) is going to be receiving the cards so that we get them. When you come back more on the menu. We look forward to you coming back. Those on the line please do not hang up. Would the next group of speakers please come forward to see me? Thank you.

((Crosstalk))

Melissa Eakle: Hello. If you could all start going back to your seats. If you have any question cards please get them to the registration desk now. If the FDA members would also return to their tables it’d be very helpful.

Again, please return to your seat. We’re going to try to get on time and we do have people on the line for the webinar all over. All right. I want to thank you for coming back. I want to - and I’m officially deaf. Okay. All ready. Before we begin I’d like to introduce our moderator Mr. Mark Burnett.

He is at the end of the table right here. He is an Associate Director for Outreach for Review in the Office of Communication, Education and Radiation Program as part of FDA Center for Devices and Radiological Health. He helps design, develop and disseminate the center’s public health messages for healthcare professionals and the public, concentrating especially on benefit risk issues involving medical devices and radiation emitting products.

He serves as moderator in live and televised programs for the FDA staff, the public and is the Executive Director and co-host of a show you may all be very much aware of called FDA Patient Safety News. So if you think he looks familiar, he does.

FDA’s video news show for healthcare professionals can be found at www.fda.gov/psm. We’re going to do our question and answer period right now. Mark - okay then. All ready. Mark will be answering/asking the questions that you all submitted.

But I will remind you that we have a microphone in the middle aisle and if you want to ask a question on air just come up to the mic, state your name and your affiliation, either your company or organization. Please speak clearly into the mic because we do have people, about 300-400, who are listening on the phone.

So if you have a question please do all those things. I now turn it over to Mark Barnett.

Mark Barnett: Okay. Thank you. Can you all hear out of this mic now? Can you hear in the back? Yes? Okay. I noticed when I sat down this morning they put this big box of Kleenex next to me.

I’m assuming that if anyone cries on the panel I’m supposed to pass this down. Is that better?

Melissa Eakle: That's better.

Mark Barnett: Okay. All right. Another thing I noticed looking at these cards is there is definitely the next FDA workshop has got to be on penmanship but I’ll do the best I can here.

Is anybody - this is not the only way to ask a question. If anyone wants to come up to the microphone and ask one live just line up there and that will be fine. So I’ll be watching for you if you do that. Okay. I’ll do these just at random. This one says will FDA create a focus group on the failure of problem investigations and capital programs?

Anybody want to handle that on the panel? Just jump in. Do you want to just pass this?

Melissa Eakle: Okay. So the question again, Mark, was would we...?

Mark Barnett: Would FDA create a focus group on the failure of problem investigations and capital programs?

Melissa Eakle: Well, we will consider it. Show of hands, who thinks that it needs to be done? Show of hands in the room. Who would like the FDA to hold a focus group on this? Okay. I see one person requesting this. Does anyone else think that it’s a good idea? Does everyone understand what a CAPA is, Corrective Action Plan? Okay. I see heads nodding. Okay.

So we will consider it but at the same point we have to make sure that it’s what you want as well.

Mark Barnett: You might as well stay there and I’ll throw another one up and someone else can go to the mic if necessary. This one, the same person says is there an estimation of under reporting in clinical and home use settings?

I think that’s really important. We don’t get information on everything that is going on out there and I guess the questioner wants to know can you in any way quantify how much we’re not getting?

Melissa Eakle: Well, I know that there is a box on the 3500A to indicate that the device was used in the home setting. That is not a mandatory box to check. That was addressed yesterday in the home care workshop but that is not a mandatory item to check off. We do know that there are medical devices used in the home and we do not have a clear idea as to how many of them there are because that’s not a mandatory requirement from the manufacturers who have to report.

Mark Barnett: I think the question was also asking you simply about the degree of under reporting, whether it’s home or clinical. Do we have any idea?

Melissa Eakle: It’s both actually. We know that it’s both. We had estimated that we’re receiving about 10% of all adverse events to the FDA. So we know that there is a tremendous amount of under reporting.

When we cross compare our voluntary reports to the user facility reports and then also to the manufacturing reports there’s a huge discrepancy of rates. So we do know that there is a tremendous amount of under reporting. Our estimates are we’re receiving about 10% of adverse event reports.

Mark Barnett: This questioner is asking about when the slides will be available and I’m told that’s Tuesday. Is that right? It’ll be up online?

Melissa Eakle: The slides will be available Tuesday.

Mark Barnett: On the FDA Web site?

Melissa Eakle: On the FDA Web site.

Mark Barnett: Okay. This one says is there I think this says concurrence at FDA for a broad brush approach for regulating pumps? Do all pump types need the same strategy? How do you accommodate the specific needs of different systems/electronics for example, Elastomeric versus mechanical pumps and so on?

Anthony Watson: Yes there is concurrence. I’m Anthony Watson and my title is long but let’s just say I work there and I’m a director. There is concurrence that there needs to be a broad stroke.

The process to go through each individual pump and determine what each individual pump requires is too cumbersome, too long. We believe the guidance document that we have produced explains why we think a broad brush needs to be done. But however, it doesn’t preclude a specific solution for whatever type of pump that you might manufacture.

So yes, we do believe that within FDA there is concurrence. As I mentioned, there was a lot of discussion about how we got to that point. But we did come to the conclusion that we needed a broad sweep because the problems are pervasive even though there might be a small percentage where a particular type of pump’s PCA, we still believe we still see problems with them and those also can be improved. This is about improvement, not perfection.

Mark Barnett: Okay. Thanks. This question says with regard to adverse events and MDRs, it’s sometimes difficult to acquire information for an adequate root cause analysis. What’s the expectation for a manufacturer to identify causes when there is little to no information available?

Mary Brooks: Hi. This is Mary Brooks again with the Office of Surveillance and Biometrics. We feel that the manufacturer should do the best job that they can and document that.

We do realize that there are several, several components into getting additional information. Yesterday one particular manufacturer said there could be as many as 11 different streams of information and the difficulty that it takes to get that information into the 3500A for reportability.

We ask that you do your utmost best in getting the information to the best of your ability and that you have protocols in place on how you will receive that information rather than ad hoc, that you have some sort of process in place, that we absolutely understand that it’s difficult to get all the information to come up with a root cause.

We do acknowledge that and we do appreciate the fact that we have manufacturers who are reporting and doing a fabulous job at it. But there is some room for improvement.

Woman: And Melissa would like to say something.

Melissa Eakle: Wow. You can tell that we feel very strongly about this. Prior to becoming a network leader I was an analyst in the Office of Surveillance and Biometrics.

And what I will tell you is in reading reports very often when I would call or email the manufacturer or the hospital or the voluntary - because we do call all of the voluntary reporters, they would say well, we did this, this and this. And I would say but it’s not in the form. If I had known that I wouldn’t be calling you and taking up your time.

So if a manufacturer has made multiple attempts to try to get the information to us it’s very important that in their reporting they say we have made multiple attempts, the dates, who they tried to reach so that we can see that they actually did make the effort.

The same thing with the user facilities. They have unique problems about getting the information. Especially where a death is concerned it’s very important that you tell us what those are. I can tell you as a nurse practitioner and a former critical care nurse that I often get asked questions about HIPPA. And HIPPA does not apply to us when we are investigating a death because it’s for the public health.

So you can get the information normally from the discharge note or from the death certificate. I have called coroner’s offices for hospitals. I have called police departments if it was a criminal investigation. So we will help you if you find in a death report that you’re having problems. You just need to let us know what information you can’t get you feel is necessary.

And if we can we’ll bring the FDA’s presence to help you get that information because it’s in all of our best interests to make an accurate evaluation of what happened. So thank you.

Al Taylor: Thank you. And be careful. They said this was working but it’s still kind of - that was low. So from an engineering point of view we see many cases where it’s very difficult to figure out what happened.

When you get one report without a lot of evidence associated with it, the people who are using the pump are focused on the patient. They’re not focused on preserving the evidence of what went wrong. And so these things can be very difficult.

As I said in my talk, there are those situations where the opportunity existed to find the problem before the product got on the market. But after the product is in the field it is difficult. On the other hand there are quite a few cases that we have looked at where the consequences were serious enough that we were in a leave no stone unturned situation.

And when we looked at what the manufacturer did there was a lot of low hanging fruit there that the manufacturer didn’t follow up on. So that’s of great concern to us when that happens. And last but not least, we’re looking at the idea of using what we call flight data recording technology to capture more information about what happened.

And that’s something that we’re going to be exploring with industry in the very near future.

Mark Barnett: Al, why don’t you stay there because there is another question here that I think is addressed to you. It says do you think the FDA will be able to predict the failures - would have I guess it should say would have been able to predict the failures of your examples in a review capacity?

Al Taylor: That’s a really good question and the answer is it’s well known that you can’t inspect quality into a product. And it also means that you can’t review quality into a product.

But the kinds of design deficiencies that we are seeing, there is no way the review is going to catch all of them. But under the quality system regulation every manufacturer is required to do a design review with independent, qualified reviewers from all the applicable disciplines before the product is put on the market.

And FDA’s pre-market review ideally mirrors that review that has already been done by the company and I have participated in many, many design reviews both when I worked in industry and now that I work at FDA and we do find problems. It’s really a value-added process. The trick is that you’d like to do that review when there is still time to fix it.

Mark Barnett: Stay there. There’s more. Here’s another one. It says out of three technical examples only the first one can be easily found by inspections. The other two require wear and tear to get to the fault detection.

Can you comment on how you propose detecting volume and wear type defects?

Al Taylor: I would say that in looking at the third example was the burnt connectors that we could look at that design and knowing the principles of docking connectors we could see that that one component was trying to do the job of three components and that that’s putting a lot of eggs in one basket.

And really with no provision for there is no flexibility in that design for misalignment, no provision. So I think we could actually have spotted those problems.

Mark Barnett: Okay. Why don’t you take a break and we’ll maybe come back to you? This one says is there any evidence that increased vigilance since 2005/2006 has made infusion pumps any safer? And if you have those statistics, can you share them?

Anthony Watson: I’ll start out by saying that it’s anecdotal evidence but I will say that some of the products that we’ve seen definitely have gotten better but improved vigilance.

I remember there was a manufacturer who was making a product and they came to us and this was in the days when we had just started asking for certain bits of information about more human factors testing and more software testing or at least more documentation about software testing.

And the manufacturer was a little resistant to give us what we wanted. But we were persistent. They provided us with what we wanted and then we started going round and round with problems they were finding in their testing. And in the end they wrote a letter after it was all cleared. It took a while to get through the process.

After it was all cleared the manufacturer sent a letter to me saying it was a real pain. But thank you because our product is better now than it was before. So we see from - I mean that’s an example where the manufacturer had come back to us and said if you hadn’t - basically saying if we hadn’t pushed them the product might not have been as good when it went out the door.

We also noticed in various products as we review them and we start to ask questions about anomalies and things of that sort, when the company goes back and tests it they find more problems. So there is evidence. It may not be public, it’s anecdotal perhaps. But from our perspective we definitely see some improvement in the products just from what we have done so far over the last four or five years.

This will focus the spotlight on the actual depth of thinking for the design process for the product and perhaps the documentation will help not only FDA see sort of the trends in the products - I mean that’s part of it. Part of our job, we think we’re in a unique position.

And we have heard this from other manufacturers for all kinds of products. We’re in a unique position. We see everything that gets through here. We see different designs. We see the thinking behind them sometimes and we’re in a position to say this is what our collective knowledge brings to the table.

By giving that type of depth of information as I said, it’s going to be difficult to bring it together, it’s going to be difficult to review it but it’s necessary for us to really get a good feel for what are the real root causes of these problems. I think we’ll get to some root causes too.

So we do believe there is value in the process and we do believe that the products that are coming out are even better than they were even from a slight change in the review process.

Mark Barnett: Thank you. There is a follow on question just occurred to me. We have a lot of clinicians in the room here and a lot of them watching remotely. And I think one question they would have is okay, you’ve identified a problem and a situation that needs correcting.

And you have got a plan now to do that. When are we going to see the better pumps that result from this program? Is it going to be next year? Is it going to be in five years? I know that’s a hard question but does anybody want to take a stab at that?

Anthony Watson: A stab. I’ll just tell you from my perspective the folks in the audience know better than us. I think though we’ll start to see better information. Better information will lead to better questions.

Better questions will lead to better products. I think it’s going to be a few years to be honest with you. That better product is not going to get on the market for a few years out the door. What our role in this is to help facilitate that process. I have heard even on the break some great ideas about what people are thinking about to improve the infusion safety period not just for infusion pumps.

But the thought around the whole system, that’s sort of what we’re looking at in this guidance document is the system approach. Looking at the infusion system, how do we improve that? That is a big undertaking. It’s users, it’s the manufacturers, it’s the regulators, it’s people buying the products forcing change.

It’s a big deal. It’s going to take a few years. We realize that. But we understand that getting better information now will make the product better later. I think Mary has something to say too.

Mary Brooks: This is Mary Brooks again with OSB. When we were - that was a question that we actually had internal to the FDA is when do we think that we’ll see better pumps in the market?

And Tony, I really do like your response. But what we have done in the interim on our Web site is we did come up with a risk reduction strategy for the various clinical communities. So if you are a hospital administrator what can you be doing to make sure that you’re reducing the risk? If you’re a clinician, what you can be doing?

So we do have and I mentioned this in the previous slide. We do have outreach to the clinical community to reduce your risk in the interim while better pumps are being manufactured.

Mark Barnett: Thank you. We’re running out of time in this session. Let me ask this is a good question I think from a clinician. Let’s get a fast answer on it. Is there a dedicated representative from the FDA responsible for specific geographical areas that could be available to help hospitals with infusion device recalls and guidance with how to transition to new pumps? Make it a fairly fast (answer).

Anthony Watson: I’ll ask Valerie Flournoy to address that question. And then we have a gentleman standing here.

Mary Brooks: Okay.

Valerie Flournoy: We have the district offices and just I guess you could call the center and wherever you’re located we could actually put you in contact with the district offices because FDA is broken up into districts.

And usually these districts are the ones that actually deal with the manufacturers as it relates to the recalls.

Mark Barnett: Okay. Thank you. We have a live questioner here. So these cards were all fake but this one is real.

(Brad Von Adam): I actually wrote all the cards too. So it’s all me. My name is (Brad Von Adam) with (Ekry) Institute and my question is related to the previous question.

So the previous question was when will we get better devices? My question is how will we know that this process has been successful? How will we know we have better devices?

Anthony Watson: We are right now in the process of trying to figure that out. But we also know that it’s experiential a little bit. You experience these pumps. You experience using them.

The users will let you know if these are better. They’ll let us know if they’re better. But we do know that in order - it’s been proven over and over again if you want to fix a problem you’ve got to focus on it. So our approach is to focus on the problem, get some information on these devices and see where this is going.

I don’t think it’s going to make a worse product. I can’t imagine it will. You could argue about whether it’s too burdensome of an effort to undertake based on the possible rewards. We would argue against that.

(Brad Von Adam): Are you planning to release those criteria when you come up with them?

Anthony Watson: I think we have no problem doing that. Of course I’m speaking totally extemporaneously and I’m blaming Melissa right now just by looking at her. No, I don’t know why we wouldn’t. I honestly don’t know why we wouldn’t.

I mean if you need to know how you’re doing we’ve got to tell you, right? I mean that’s fair. So yes, I think Dr. (Shoran) would probably agree that we need to be as transparent as we can. Probably our Web site. I can’t tell you what form it’ll take but we do have a process in place to look at that. I see this light is flashing saying stop so I’ve got to stop.

Mark Barnett: Okay. As they say on TV that’s the last word. Thank you all very much for your questions. Melissa, anything else before we go any further?

Melissa Eakle: Real quickly just a reflection of the work that the infusion pump working group has done over the last five years, when we first started getting together in 2005 we didn’t know a whole lot about what the other office did.

I knew Al was an engineer of some type. I thought Tony was, wasn’t sure. Valerie and some of the other people, we had to find out what it is that they did and what their perspective was on the problem as we focused on it. So we finally came to consensus fairly early after looking at the recalls, adverse events, the quality of the submissions that Tony’s group was getting, the investigations that Al was doing in his group is that there was a problem and that it was in many cases a preventable and fixable problem.

We are looking at metrics. I’ve got to tell you already we have people calling us. Part of the metric is now people are aware of what we know and you’re becoming more aware of what we know. And we are starting to talk about the issues, which is the start. So we’re really thrilled that that’s occurring and you all came and we had such a high response.

The second thing that we have sort of looked at and if you have ideas for metrics, please send them forward. You can email them to us. We have a Web site, a Web page. But we looked at less recalls, fewer adverse event reports. We realized that if you all start reporting as we hope you will there will be an initial bump.

But we’re hoping that as we go forward there will be less adverse events that you have to report about. So that’s something. And also that the severity of injuries and/or deaths that we see would not be occurring. So those are some of the metrics we’re beginning to look at. Recalls - did I not mention recalls? I thought I said less recalls. Okay. Thank you Valerie. Less recalls.

So there are a whole lot of metrics that we’re looking at. The very biggest one is that you’re all here. We didn’t know how many people were going to be interested in infusion pumps. And this makes us very happy to see you all and all the people who are on the phone. So thank you.

Now I want to move into our next section, which is the user’s perspective and I’ve got to thank several people here for getting these speakers to come. And then I’m going to introduce them as they are going to talk. First of all we have on the group today representatives from Johns Hopkins Hospital.

We have a representative from Brigham Women’s and Mr. Grant, I want to get this correctly because I’m getting a little tongue tied, West Virginia University Hospital. That is correct. Okay. And they have traveled so that they can give their perspective on infusion pumps.

And they were brought together by an organization within the FDA called Med Sun or the Medical Product Safety Network. And as people have alluded to, if you don’t know about it, it is a group of 350 hospitals. The program was launched in 2002 by our center and the primary goal was to work collaboratively with the clinical community to identify, understand and solve problems that occurred with the use of medical devices.

We had those 350 facilities, mostly they’re hospitals, but we have other labs, home care, other groups that also participate. And we get representatives from each of those facilities from risk management, patient safety, quality improvement, biomedical and clinical engineering, physicians, nurses, material management - you name it, multiple staff at those hospitals participate.

And what it does for us is in a very small universe it gives us a partnership between clinical sites and the FDA. It’s a powerful two-way channel for us that we can start identifying problems earlier, start getting feedback and working with them. And we have lessons learned that we share as we go forward.

So I want to thank MedSun for bringing everyone together and I would like to introduce our first group of users. Each of the groups is going to speak for about 20 minutes. We also have Ms. (Polacheck) who is a clinician herself in infusion and Mary Brady who heads up the home care initiatives over here that are going to speak.

So let me begin with Johns Hopkins Hospital and there are several people here so I’m going to try to make this a little brief. They were wonderful. They brought the big guns. So Maria Savak is Assistant Director of Nursing at the Johns Hopkins Hospital with 30 years of experience in the field of critical care, intermediate care, general care and cardiac rehab nursing.

She began practice in the coronary care unit at Johns Hopkins in 1981 and obtained a graduate nursing degree as a clinical nurse specialist. She has worked as a cardiac clinical nurse specialist for three years and in 1987 she began her role as a clinical care nurse educator at Hopkins and pursued this for 17 years.

Her interests have focused on cardiac care including ECG interpretation, cardiac monitoring and complex cardiac drug therapy. In 2005 she accepted her role as the Assistant Director of Nursing for Clinical Standards. She is interested in patient safety issues such as fall prevention, pressure ulcer prevention, alarm fatigue and product safety issues.

Currently she co-chairs a PCA pump steering committee, the monitor alarm management task force and is the central nursing representative to Hopkins’ products committee. We also have Mr. Doyle here. He is a human factors engineer with a PhD with over 30 years of experience in the fields of nuclear power generation, defense contracting, simulation, usability testing of Web sites and cell phones and most recently in health and medicine.

Dr. Doyle has enjoyed human factors analysis, design and test activities since receiving his doctorate in applied experimental psychology with an emphasis in human factors from the Catholic University of America. Given his interest in analysis and design, he orients his efforts to apply solutions for operational challenges.

He presently works within the clinical engineering services department at Hopkins Hospital in Baltimore addressing patient safety issues using human factors techniques. Okay. Julia Faum. All right. Dr. Faum is trained both in emergency medicine and critical care at Hopkins. He provides resident education and supervision in the emergency department as well as the intensive care unit. Because of this cross training he provides expertise in the management of critically ill patients in the emergency department.

His area of research involves evaluating the quality and safety of critical care and emergency medicine. Recent publications involve studying ambulance diversion, EDP management, rapid response teams and error reporting systems. Current work also involves studying medication errors, quality markers, culture safety and the association between information technology, ED staffing and standardized protocols on quality of care.

Mr. Ravitz is a principal professional staff member at Hopkins applied physics laboratory who received his BS in biomedical engineering in 1987 from the Johns Hopkins University. He earned his MS in electrical engineering in 1989 from the University of Miami and an MS in technical management from Johns Hopkins in 2004.

Since 1989 Mr. Ravitz has been at Hopkins in the applied physics laboratory working on programs spanning undersea warfare to biomedical weapons systems engineering challenges. He is an IEEE certified software development professional. Please give a round of applause to these people who have traveled to talk to us from Hopkins.

Maria Savak: Okay. Well, thank you everyone. I didn’t know actually who our audience was until I walked in here today. Ten days ago I was asked by our risk management committee to present Johns Hopkins infusion pump issues at the FDA meeting.

And I thought what can I talk about for 20 minutes? That’s a lot of time. And then Pete walked into my office and told me about the projects he was working on related to infusion pumps. And I thought voila, this is great. I could talk about the issues and then really look at some of the solutions that are occurring at Hopkins. And that’s why I invited Pete, Dr. Faum and Dr. Ravitz to come and help with this discussion.

So what I’d like to talk about is the user perspective and I am a practicing nurse. I do work still on the units although I also work in administration. I do have thoughts about infusion pump issues. I didn’t want to give you exact issues because I didn’t have enough time to pull the exact issues. But what I’m going to be giving you is examples of issues that have occurred.

And I didn’t give you all of them, just about five or six that are common occurrences at our hospital that I thought you should know about. So if you could go to the next slide please. Next slide. What I thought would be important is to talk about cases and how you can draw your own conclusions about these cases and whether it was a pump issue or whether it was a user issue or whether it was both.

But these are the types of things that have happened to us. So the first one is an example of a nurse who goes into a room, hangs a bag of Heparin and when she hangs the bag of Heparin she does everything correctly, sets the infusion pump up correctly. She goes in at 10:00, at 11:00 comes back into the room to check on the patient and what do you know, most of the bag is infused in an hour.

So this is an example of a very smart nurse, very concerned. This shouldn’t happen. She appropriately takes that pump out of service, calls clinical engineering, has them check on the pump, which is what we want them to do and of course what we found in this particular case was an over infusion and it was a result of a faulty door.

And this actually was the impetus for finding the faulty doors on the pumps that we use. Next slide please. The next incident that we had, this was a recent one. And because of my interest in alarm fatigue this one actually came up on my radar screen.

This is an example of a nurse who was in the ICU. She was working with a very potent beso active drug called norepinepheron or otherwise called Levophed. And the patient was dependent on that for blood pressure management. The nurse went in and hung a new bag of Levophed but forgot to put the volume to be infused in.

And as a result after she hung the new bag although there is a new bag hanging she failed to put the volume in. In about 20 minutes the infusion pump starts alarming that the volume has been infused. And so the nurse is in another patient’s room at that time and doesn’t have time, she doesn’t even hear the alarm to be very honest with you.

And it’s a very busy ICU but alarms on pumps do not trigger the same reaction that a crisis alarm on a ventilator or on a monitor might cause. So what happened is everyone ignored it. And then later on within a few minutes the cardiac monitor started to alarm and unfortunately it wasn’t a crisis alarm because people react to crisis alarms very quickly.

This one was a medium priority cardiac monitor alarm and no one reacted to that one either. And what happened was the low blood pressure alarm was occurring from the (arm line). What happened then ultimately was that the patient went into cardiac arrest. And then of course people reacted.

When we investigated the situation what we found is that the blood pressure was actually low for quite a while and we also were able to investigate that it was the volume to be infused just needed to be reset. That was all that had to be done and this could have been avoided. So that would be my second example. Next slide please.

This one and again, as I said, these are examples. These are not exact but this is an example of something where you might think what was that nurse thinking or you might think boy, she really had good critical thinking skills to really come up with this solution.

What happened in this case is the patient is on a nurse managed protocol and in this case insulin. And the insulin is running at let’s say 24 units per hour. The nurse checks her blood glucose and what she realizes when she compares it to the grid is that the patient needs an increase in dose. And in this case the patient needed an increase of ten units per hour.

Now there is a max limit on the particular pump. So she hit that hard limit and she couldn’t go on. So rather than stopping at that point what she did is she ended up getting two pumps and she split the infusion of the insulin into two bags and delivered the infusion on two separate pumps to be able to deliver that.

So again, you might be thinking wow, why would anybody do that? But when you’re busy and you’re thinking that you’re doing the right thing, they don’t always go through those safety checks. And in this case it turned out to be the wrong thing. Next slide please.

This is an example of a PCA issue that we recently had. A nurse is on the unit, she gets an order for hydramorphone, 20 milligrams in 100 mills for PCA. Upon start up of the pump there is a safety feature and it asks is this a new patient. Well, she inappropriately said no and what happened is it defaulted to the previous patient’s settings.

And as a result the pump was not set for an adult, it was set for a pediatric patient and so it defaulted to morphine as opposed to hydramorphone. Next slide. This one is very recent to me and this was we just recently switched PCA pumps at our institution. And we typically when we switch pumps, we usually try them for a set period of time on patients.

But in this particular case we didn’t do that. We did a simulation. And when we did the simulation we did not have the error line detector turned on during the simulation process. So we totally missed it and really I think that had we actually done the simulation on real patients we might have picked this one up but we didn’t.

So we did massive in servicing for our staff on the new pump. We had great vendor support. The vendor was terrific and we decided we were going to implement the pump on a given day. Well, within about a week I started getting incident reports about error line detectors on PCA pumps going off. And we decided we were just going to watch it and watch what was happening with that.

After a month I queried our database of incident reports and we had 68 incidents of error line with this particular pump and we knew that it was a problem. So the company was very helpful in coming in and helping us to sort things out. But what we found is that it really was a design flaw with the error line detector.

And unfortunately the company did not really have an action plan for that error line detector problem. We further investigated and found that the other accounts that were using this particular pump had the error line detector turned off. And so that they had already identified it as a problem but there was again no solution for that problem.

So our staff was very, very frustrated that we had turned the error line detector on and it was actually causing significant problems in terms of patients not getting their therapy. So what we ended up doing is going to our risk management committee and making and institution decision to turn the error line detector off.

Here it is a known safety feature and we’re deciding to turn the error line detector off because of numerous problems. Next slide please. So what is the impact of these issues? Well, there are a number of them. Number one, you design a smart pump and when you design a smart pump people tend to rely on that technology for safety checks.

But the technology may not be such that it still may need the human factor. It needs that person to critically think about what they’re doing. You can’t just rely on those safety checks. Second, reliance on alarm - infusion pump alarms do not convey the same sense of urgency that other alarms such as event alarms would convey.

And in fact I use this as an example. A cardiac monitor has hierarchy of alarm. An infusion pump, at least the ones that I use, do not really have that hierarchy. And so they all sound the same and you really can’t tell if it’s a Levophed that’s alarming versus a normal saline that’s alarming. Another problem is frustration.

When something is occurring and they can’t troubleshoot it, that tends to lead to the fourth problem, which is developing a back up system to try to bypass the safety feature. An example is that hard max when the nurse hit that hard max what she did is she came up with a troubleshooting mechanism of getting two pumps to deliver it as opposed to one pump.

And then fear of your equipment being faulty - so example of that is that very first incident where the free flow occurred. And the staff was very paranoid about whether that was going to happen again. So that is kind of the impact on our staff and there are probably others that I’m missing.

Now in terms of patients and families, this really poses a lot of problems. The PCA issue was a significant problem for our patients and family. The reason is here we’re putting them on a drug via patient-controlled analgesia so that they can push the button and control their pain.

But they’re afraid to push the button because every time they push the button it alarms error line or periodically it alarms error line. So we had patients who said they didn’t want to push the button because the pump is working and they’re afraid if they push it they’re going to cause a problem.

We also feel that when you have these alarms occurring for no apparent reason it creates significant delays in particular in this case pain management for that patient. And frustration - the patients’ families and the patients get very frustrated. Next slide please. And lastly, our facility, the issues that involve our facility are risk management issues.

So patient harm first and foremost, of course indemnity or any kind of incidents that occur and then a lot of the pumps are designed with generating reports but the reports are not necessarily able to be interpreted because there may not be a denominator involved and you’re not exactly sure what you’re looking at when you look at these reports. And Pete can probably talk more about that.

I think we have been very fortunate in terms of the vendors we have been working on have been very helpful. And I would honestly say that I do trust the vendors to have good intentions. But there tends to be a lack of trust when these incidents occur and we hear that they were occurring in other institutions and we didn’t know about that.

So there is that trust issue that could occur. Now that really ends my portion and what I’d like to do is call Pete up and he’s going to talk a little bit about Johns Hopkins applied physics lab and Johns Hopkins School of Medicine and what he and Dr. Ravitz are doing regarding that.

Pete Doyle: Okay. So hello. Alan Ravitz and I are going to tag team here for a minute and I just wanted to start off by introducing - there we go - a little bit about some things that Hopkins is doing.

There are a lot of issues that we’re discussing today and that are common to other medical devices. There are some specifics about the pumps of course but the issues of systems engineering and human factors, usability and safety, they have applied to many other medical devices of course.

So some years ago Dr. Peter Pronovos and Dr. Julia Sam, having realized this, they came up with a concept of establishing a public/private partnership to address some of these safety issues and the idea is to get many stakeholders involved. If I could see the next slide we could show you just basically this initiative was instituted by the Hopkins Quality and Safety Research Group.

It was modeled after something the aviation people did called commercial aviation safety team. In 1995 there was a very bad accident and they decided that it’s time to get all the stakeholders together, not only the air traffic control, the airframe manufacturers, the avionics people, the maintainers and they did this in a global manner.

In 1995 they got a great reduction in accidents because of it. So this served as a model for this, a public/private partnership to provide patient safety. And again, we’re interested in looking at the systems engineering aspects as well as just the clinical aspects. So we have a team here at the Johns Hopkins University applied physics lab.

We’re fortunate in that we have access to systems engineers there and we have our clinicians, our human factor people and structural developers at the hospital. So we’re working together on a project we started back in December and I’m going to ask Alan to explain a little bit about the systems engineering approach and aspects first. Then I’ll come back and talk a little bit about the specific new projects.

Alan Ravitz: Thanks Pete. I’ll be real brief here. Thanks. Just a quick overview of the applied physics laboratory and touch on some systems engineering to lend some context to what people will talk about about the pilot project.

The applied physics laboratory system on the other system organizations associated with the university, some of which are much better known in the applied physics laboratory, the school of medicine, the undergraduate school and whatnot.

The applied physics laboratory is chartered as a university affiliated research center or UARC and in that capacity we perform as a trusted agent providing objective technical guidance to our sponsors for their challenges. Next slide. Currently the laboratory is out in Howard County but it was originally founded in 1942 in the basement of a car showroom right here on Georgia Avenue and Colesville Road.

And the operational challenge at the time was defeat of the kamikaze threat. The APL engineers developed what was called the VT fuse to shoot at the kamikaze threats to defeat the threat. And when the Hopkins engineers took it out into the Pacific alongside the navy to test it out and evaluate its performance they found that the VT fuse that they developed was actually part of a bigger system, the weapon control system.

And the weapon control system wasn’t functioning properly so they transitioned from looking at this one item, this system called a VT fuse into a much broader system perspective addressing a system of system issues associated with the technical challenge.

Since that timeframe in 1942 the laboratory has expanded its focus on system engineering applications across a broad range of challenges facing the country. Next slide. Okay. So systems engineering has evolved as a (distinct) professional discipline in direct response to the developing complexity of new systems.

Systems engineering is itself a multidisciplinary endeavor drawing on the technical expertise of electrical engineers, computer science and what not to evolve through this life cycle that you see here. A key tenant of systems engineering is incorporating user involvement throughout the process and as you’ll hear Pete talk about the pilot project that we’re involved with, we’re really focused on that purple deployment section and the critical piece area.

I think with that I probably should pass it back to Pete to talk about the pilot project.

Pete Doyle: Okay. So we’re very interested in expanding the number of stakeholders involved if we could go on please. In my experience working with the DOD projects, we learned very quickly that getting the systems engineering, making that investment up front is very beneficial in terms of your total product life cycle cost.

It’s monies up front that can be saved by reducing exposure to risk. And we know that the cost has been an issue for some time, even in the early ‘90s when I went to visit my father in the hospital. The gentleman next door was very loudly moaning and moaning and moaning. My father mustered enough breath to comment he must have got his bill.

So we are - this is a public project with the infusion pumps. We hope to expand it to other types of areas as well. This is basically just a context diagram that shows how we’re looking at our projects. Everything from physician orders to training to drug libraries, patient clinical control - all the issues we’re trying to address.

The systems engineering approach of course is interested in looking at cradle to grave concept to disposal of equipment. So next slide please. Okay. So we’re in an analysis phase now and we had a team of clinicians, human factors persons, systems engineering and we had a structural developer there to help us identify what training requirements might come out of this analysis.

And we used a data test analysis of two legacy pumps we had in the hospital and we interviewed our subject matter experts. We are doing an ongoing literature review and walk-throughs to see how the pumps are used in different environments. We had round table discussions of course - all this kind of data collection.

And let’s see. Do we have the next slide please? In the task analysis, which is a classic human factors technique of determining requirements we were decomposing the task to the smallest steps. We did four drugs. We did sentinel delivery, D20W normal saline and fennelefpherone and we reduced them down to sub-tasks.

A sub-task is basically like the sense, the subject or verb and object, a very basic discrete task. And we identified what the prerequisite conditions were for each of those tasks because you can’t proceed down the line until your previous conditions are met, identified or defined as completion criteria for those sub-tasks.

And along the way we were able to identify a lot of these associate design considerations and training considerations. It gives you an opportunity to kind of do a hazard analysis, an SMEA type activity. So we’ve got a lot of documentation, which we’re still trying to get our arms around and summarize. Next slide please.

So these are some of the human interface findings and if I can just deviate from the slide for a minute, I spoke to a human factors person who worked for a pump manufacturer and he left about six months ago. He said that he worked for six different managers at that company.

None of the managers really gave them a legitimate role in the company. They were more interested in marketing aspects and legal aspects and I think perhaps they were overlooking the fact that improved usability or human factors design can help marketing and it can help reduce your exposure to risk.

These are front end investments again that give you benefit in the long-term, reducing costs to yourself, company, to the hospitals and ultimately to the patients. So this is something to consider. It’s difficult for a human factors person to find a legitimate role. I was lucky enough to work in the nuclear power world but that didn’t happen until after the Nuclear Regulatory Commission said you must have a human factors detail control and design review after Three Mile Island.

In the military they say you must have these requirements if you’re going to deliver a product you must meet these human factors requirements. So I think it would be beneficial if we could do something again in this context here to make sure that we’re addressing the appropriate systems engineering activities including usability and human factors parts.

So these are some of the findings and they are probably not that new to you. We’ve talked about balance before. It’s kind of hard to press that button when the device is on an IV pole and it’s running away from you. You feel like you’re doing Morse code on those. But that - we have problems looking at the association of some of the soft keys with the controls, the indication and inclusion was developing was not a very prominent.

We’ve noticed that there was only one alarm. You don’t have a separate alarm if a beso active drug is starting to run low. The nurse doesn’t know whether the alarm is for something less important. Of course there are a number of things here. At the bottom for instance for those in the back, for each drug, each med, it’s important that the right metric is used in the pumps, that we’re not confusing our metrics.

Maria talked about use of two different bags through two different pumps. In that case if you have a brain, which is operating both pumps it should be able to recognize if you were in a (duris) mode and working in a drug library, it should be able to recognize that I’m trying to push the same drug through two different pumps.

That doesn’t prevent someone from using it in basic mode but it’s just another stop gap to remind someone to make sure what you’re doing. Next slide please. These are some of the sample training considerations. It has to be able to cross multiply, has to know when to use the multi-dose option, how to use the secondary.

When some of our PSMs come back when the secondary bags are put below the pump instead of above the pump, which drugs are weight based and things like this. We have a collection of things we can ultimately put into some training requirements. If we move on once more please, we have some systems engineering recommendations that Alan can address. Okay.

So there are opportunities to integrate pumps into the wider enterprise. Of course, everybody is challenging in trying to get space on the WiFi but there may be opportunities to take orders directly from a provider order entry right into the pump. There may be opportunities for if the brains of a pump realize it’s starting to run low it can automatically call for the pharmacy to make up another bag, label it and get it delivered over.

So we’re reducing the workload for the users and we’re making these things happen in a fashion so that they’re not late for the patients. What else? We have developed institutionalized open standards to find hardware and software interfaces. We have not gotten into the destructive testing or evaluation of the software code yet.

But we’re looking at those aspects with respect to how the pump behaves to the user commands. Also finally, to establish good systems engineering practices for a medical device just as we have good practices for other areas. One more please I think. So we started this in December. We’re still trying to boil down our data we’ve gotten so far.

But we want to look further and to look at some of the anesthesia tasks, some of the maintenance issues. We know there are issues. In one of the pumps we have, we have to go to 17 places if you want to find everything you need to do to operate and maintain this pump. There are information policies, there are information procedures, there are safety bulletins.

The fixes come later and they document in different places. We need to collate all this stuff and have this all in one place so the user has a good reference and it’s not taking up and we’re complying with 15 or 17 directives so to speak. We talk about opportunities to integrate pumps and their systems.

And we want to add additional stakeholders as I said earlier. So we started in December. We’re hoping to continue. We expect also that we’ll have opportunities to do some simulation and also to do some rather prototyping for interface design so that we can very quickly go through a number of tasks and just look at different interface designs and make improvements. So I think that’s it. Thank you very much.

Melissa Eakle: First of all I want to thank Johns Hopkins. A lot of information. Look at their faces. There are going to be breaks so that you can talk to them about their programs because they are doing a lot of very interesting work that you should know about and may be helpful.

The other thing is the slides will be up on the FDA Web site on Tuesday. So if you miss anything or have any questions their names will be on those slides too. So thank you again. We appreciate it. I’d like now to introduce our next speaker from Brigham Women’s. This is Mr. Michael Fray. He is the Director of Biomedical Engineering at Brigham Women’s. Michael.

Michael Fray: Thank you. I think I also want to talk to the same (unintelligible) - and there are a lot of people in uniforms in here. So one would think that (we behave) properly. So let me do my best.

I would like to thank the speakers this morning that introduced a lot of concepts around challenges with infusion pumps and also especially the first and the last speaker who actually highlighted a lot of challenges that we see and a lot of the detailed engineering analysis.

And hopefully this will actually start to hold the vendors accountable to delivering safe products for healthcare delivery. And I think the concept of using infusion pumps is such a ubiquitous device in our hospitals as a total levered, safe and efficient there is very important.

Yet you almost have to have three PhD engineering degrees in order to operate these pumps safely and effectively. So let’s try to work with the FDA and hold these vendors accountable so that we can make incremental changes and incremental improvement in these products rather than before we have to go to these all out recalls.

It’s quite punitive and can have a lot of impact on the hospitals. So talking a little bit about Brigham so you can get an idea - next slide - about the Brigham Women’s hospital and some of the concept so you can get an idea about the framework of what happens there, this is our people from the Boston area.

This is the former Brigham’s hospital so this is now primarily the med surgical and oncology building. This is our center for women and newborns. This is our ambulatory site and this is our new cardiovascular center that we just built. And this is what we call our preeminent cardiovascular center of the world and it’s something that we’re very proud of both from a hospital perspective and from a biomedical perspective.

So to get an idea specifically around infusion pumps, we have about 3000 nurses in our hospital. So about 3000 clinicians who actually touch these devices. So the reason I say 3000 is because we have 2800 nurses and we have all the physicians in the OR environment that actually do touch these devices.

We had about 41,000 inpatient visits about a few years ago. We bring about 9000-10,000 births a year. The ED had about 36,000 patients in 2006 and we have 19,000 maybe pushing 20,000 patient care devices in our inventory. So as you will see, these pumps are quite a lot of devices because the pumps are about 16%, 3000 infusion pump devices in our institution. Next slide.

So what is the infusion pump used at the Brigham? The concept that we had developed in around 2003 biomedical engineering and nursing and the medical staff, we developed this vision of developing a closed loop medication delivery system.

So what that entitled was from our chief computer provider or entry system you were going to use our infusion pump platform to the patient in one closed loop system. So that we were quite successful in doing a lot of this. We now have one challenge left. It’s getting response from them so we are slowly getting it.

So our inventory as I mentioned earlier, we have about 1100 program modules and about 2100 of these pumping modules. We selected a pump that allowed safe medication delivery to our patients and the concept of (guardrails) so maybe that will kind of give you an idea of whom I’m talking about came to mind.

And this was a good idea because it prevented some over infusion and also prevented under infusion. So the concept was create from a concept perspective. The pumps are used both for medication, PCA and also to keep the (EBO) activations in our institution by nursing policy. Next slide.

So let’s focus on the positive before we go on to some of the challenges. So the concept was great so we now have DQI data both from the pumps, which allows us to do drug library improvement. It also allows us to do a lot of analysis. The adverse drug events have been reduced based on the DQI data that we now see.

If we look at how many drug libraries, I would say people that bypass them or don’t adhere to those, we’re now seeing a decrease in that. The pumps are also widely used in the institution to the point that people have accepted the technology and are using it widely in our institution. Next slide.

Some of the challenges that I will speak about today and these are some of the challenges I will not speak about the final integration set, which are the WiFi challenges that I think somebody previously mentioned. If there are questions about I would be glad to address those as well.

So there is a high total cost of ownership for the hospital as a basic lack of transparency from there and a lot of it has to do with that these vendors now are starting to struggle with being either an infusion pump company or a software company and a hybrid in between. There is a lot of lack of transparency with that and that poses a lot of challenges for the in-house biomedical engineering department let alone for the people who need a lot of the data for continuous improvement.

The recalls have a long lead time from vendor feedback and implementation so when you tell us there is a recall obviously there is something that is not so good happening. So when it’s a long lead time it’s the risk fostered for the institution, it’s a big concern.

So let’s drill into each one of these and then if there are questions please stop me and I’ll try to address those (if I can). So to give you an idea about why I say a high total cost of ownership so in FY’09 we spent about $500,000 on the manufactured service costs per year.

So to break this down further, $180,000 was spent on parts, $165,000 on outside service and 155,000 in labor. So how much is this? This is almost 45% of our department service costs for pumps that are less than 16% of our inventory. So if you look at a comparison there is not a comparison. And if you look at other devices that are higher in cost that are used even more or in harsher conditions you would expect these costs to be a lot lower.

So CPD devices, they are like the devices that are used in 24 hours a day and high pressure, high temperature environments. They don’t even come close to the $500,000 that we spend on these pumps per year. So from a hospital perspective and the current budgetary situation that all of us in this room are under, this is a big, big challenge.

As a matter of fact, on Thursday I’m about to go into budget steering. This is probably what will be the one thing that will be talked about a lot in our institution. Why did it cost us so much to maintain them? Now you may challenge me and say is this constant.

If you look over the last three years this cost has been almost consistent from looking at it from a broader window perspective. So from a lack of transparency that the service goes without explanation both from the service manual perspective, you look at the service goals, one, the service goals that actually show up on the device are not even in the service manual.

From a software perspective it may be that the vendors are really struggling with the software and they don’t quite know what it is. But from our perspective when you can’t explain to us what it is we cannot put processes in place to avoid failures and compromise the safety of our patients.

So transparency is critical for us biomedical engineers to understand what’s going on and it also means that maybe we need to use the pump in a different way. Maybe we need to develop different processes with our clinician colleagues. We often get service reports that are not clear about the repair.

So the concept here is that we get back and it’s all code, code, code, code has been replaced, right? So somebody said earlier today when you label it correctly, don’t just say broke or broken on it, this is the same thing that we get from the vendors. So it makes this incredibly difficult for us to support these devices in the hospitals.

Now from a technology that is so ubiquitous one would expect that there is a lot of transparency about how to support it and to keep making the device a safe tool for patient care delivery. There is a high number of devices that have to be returned back to the vendor for the same service. So we send the pump out. It comes back within a week or two it goes back out for the exact same thing, which makes you ask what is actually being done on these devices?

So again, transparency is important. Now there is also a lot of imbalance of the flat rate service costs versus the parts costs. So for example, we get a flat service cost of $480 just to send the pump out. But if we were to buy those components separately it would cost us $520.

So it’s impossible that cost of bagging the pump up, boxing the pump up, sending it out, a manufacturer repairs it and then you ship it back to us - it’s cheaper than our cost of repairing this in-house. So there is a lot of imbalance there about what actually happens.

So these are some of the concerns we have. I’m sure that if I was to ask my other colleagues in the audience they probably would keep on going down through the first door and keep on going down in the ground. So for recalls and corrections that actually happened - so after the recalls are issued it takes several months to hear from the vendor as to when the correction then will be implemented.

So to us it’s very, very concerning because when you tell us there is a recall that has been issued it means there’s something wrong. So from a risk perspective I think that previously I was speaking from a risk management perspective, the hospital starts to wonder what is going on. What is it that we have to do to maintain a safe environment for our patients?

So we need to make sure that the manufacturers have a proactive plan in place before they issue a recall and they actually have to do it with expediency. It can’t be that you issue a recall like one of the last recalls that were issued and it took almost four months before we even saw the manufacturer to show up in the place.

As a matter of fact, there just has been another recall issued two weeks ago from another vendor. The vendor doesn’t even have the corrective plan yet in place for that. So we need to make sure that that is in place. The impact on hospitals during the recall is significant.

Hospital operations are often compromised related to this recall because the vendors come out to do the recall and they don’t bring enough pumps to do the replacement. So what happened was in the last recall we did I’m sure people in the audience probably will know what I’m talking about.

And so the vendor sent out a handful of pumps. Luckily Brigham had about 80 pumps in stock that we were waiting to deploy on a new installation. Even with those 80 pumps and what the vendor had sent out it took us almost a month plus to get this recall done.

So during the implementation of that correction what does it mean for the hospital? It meant that a lot of biomedical engineers had to be pulled from something else to do the recall. We had nursing involved because we had to go to floors and actually ask a nurse to stop the infusion, to start another infusion when we changed the pump.

And sometimes we even got pharmacists involved because we needed to get new medication to get this correction done. So the burden on the hospitals to do this is significant. Also the logistics involved with this because we have to coordinate all the pumps getting in, we have to all the coordination of tracking which pumps were done.

And inevitably you’re not going to get 100% of pumps. If you get 85% you’re pretty lucky. So the last few pumps, the burden is significant on us to track it and make sure we’re actually following and getting all 100% of the pumps. So let’s look at it from the medical devices and regulation level. The devices have to meet the regulatory requirements and standards from a design perspective. Next please.

The devices also have to meet a clinical use model. So infusion pumps, I think somebody had said earlier about an infusion pump is not used in a wet environment. So this is - I just took this picture. If you tell me an infusion pump is not used in a wet environment I really would challenge you on that because what’s right above the infusion pump? It’s a bag of saline or a bag of medication.

So inevitably there is going to be some fluid that is going to drip here and I don’t think you can hold a clinician accountable to spiking that bag without anything coming out of that bag. So whenever I think the previous speaker, the engineer that actually said we got some fluid that actually shows up in the connectors.

Well, absolutely you’re going to get some fluid to show up on the connector because when you spike the bag there is going to be some fluid that comes out of the bag. So if the device meets the regulatory and the standards requirement, it may not be well enough to work in a clinical environment because it has to meet the clinical use model of where the device is actually used.

Devices have to be tested in the actual environment of use. So an inpatient room and an OR environment are quite different from the lab that the devices are tested with. If somebody was to tell me that a lab is a controlled environment, you control everything including the temperature of that environment. In an OR or an inpatient room the level of activity of what a nurse actually has to go through in a cardiac room in an OR environment is quite different than what happens in the lab.

So we need to make sure that we test these devices in the actual environment that they will be used to make sure that they are safe before we deploy them. There are a lot of devices that we actually get now are released to the market and when we test them in the environment of use without going through the extensive testing that Johns Hopkins is doing, we find all these gaps and all these issues with the devices.

So we can hold these vendors accountable for testing in the environment that the devices will actually be used. And finally, a recall is recall. So if you tell us there is a recall it can’t last three months, two months before we even get the corrective plan in place.

It makes the risk foster of the institution actually becomes very, very concerning. Whenever a recall is issued and if they all come out Friday in the afternoon - so what happens is it’s a lot of effort that actually happens in hospitals to go from a risk management perspective, a biomedical engineering perspective and the hospital leadership. What is the risk that is going to happen for the institution and how are we going to minimize our risk?

And remember that this is the age of the Web environment so everything you tell us goes to the Web. The same way as we go to the Web in the environment, the patients also go to the Web to know what is going on. So somebody will come to the institution and say why has this recall been acted on devices being used on me? And it’s not unheard of for us to get these questions. Questions?

Melissa Eakle: We’re going to do the same type of thing. You’re going to get cards when you go to lunch and I’m going to ask the speakers because when we go to lunch we’re going to ask them to answer all the questions you didn’t get to ask them at lunch. All right?

So please write down your questions and give them into the registration desk because I know from both of your presentations people have a lot of questions. So thank you very much to Brigham Women’s. We have another user facility who is going to be speaking.

This is Mr. Lee Grant who is the Director of Biomedical Engineering and the Laser Safety Officer at the West Virginia University Hospital. He also serves as the lead MedSun representative there. Prior to becoming director Mr. Grant was also a biomedical engineering supervisor and senior biomedical equipment technician at the facility.

He began his career at Luther Hospital, Eau Claire, Wisconsin - I hope I said that correctly - and received his degree in biomedical equipment technology from Hennaton Pharmaceutical College in Eden Prairie, Minnesota. Thank you Mr. Grant.

Lee Grant: Thanks to the FDA for letting us speak today. Sorry about that. To let you know a little bit about WVUH, we’re located in Morgantown, West Virginia. We’re a little bit over a 500-bed hospital, 30 ORs.

We can run about 1000 lines of IV therapy at a time. So from this user’s perspective infusion pumps are the most problematic device in the hospital. Not a lot of news, I think that’s why we’re all here. I have the same issues a lot of biomedical people have.

I’d like to see a design that’s less susceptible to food intrusion during cleaning or during contamination of fluids. I want to see a longer battery life. I want the pump to be easy to use with big, bright displays, minimal screens to go through during programming, large, easily read information.

When the pump is in operation - this isn’t going to work without these - I’d like to see fewer break downs. Should have known that one. I would like to see devices that are repairable in the field. I hate, absolutely hate the down time that is automatically required by forcing us to send the stuff back with shipping, the shipping time.

But today I want to talk about what I thought was going to be the easy part of IV therapy integration, the whole personal jet pack of infusion pumps, IV therapy integration. Seven years ago when WVU hospitals was evaluating what pumps we wanted to standardize on, we were told we were looking just a little bit too soon.

Right around the corner there were going to be systems where the physician electronically enters an IV order, the pharmacist prepares the meds, they’d be delivered to the RN. The RN barcodes the meds, the patient and the pump and the pump then is wirelessly programmed for what the physician ordered.

And when the infusion is complete the pump sends the appropriate information to the medical record. We were just looking too soon. Well, we had a physician order entry system at the time. I didn’t think that was going to be a big deal. We now have an EMR and currently we’re sending information from our ventilators from our monitoring systems and that is going directly to the EMR.

I don’t think that’s going to be a big problem. The wireless part is. That does still scare me. But so I want to talk about the easy part of this whole system, the drug library. 18 months ago we decided we were going to enhance in force the use of the drug libraries in our infusion pump by all the clinicians.

Now let me start off by saying this isn’t anything derogatory about the pumps that I’m going to be talking about, the ones I happen to have experience with. I think they’re probably all going to be similar. Due to some personnel changes in our pharmacy we weren’t able to locate the original programming software and the files for the existing drug libraries that we had.

This was okay. We were upgrading. The company was going to be sending a new upgraded software for our pumps. This is our syringe pumps by the way. We were going to get a new programming toolbox and I figured I can download the existing system from the pumps and we just move from there.

Not quite as easy as it did and I’ll come back to that in a minute. So pharmacy was going to start revising our drug library. They were going to start this a year ago this last January. Here is a recommendation for the manufacturers. When you send software for something like this send a person along with it.

Every - it was easily a month from the time we got the first attempted to load the toolbox until the pharmacists felt comfortable loading the drugs in to start the process. Support was entirely over the phone and once pharmacy thought they understood the program, everything seemed to be moving quite nicely.

They worked with nursing, physicians, the company - everything was going according to plan. This was the plan. We were going to update the pumps with the new software and drug library during biomeds routine inspection. We were only going to have to touch these pumps once. This was going to be great. The timeframe was to have 170 pumps completed during the month of April. I really wanted to get started mid-July but the plan wasn’t perfect.

We forgot a step in our plan and that step was to see how the pump presents the information to the user. When we loaded the first pump for testing in mid-July we started to notice irregularities in the display. The toolbox allowed for wrong file names in the programming process.

The pump was limited to 30 characters. So we had to start figuring out some work arounds and again for the manufacturers, make the programming tool have the same limitations as the pump. This is something and I was a little late getting here today because I was at a MedSun meeting, which is a group that is supposed to be reporting device incidents to the FDA more voluntary reports.

The number of times I’m going to say here this afternoon we didn’t report is embarrassing. But this is something we probably should have reported. We didn’t report. Just the programming error because of the potential for problems with that. After wading through multiple screens on the pumps just to get through the drug concentration selection process the limited display became an issue.

The pump only displayed four drugs at a time. So to help prevent selection of the wrong drug we started inserting blank lines on some of the pages to ensure that we have the same drug or all the concentrations of a drug occurred on the same page. We didn’t want one concentration of (debevene) on one page and then the other concentrations on the other.

Again, probably should have reported that. We started modifying display names. We used caps and small letters to try to distinguish between drugs, (debutaline) versus (dopamine). We used capital BUT in (debutaline), capital DOP in (dopamine) just so that on the display it’s a little more recognizable.

We also started finding problems with (alphabetization) getting them in the right order. I knew I couldn’t say alphabetization. Yes. We wanted drug names in order of concentration. 200 micrograms per milliliter comes before 2 milligrams per liter, which is before 20 milligrams per liter. So we had to be careful. You have to have proper leading space in order to get the orders right.

Yet didn’t report - these are all work arounds that we developed to try to prevent selections of the wrong drugs and wrong concentrations. Now I said I was going to get back to downloading our previous drug library to the pump. Here we go. It turns out the two files weren’t exactly the same. What we had originally and what I downloaded weren’t quite the same.

There was a glitch or something going on again should have been reported. Depending on the program infusion type, the software presented inappropriate programming screens. Loading those screens were showing up for the users when they shouldn’t have been showing up there.

Apparently there is some special meeting going on at the companies to try to figure out what the difference was because they weren’t aware. They thought downloading from the pumps was going to work too. So monumental efforts by our pharmacy and some phone support from the company - it does work sometimes - on April 2 I was able to send an email out to tell people that we were going to get starting.

And yes, the final decision was made on April Fool’s Day. I should have known better than to move forward by that. By April 14 we had half of the pumps done, half of our 170 pumps were upgraded - not bad. On April 28 the report came down. When (remisentinel) is loaded into the pump and is selected from the drug library, the pump locks up.

(Remisentinel)? We had not made any changes to (Remisentinel) in our drug library. (Remisentinel) had worked fine in the previous software. In conversation with the company they told us that the new software required some entries in some fields that had not been required previously. Nice to know.

They also told us that there was another drug that would do the same thing - also nice to know but incomplete. Once we knew what we needed to look for we found that there were actually four drugs that would lock up the pump. Okay. Time for a scene change a little bit here. Going back to early March on a parallel track.

Another pharmacist was looking at the drug libraries on our large volume pumps and for some well publicized reasons we found ourselves in a position where our single channel pumps and our multi-channel pumps, both of the same model family, had different libraries. After our single channel pumps had received an upgrade and expanded library capabilities at least one of our standard drug concentrations had changed.

Now unfortunately we neglected to make that change in our three-channel pumps. So we had two pumps that looked alike but didn’t act alike. Again, something we probably should have reported. Multi-channel pumps had an old 96-drug library. There were labels that are just labels only. There were some predefined things where we can’t change the names of the labels.

And there was the remainder were fully adjustable. The new drug library had the same sort of predefined labels, function only labels but it had over 400 fully adjustable labels. Somehow we needed to compare these two labels, these two drug libraries so that we could get at least 96 that was in the old system identical to make sure that those were absolutely the same and then the additional ones that were available on our single channel pumps wouldn’t conflict or cause problems.

So we have the drug library, which is a proprietary database from the company. Another recommendation - when you send a proprietary database send a way to at least be able to download it to Excel or something so that we can compare things, we can look at things, we can sort and order and figure out how it’s working overall and look - okay.

We were able to get fortunately I have an excellent senior technician at my hospital and he was able to get the document or export the documents into a spreadsheet. The man has the patience of a saint. We were able to get data into the tables and comparisons to be made within the libraries.

It was critical to find a way to do this electronically so that we didn’t have - we weren’t introducing our old data entry errors. We were a little bit concerned about fat fingers before the stock market was. Although we couldn’t provide a unified drug library across all of our large volume pumps, by mid-May we were ready to at least have a consistent program, consistency between the drug libraries.

Again, this as we were going to do this update, it coincided to our scheduled maintenance inspections. We were only going to have to touch these pumps once. Now getting back to my syringe pump problem, obviously we needed to get word out to the users not to select one of these fatal four drugs combinations.

So after notifying the users about the problems and some fairly lively discussions got started as to what our next steps were going to be, there was a camp that said we need to fix these final four. We need to change the program on those final four and we need you to do it on all the pumps. We need you to do this immediately.

It’s a patient safety issue. Another camp that said let’s tweak the one or two other things that we have noticed in time. It’s still a patient safety issue but hey, we can solve some other problems along the way. And then there was the camp that said let’s get a good solid well tested library together before we create more problems and that was my position.

Then came our one-CC syringe backlog of problems. We were unable to obtain our standard syringe in the available one-CC syringe was now recognized by the pump as a three-CC syringe due to its (needle size), just the diameter of the syringe itself.

So the discussions changed. Obviously we were going to need to see every pump again, we have to reprogram them. We were going to have to validate. We had to go through the entire senior process again. I finally got smart. I gave the job to my senior technician.

I felt better but the problems didn’t go away. In mid-September we started programming a new library that corrected the lock up issues and added recognition for the one-CC syringes. Two days later we found that one of the whole profiles that we did was inoperative. This was getting silly. There were all of these things should have been reported to the FDA; again they weren’t.

A few more false starts and finally after a full year of trying we were able to view the completed drug library system into our syringe pumps. The drug library was supposed to be the easy part of an integrated IV therapy system I thought. We had this many problems. How in the world are we going to get a fully integrated system and get it to work wirelessly?

I don’t know but I think I’m going to find out. There have been other well publicized events recently that are going to force us into looking for some new pumps. Still working in a timeline vacuum but I’m guessing we will have to make a decision on new pumps and libraries and this is for our large volume pumps before some time next year.

We’ll probably be looking again just a little too soon for this although from what I heard earlier it’s probably four or five years away from this wonderful integrated IV therapy system. Well, what’s my point in all of this? Complicated - IV therapy is complicated. Integrated IV is complicated.

To pull it off will require communication and lots of it. Hospitals will need to be able to explain their needs and wants to the vendors. If those needs and wants aren’t met in a safe and easy to use way we have to report it to the FDA and the manufacturers. Manufacturers need to be up front with their products’ capabilities.

What can’t the systems do? Let us know what the risks are. Help us figure out ways to mitigate and eliminate these risks. The answer is not always more training. Ultimately I think we can do it. Hospitals, manufacturers and the FDA can come up with a safe and effective IV system. I don’t know when it will happen - certainly before we get jet packs. Thank you.

Melissa Eakle: Thank you Mr. Grant. And again, please your questions for Johns Hopkins, Brigham Women’s and Mr. Grant from the University of West Virginia. Please write them during your lunch break and make sure that we have them.

I want to actually acknowledge someone. We ask her to come if she has a few words to say. Her name is Mary Brady. Okay. And we always ask Mary to come because there are pumps in home care have a lot in common. She has worked in many different venues, hospital, public health, peace corps, long-term care and home health settings.

She received her Master’s degree in nursing administration and international nursing from George Mason. She has been with the FDA for over 20 years and in 2001 she began the Center for Devices and Radiological Health Home Health Care Committee to monitor the safe migration of medical devices that go into the home.

She continues to chair the Home Health Care Committee and the FDA Home Use Initiative. I’d like you to give a welcome to Mary Brady.

Mary Brady: Thank you and thank you for the opportunity to talk about this wonderful environment that people tend not to think about when they think about infusion pumps.

You heard Maria talk about case studies of RNs who are professionally trained on these pumps and bad things happen. Now put yourself in a home with a 75-year-old patient who is starting an infusion on themselves after only one or two trainings on that pump.

Or a parent with a newly diagnosed 12-year-old who has Crohn’s disease and now they have to overcome their own emotional aspects and the aspects of their child to learn how to operate that TPN and to teach their child how to operate that TPN. Home care is the fastest growing field in healthcare.

I have heard and this is anecdotal. I read it somewhere but I’ll have to call it anecdotal at this point that 50% of infusion therapy goes on outside of the hospital environment. Pumps are portable, they’re compact and they are smaller. People are going to work, they’re going to camp, they’re going to school, they’re going on vacation, they’re playing in sports. They are not lying in bed.

Pumps are developed to substitute for something a person can no longer do for themselves such as total parental nutrition or insulin therapy or internal feeding or they are there to provide relief for something that a person needs through an infusion such as antibiotics, pain management, chemotherapy, antispasmodics.

Therefore, people expect these pumps must work consistently, accurately and simply as what the person’s body would normally do for themselves. Some take home notes, we had a home care meeting yesterday on the home use of infusion - not just infusion but all use of home use devices.

And this is a home use initiative that we introduced back in April, right around the same time that the infusion pump initiative came about. Some of the take home notes that I want you to think about before you go to lunch are manufacturers, you need to seriously consider the existing standards for devices and adopt them where you can.

We just had one approved for home medical equipment, it’s IEC-60601-111. Also, look at the guidance’s that FDA already has for home use products. There will be the one for infusion pumps now and we are also developing a home use guidance labeling.

In the area of a care giver and recipient, we need to look at the barriers that they face. You were talking Michael, about a nice controlled environment and to make things operable in an operating room or an emergency room or a patient’s room. Now let’s think about an uncontrolled environment and that is what home use is considered.

It is uncontrolled. You have to factor in temperature, humidity, air quality, particles in the air. You have to factor in safety. You have to factor in the location where that infusion is taking place whether it’s in a rural or urban setting, whether it’s in a high altitude or low altitude setting.

You have to think about if it’s in an apartment building or a single family home. You have to think about the electrical considerations there. Do they have two- or three-prong outlets in their home? You have to think about safety with needle disposal. You have to think about waste control and how people handle biomedical waste.

The big ones that I like are what we call the pest, pets and children. And you have to think about that there are pets in the home and they are going to interfere with any type of therapy going on in that home. You have to think about children playing with the different parts of the infusion pump. Look at that button. That’s cool.

So you have to think about these things and you also have to think about vermin. One of the big things that we have is that when a person is on some type of therapy in the home and they get re-hospitalized, they can’t take their device in with them because there might be bed bugs, there might be cockroaches.

There might be all sorts of things that that person is bringing from the home into the hospital environment. So when you are designing your products you have to take these things into consideration. You also need to think about the care recipient themselves. They are not professionally trained.

They will never be professionally trained. These are people who are coming in with a lot of baggage as well. There is the emotional aspect, there is the cognitive aspect. There is a physical disability. Pretty much anybody over the age of 50 has at least one type of chronic illness and that has to be taken into account because they are going to be giving themselves care or the caregiver will be giving that care recipient some care.

So think about these things as you are designing these devices. There are sources of training. The training needs to be geared toward the end user. It’s either going to be self directed or facilitated. When I was listening to people yesterday one thing that crossed my mind was something that I had done this weekend.

And I had gone to one of those self check out areas at the grocery store. And I absolutely detest those but I am bound and determined to beat those things. And I’m sitting there going okay, I have three things with bar codes on them to be able to handle this. This was self directed training. I can do this. I can follow those buttons.

So I do a card. It doesn’t recognize the card as it’s going down the line. Put your card by the bag. I did. Okay. I picked it up and I moved it down there. Put your card by the bag and it kept telling me that over and over again. What was self directed now became facilitated because it stopped my transaction completely and sent somebody to come and help me.

So you need to think about these types of things. If you think it’s self directed, is it really self directed or might there need to be some sort of facilitation there? Know that there might be multiple people interacting with that device. If that child is going to school with their insulin pump there is probably going to be a school nurse, there is going to be an after care person.

There is going to be a teacher involved. If they are going to camp there are going to be counselors involved. Know the user. Keep the user involved in the complete design process. Get human factors started at the beginning of your design process and use it throughout the design process. I can’t emphasize that enough.

Find methods to track your devices in the home. Guess what usually goes home because we tend to sell our newer products to the hospitals so what goes home? The older products, the legacy products, things that are out there that aren’t supported so well any more.

There are also newer devices that are marketed toward the home environment however, we have a lot of legacy devices that are out there. So we need to find some way to keep track of these when problems are occurring. I would recommend that you get really good customer service and this was brought up yesterday as well.

It has to be good customer service. The customer service person needs to check with the person and say what is going on? What was happening in your environment while you were using this device? What failed? Did you have problems with pressing a button? Was it something that you were doing that caused the problem to occur? Is the problem with the design interface with the pumps?

So I would recommend that you have really good customer service. That’s your one and only opportunity really to talk with these people who are using these devices in the home. I think there are opportunities in electronic health record area. I think we can find ways to build an algorithm to find problems and say okay, this is what we want electronic health records to help us track these devices.

There are also different paths that devices get into the home. It can go home directly from the physician’s office. Sometimes they go home with the patient from the hospital. Other times upon discharge they call a DME or they’ll call somebody, a supplier, some type of supplier who will then bring that directly to the home.

Sometimes it’s shipped to the home. Sometimes a person goes on eBay or another Web site and just buys their prescription device on the Internet, typically without good instructions for use, without accessories that are compatible and without cleaning instructions.

Let’s see. In the area of post-market I think there are a lot of issues here where people do not want to report problems with devices, especially infusion pumps in the home. The interface is very scary. They tend to blame themselves when something happens with the device and so they don’t want to report.

There are issues with privacy, they don’t want their names to be known out there. There might be insurance issues, they’re afraid they won’t get reimbursed for that. So this tends to lead to no reporting. So I would encourage manufacturers also when you are doing your talking with the person through your customer service that you have a good decision making tree, that you enhance the decision making tree to ask some more probing questions as to what was going on while you were operating that device and it failed or that you got injured or something else happened?

So these are some take home points. All I want to say is know the user and know your environment. Thank you.

Melissa Eakle: I want to thank Mary because she has been very instrumental in making this environment so well known and really moving the process ahead. And she’s very passionate about it, which is what I really love about it.

All right. Our next speaker is (Janice Polcheck). She is currently the nurse manager for InfuScience, a home infusion provider serving patients in the Washington metro area who have acute and chronic IV needs. She has been a nurse for 32 years and 25 of those years have been specifically in infusion therapy as part of a hospital IV team, infusion suite nurse and as a home infusion therapy nurse.

She is a member of the National Infusion Nurses Society and current president of the Virginia Capital Chapter. She holds a BSM from George Mason University and I ask you to give a warm welcome to (Janice).

(Continued in Part 2)