UNITED STATES OF AMERICA

P R O C E E D I N G S

(9:04 a.m.)

ACTING CHAIR LASKEY: Well, good morning. The Circulatory Systems Devices Panel is meeting today to discuss the pre-market notification for the Philips Medical HeartStart Home, K040904.

Ms. Wood, if you can read the conflict of interest statement, please.

MS. WOOD: Before I read the conflict of interest, I'd just like to clarify something on the agenda. There will not be a vote today since this is a 510(k) device. The vote was inadvertently left at the bottom of the agenda. So please disregard that.

The following announcement addresses conflict of interest issues associated with this meeting and is made a part of the record to preclude even the appearance of an impropriety. To determine if any conflict existed, the agency reviewed the submitted agenda and all financial interests reported by the committee participants.

The conflict of interest statutes prohibit special government employees from participating in matters that could affect their or their employer's financial interest. However, the agency has determined that participation of certain members and consultants the need for whose services outweighs the potential conflict of interest involved is in the best interest of the government.

Therefore, waivers have been granted for Drs. Mitchell Krucoff and Joseph Ornato for their interest in firms that could potentially be affected by the panel's recommendations.

Dr. Krucoff's waiver involves consulting with a competitor on an unrelated matter for which he receives an annual fee of less than $10,001.

Dr. Ornato's waiver involves consulting with a competitor on an unrelated matter for which he receives an annual fee of less than $10,001.

The waivers allow these individuals to participate fully in today's deliberations. Copies of these waivers may be obtained from the agency's Freedom of Information Office, Room 12A-15 of the Parklawn Building.

We would like to note for the record that the agency took into consideration other matters regarding Drs. Mitchell Krucoff, William Maisel, Joseph Ornato, Richard Ringel and John Somberg. These panelists reported past or current interests involving firms at issue, but in matters that are unrelated to today's agenda.

The agency has determined, therefore, that these individuals may participate fully in the panel's deliberations.

The agency also would like to note that Dr. Warren Laskey has consented to serve as Chair for the duration of this meeting.

In the event that the discussions involve any other products or firms not already on the agenda for which an FDA participant has a financial interest. The participant should excuse him or herself from such involvement, and the exclusion will be noted for the record.

With respect to all other participants, we ask in the interest of fairness that all persons making statements or presentations disclose any current or previous financial involvement with any firm whose products they may wish to comment upon.

ACTING CHAIR LASKEY: Thanks, Geretta.

If I can have the panel members introduce themselves beginning with Dr. Zuckerman.

DR. ZUCKERMAN: Bram Zuckerman, Director, FDA Division of Cardiovascular Devices.

DR. KATO: Norman Kato, cardiovascular surgery, private practice, Encino, California.

DR. ORNATO: Joe Ornato, cardiologist and emergency physician, Chairman of Emergency Medicine, Virginia Commonwealth University Medical Center, Richmond, Virginia.

DR. RINGEL: Richard Ringel, Division of Pediatric Cardiology, the Johns Hopkins School of Medicine.

ACTING CHAIR LASKEY: Warren Laskey, interventional cardiologist at Uniform Services University here in Bethesda.

MS. WOOD: Geretta Wood, Executive Secretary.

DR. NORMAND: Sharon-Lise Normand, Professor of Health Care Policy and Biostatistics, Harvard Medical School and Harvard School of Public Health.

DR. SOMBERG: John Somberg, Rush University, Chicago.

DR. KRUCOFF: Mitch Krucoff, cardiology, Duke University Medical Center and Director of the Cardiovascular Devices Unit at the Duke Clinical Research Institute, North Carolina.

DR. MAISEL: William Maisel, electrophysiologist at Brigham & Women's Hospital in Boston.

MS. MOORE: Christine Moore, consumer member.

MR. MORTON: Michael Morton. I'm the industry representative and an employee of Sorin Group.

ACTING CHAIR LASKEY: Geretta, if you could please read the voting status statement.

MS. WOOD: Pursuant to the authority granted under the Medical Devices Advisory Committee charter, dated October 27, 1990 and as amended August 18th, 1999, I appoint the following individuals as voting members of the Circulatory System Devices Panel for this meeting on July 29th, 2004:

Warren Laskey, M.D.

Norman S. Kato, M.D.

John C. Somberg, M.D.

George W. Vetrovec, M.D.

Joseph P. Ornato, M.D.

Richard E. Ringel, M.D.

For the record, these individuals are special government employees and are consultants to this panel under the Medical Devices Advisory Committee. They have undergone the customary conflict of interest review and have reviewed the material to be considered at this meeting.

This is signed by Daniel G. Schultz, M.D., Director, Center for Devices and Radiological Health, and dated July 23rd, 2004.

ACTING CHAIR LASKEY: Thanks, Geretta.

Before we proceed with the open public session portion today, I just wanted to introduce Dr. Oscar Tovar who will give us a short presentation on adverse event reports on the AED.

Dr. Tovar.

DR. TOVAR: Hi. I would like to apologize because Ms. Beverly Gallauresi who worked with me wasn't included in the agenda, but she is the first presenter.

MS. GALLAURESI: Good morning. I'll overlook that little oversight. I won't take it personally.

Good morning. My name is Beverly Gallauresi. I'm a nurse analyst in the Division of Post Market Surveillance, Office of Surveillance and Biometrics in the Center for Devices and Radiological Health.

I'll present a brief overview of the medical device reporting system and an abbreviated summary of adverse event and product problem reports associated with automatic external defibrillators.

The medical devices reporting, or MDR, system is a nationwide passive surveillance system which includes both mandatory and voluntary reporting. Since 1984, manufacturers and importers have been required to submit reports to the FDA of device related deaths or serious injuries, as well as events involving device malfunction, that may cause or contribute to a death or serious injury.

The Safe Medical Devices Act of 1990 introduced mandatory reporting or device related deaths and serious injuries by user facilities, most notably hospitals and nursing homes. Voluntary medical device adverse event in product problem reports are most often submitted by health care practitioners, consumers, patients or family members and are received through FDA's MedWatch program.

In general, approximately 95 percent of medical device reports received by FDA are from manufacturers, one percent from importers, and the remainder equally split between voluntary and user facility reports.

Under the medical device reporting regulation, an adverse event is an event whereby a medical device has or may have caused or contributed to a death or serious injury. This includes events associated with device problems or failures, as well as those events involving use error.

The manufacturer and user device experience, or MOD, is a database that includes all voluntary AED adverse event reports received from December 1993 to the present and mandatory adverse event reports from August of 1996 to the present.

Now we'll describe the search methodology we used to obtain the data set of automatic external defibrillator device reports.

All medical devices approved or cleared for marketing have a unique three-letter identified called a product code. We searched the MOD adverse event database by product code for AED. As I previously stated, the MOD database includes voluntary AED adverse event reports from December 1993 to the present, and mandatory adverse event reports from August 1996 to the present.

However, for this analysis we included only mandatory manufacture reports from August 1996 through December 2003.

Medical device adverse event reports contain information about adverse event or product problems, including where and how the event occurred, who was involved, and consequences associated with the reported event. Reports submitted by manufacturers contain their evaluations of the adverse event, including coded conclusions drawn from investigations.

These numbers represent adverse event reports associated with AEDs submitted by all manufacturers for the eight-year period from August 1996 to December 2003. As you can see, the FDA has received 7,644 manufacturer adverse event and produce problem reports associated with AEDs. The number of death reports, 590; injury, ten; and malfunction, 7,044.

These reports are reviewed in detail to assess signals of actual or potential device related problems.

The MDR system, while providing signals of actual and potential device related problems, has some limitations. Under reporting of adverse events to hospitals, manufacturers, and the FDA by health care practitioners is a well known and recognized phenomenon. Thus, events reported to the FDA represent a subset of the total occurrence of events.

In addition, manufacturers are not required to submit denominator information, such as the number of devices manufactured, distributed, and implanted. Thus, due to under reporting and lack of denominator data, accurate incidence rates are unable to be determined based on these data.

Furthermore, reports received may not be representative and reflect a variety of reporting biases. Thus, for example, reporting may vary by manufacturer and by the presence or absence of publicity. Because adverse event reports vary in completeness and details, causality often remains uncertain.

Dr. Tovar will now discuss in more detail reported problems associated with AEDs based on review of mandatory manufacturer adverse event reports that have been submitted to the FDA.

DR. TOVAR: Thank you, Beverly, and again, please accept my apologies.

I am Oscar Tovar. I am a medical officer in the Office of Device Evaluation and in the Office of Surveillance and Biometrics.

This morning I am going to present a descriptive analysis from adverse event reports on automatic external defibrillators from 1996 to 2003. The benefits of early defibrillation in public places have been shown in numerous studies as the PAT trial (phonetic), the Chicago area airport, and the Las Vegas casinos.

Along with this, there is a steady increase in the deployment of automatic external defibrillators. The estimated AED growth rate for the United States was 8.2 percent for 2000 and 2001; 11.5 percent for 2002; and 22 percent for 2003. It is estimated to be about 20 percent per year in the next five years. This data was obtained from Cross and Sullivan.

These are the AEDs shipped in the United States for years since 1999 to 2003 and the forecast for the next six years. The plot shows a progressive increase in the number of AEDs shipped.

The success of early defibrillation implies that the AED works in the first attempt and consistently in the following attempts, if necessary. The AEDs, as any other device, are subject to failure, but an AED failure to deliver a defibrillation shock decreases significantly the probability of survival of a patient in ventricular defibrillation. This association of device failure and survival highlights the importance of the awareness of these failures for wherever. There is scarce information about adverse events associated with AED use.

The goals of this study were, one, to assess adverse event reports, particularly death, associated with AED failure from 1996 to 2003; and to determine AED component failure or factors that resulted in failed defibrillation associated with death.

For this purpose, Beverly and I review medical device reports submitted by AED manufacturers to the FDA for AED related adverse events. The MDRs were received from August 19, 1996 to December 31st, 2003. We analyzed the MDRs using the manufacturer and user facility device experience or MOD database from the FDA.

Ms. Gallauresi mentioned before that the conclusions and the determination of the component failure reported by the manufacturers were grouped in categories and were used to assess the association of device and component failure with the patient death.

I have arbitrarily separated data from 1996 to 2003 in two groups of four years each. The early years, that's the way I call it, the first four years, from 1996 to 1999, and the recent years, from 2000 to 2003 because AED availability and technology was somehow different.

From 1996 to 1999, we have 191 deaths associated to an AED failure. Of course, these are the deaths that were reported.

We have also 1,579 malfunctions and only six injuries and the category that classify things that could have fit into these previous categories as other.

From 2000 to 2003, 399 deaths were reported related to an AED failure; 5,465 malfunctions; and four injuries.

This view is intended to show the difference between the early years of AEDs, of AED deployment, and the four more recent years of AED deployment. The malfunctions have increased almost three and a half times and death more than doubled.

But we have to have an account, increasing numbers of AEDs. The ratio of death to malfunction is about 12 percent for 1996 to 1999 and seven percent for 2000 to 2003.

Now I am going to present the different categories per year. The report of malfunctions increased from 105 in 1996 to 1,917 in 2003. It is easy to associate this increase with increase of AED deployed. Maybe better device self-agnostic. With this I was to say that with self-diagnostic, want to imply the device detects the malfunction before use by a patient. You're in daily, weekly or monthly self-diagnostics. That means that it's not during the use of the patient -- on the patient, but during the self-diagnostic.

There were very few injuries per year. The maximums were three injuries per each year, and as you can see, there are very few per every single year during this period.

From 1996 to 2000, there were below 70 deaths per year, but in 2001, there was an increase over 100 deaths per year and have remained about 100 deaths per year associated with the increase probably in the AED numbers of probability reporting. There are several possibilities.

Twenty-six manufacturers reported during this period. The results are reported as percentages because in some instances there were multiple conclusions per report. This is a retrospective and ‑- I'll say it again -- descriptive analysis because the absence of an accurate denominator, even if we know the number of AEDs that have been employed, it is extremely difficult to determine the number of devices used during this period.

That's why it is important that we take this data with caution, and I'm going to mention this again later in the presentation.

The following slides will show the report of results of the failure analysis in the two periods of time. These were the most frequent conclusions.

No conclusion with 32.3 percent. No device failure with 26.7 percent. Device failure cause or was related to event in 22.4 percent. Unknown if the device contributed to event in 12.2 percent. And user error caused or contributed to event with only 4.4 percent, and device maintenance contributed to event in two percent.

There is a little confusion, what is every single category. For this I thought that the best way to explain it was with real examples. I had edited the reports in an effort to remove any identifiers or note that it is not literal, and what I am going to read are event descriptions.

For example, an example of no conclusion can be drawn and also device failure occur and was related to event. This was the report. Reporter alleged that while attending to the defibrillator, a patient who had been in a car accident and was in full cardiac arrest, the device delivered two shocks but then did not deliver more shocks. The patient subsequently expired. The report indicated the patient outcome was not a result of the reported malfunction.

An example of device failure, this one. Device failure directly caused event. The reporter alleged that when attempting to defibrillate the patient believed to be in ventricular defibrillation, the defibrillator did not deliver a shock. The defibrillator displayed an error message, and the defibrillator could not be switched on.

An example of device failure occurred and was related to event. The reporter alleged that medics were attempting to defibrillate a patient, but the device would not discharge. The medics attempted to shock the patient a total of four to five times, but the device continued to not discharge. The medics then obtained another device and defibrillated the patient. The patient subsequently expired.

And the last example is an example of user error caused event and device perform according to specifications and order. The Complainant alleged that while attempting to defibrillate a patient with paddles, the user charged the device with the paddles and the paddle container and the device appropriately displayed an error message and failed to discharge.

Second, third, fourth and fifth devices were used in an attempt to continue treating the patient with the same result. Each device had a set of paddles previously attached. The user did not remove the paddles from the container on any of the devices prior to charging the energy. That's the way you read this report.

Now I'm going to show the results from manufacturers' conclusions from 2000-2003. The no conclusion has increased to 60 percent. The no device failure here is 12.1 percent. The device failure caused or was related to event in 9.6 percent, and the remaining categories add to three percent.

This is just for final comparison between the two periods. The major difference is where in our conclusion and in none if device contributed to event; the non-conclusion slice, as I said before, has increased from 32.3 percent to almost 60 percent. On none if device contributed to event decreased from 27 to 15 percent.

As I said before, what I presented were the conclusions about the events, and now I am going to present what were the components of the device that failed during those attempts. These are the results of the failure analysis in terms of the component or the factor that failed in the case of an AED failure.

Electrical component was the most frequent result reporting with 42 percent. The electrical component included, for example, diodes, relays, circuit boards, switchers, capacitors, et cetera. There are something like 21 components reported in the category of electrical component.

Device performed according to specifications was reporting 32 percent. Device operating outside of specifications in 4.4 percent, and mechanical problems in almost two percent.

From 2000 to 2003, electrical component was 36.5 percent. Other was 30 percent. Device performed according to specifications, 21 percent. The defibrillator subassembly is a new category here and includes, for example, pads, cables, et cetera, and the remaining categories add to five percent altogether.

Electrical component decreased from 42 to 36 percent. Device performed according to specification decreased from 32 to 21.1 percent.

We have to keep in mind again that there is an increase in AED's availability.

The AED recalls for this period are shown here per year, and as you can see, the recalls per year were between zero in 1998 to six in 2000 and 2003.

These results suggest that the number of reported deaths associated with AED failure is actually more frequent than injuries. The number of reported AED failure is increasing along with increase in AED reporting or deployment.

There is a relative decrease in reported electrical component failure. There is a relative decrease in reported device operating outside the specification, which includes use area, and increasing number of reported deaths over time associated with AEDs may have several contributing factors, including increased device availability.

Thank you for your attention.

ACTING CHAIR LASKEY: Thank you very much, folks.

Are there any questions for the presenters from the panel, questions, comments? Several. Good.

Dr. Maisel.

DR. MAISEL: Oscar, just so I want to make sure I understand what you presented, it looked like there were about 7,600 events and about at least in the early portion of your data 25 percent or so were identified as being caused by the device. Does that mean the total number of device related events was 7,600 divided by four or around 1,800 or 1,900 that we can conclude were due to the device?

DR. TOVAR: No. It's not that you can divide it by four because that's why I presented per year. There is a progressive increase per year. So the devices or the malfunctions reported, including death, are increasing per year. So injuries, malfunctions, injuries not too much, but malfunctions and death were increasing per year.

DR. MAISEL: Right. What I'm trying to get at is you have reported 7,600 events, and I am trying to understand. I understand that many of those events might be reported, but not due to device malfunction.

DR. TOVAR: That's correct.

DR. MAISEL: And it appeared that about 25 percent were due to device malfunction or that was the manufacturer's conclusion.

DR. TOVAR: Yeah, if I understand your question well, I said during the presentation that some of the report or some of the malfunctions were caught during the device diagnostics. For example, it was not during the use, not during the use on a patient. It was, as I say, during a daily diagnostics, weekly diagnostics, yearly/monthly diagnostics. But it was not all the malfunctions were -- actually there are not too many during an actual use.

ACTING CHAIR LASKEY: Mitch.

DR. KRUCOFF: Thank you both for just sort of giving us the perspective. I wonder if you could help me understand, given the level of illness that patients have in order to deploy this device in all the other vicissitudes of putting an MDR type of data set together, can you help us understand how you would decide or try and understand whether a patient who is fibrillating would or would not have survived if the device had functioned well, i.e., how do you conclude that the device's malfunction is associated with a death event when essentially the sudden death is the presentation?

How the device causes a problem versus how it just fails to turn around a problem?

DR. TOVAR: Right. I think we both can answer this question, but Beverly is always ready to jump.

MS. GALLAURESI: That's kind of the unknown. Sometimes it's very obvious when the device completely fails and doesn't work. I mean obviously the patient is in fibrillation and so then you don't know if the patient could have been resuscitated or not.

I mean, is it answering your question or am I misunderstanding your question?

DR. KRUCOFF: Well, somewhere in there you're coming at least in some events it looks like to a conclusion that the device malfunction was somehow ‑-

MS. GALLAURESI: May have caused or contributed.

DR. KRUCOFF: Right.

MS. GALLAURESI: And then it's the manufacturer that evaluates the device and they report their conclusion codes, and then we have to take that information as the manufacturer reports it. We can't really assess.

With these reports you can't really have cause and effect. We can tell when a device fails, and then a patient wasn't resuscitated. That's the information that we have. So it's the great unknown.

I don't know if perhaps a patient never would have been saved.

DR. KRUCOFF: Right. Even if it had worked ‑-

MS. GALLAURESI: Even if it had worked.

DR. KRUCOFF: -- would they have been successfully defibrillated.

MS. GALLAURESI: Yes.

DR. KRUCOFF: That's what I find would be very -- so there are no criteria per se to understand that because I can't imagine what it would be.

MS. GALLAURESI: Have to look above.

DR. TOVAR: I would like to add to this that, yes, the manufacturer reported a death that was associated with the use of the device. It doesn't matter the situation. However, sometimes it's not related. That's why I brought up an example in the car crash. It looks like the victim was really injured, and it could have had several, multiple causes to lead to the cardiac arrest of this patient. The AED was used, but the patient died. The manufacturer reported this event, but probably the cause of the death was not the ventricular defibrillation or cardiac arrest.

DR. KRUCOFF: Oscar, all I was trying to understand in your two pie charts over seven years, as the device failure caused or was related to event drops from 22.4 percent to 9.6 percent. I guess what I was trying to understand is is that a difference in reporting, a difference in interpretation. Are the devices doing better or just what? How do you interpret that change in percentage over seven years?

DR. TOVAR: Right, right. Probably we'll leave it at that unless you have a --

DR. KRUCOFF: No, no. Thank you.

DR. TOVAR: Okay.

DR. SOMBERG: In the device caused the death outcome, do you have any direct data where the device fired and caused fibrillation or is most of the data the example you gave, that the device being placed on the patient did not successfully work by having a save?

DR. TOVAR: Yeah, that's a very good question, and that's what I was trying to remember, that I was going to comment on the previous question.

These devices when they say directly cause is when they did not fire, when they did not shock. At least I have seen many of these individual reports; in any single occasion I saw something that the device caused by the shock. It caused death because it didn't fire.

DR. SOMBERG: I see, and is it also correct to state that from my understanding the engineering behind these is that there's a default mode and that the device has question of the arrhythmia due to motion, due to other intervening confounders? It goes to default and does not fire as a safety component?

DR. TOVAR: Yes. If the device diagnoses a non-shockable rhythm, it won't fire even if the patient has been in cardiac arrest, and if the patient has been during a long time in ventricular defibrillation, for example, it has a great or high probability to grow into a system with --

DR. SOMBERG: We're not finding for asystole. We're not -- but if there's motion or other artifact --

DR. TOVAR: For those things, that will depend, yes.

DR. SOMBERG: And there's a question. It doesn't fire to say given the benefit of the doubt the algorithm is not to fire. So the device may have actually worked and the vicissitudes were that there was artifact and it could not identify the proper rhythm.

DR. TOVAR: Yeah, that's correct.

ACTING CHAIR LASKEY: Well, thank you both for this exhaustive effort, and I guess just to quickly sum up, you're doing the best you can with extremely limited data set. I wouldn't say that there's a lack of a reliable denominator. There is no denominator. It's not that it's unreliable.

And I think if we wanted to fill in the other three cells in a two-by-two table, we can. You're just giving us one cell.

So what can we do with this? I think it's a blip and it's something important to note, but it just underscores the importance of getting accurate statistics if we're going to make use of them.

But we certainly appreciate and applaud your effort, and I guess on behalf of the panel we would just ask the agency to continue to seek out ways to improve the data collection so that we can get more accurate handles on these things.

DR. TOVAR: Thank you.

ACTING CHAIR LASKEY: Thank you much.

We have a busy, busy morning, and I'd like to commence with the open public hearing, and before we proceed with the roster of speakers, I have one brief statement to read, which is that both the Food and Drug Administration and the public believe in a transparent process for information gathering and decision making.

To insure such transparency at the open public hearing session of the Advisory Committee meeting, FDA believes that it is important to understand the context of an individual's presentation.

Are you timing me on this? Is that the ‑-

(Laughter.)

ACTING CHAIR LASKEY: For this reason, FDA encourages you, the open public hearing speaker, at the beginning of your written or oral statement to advise the committee of any financial relationship that you may have with the sponsor, its product and, if know, its direct competitors.

For example, this financial information may include the sponsor's payment of your travel, lodging, or other expenses in connection with your attendance at the meeting.

Likewise, FDA encourages you at the beginning of your statement to advise the committee if you do not have any such financial relationships. If you choose not to address this issue of financial relationships at the beginning of your statement, it will not preclude you from speaking.

With that, I would like to being the open public hearing session this morning. Speakers, as previously forewarned, will be limited to we're now at five minutes for their presentations, and we have the little electronic timer up here just to keep everybody honest. As I say, we have a lengthy schedule.

The first speaker of the morning is Dr. Mickey Eisenberg. Dr. Eisenberg.

DR. EISENBERG: Thank you.

My name is Mickey Eisenberg. I have studied out of hospital cardiac arrest for almost 30 years as a clinician and researcher at the University of Washington. I'm also the Medical Director of the EMS Program for King County, Washington.

MS. WOOD: Pull the mic up just a bit, sir. Thank you.

DR. EISENBERG: As to financial disclosure, I am here on my own coin. I have receive no salary support or honoraria from defibrillator manufacturers. Two defibrillator companies have contributed to a University of Washington research fund which I have used in the past to support the salary of a research assistant to study out of hospital cardiac arrest.

Let me start with a few facts. Eighty percent of cardiac arrests occur in the home. Defibrillation is the only effective therapy for ventricular fibrillation, or VF. VF usually occurs with little or no warning. Defibrillation, if delivered quickly enough, leads to a very high survival rate. When delivered in two or three minutes from collapse, 75 percent of patients survive. When delivered in ten minutes, survival rate falls to ten percent or less.

And I might point out that that's the situation in most communities throughout America.

AEDs are effective, safe. Their operation can be readily mastered by lay persons. We've trained several hundred seniors in the use of AEDs using only a ten-minute video.

Patients resuscitated from VF generally make good recoveries, and the shorter the time from collapse to defibrillation, the better the neurologics outcome.

You will undoubtedly hear in the coming few hours about the efficacy, safety, and labeling of AEDs, and rather than talk about these issues, I would prefer instead to address the issue of dissemination.

I believe widespread dissemination of AEDs, especially in the homes of higher risk patients, offers the means to improve the current grim mortality statistics. The question is how best to achieve this. Do we use the current medical approach or do we use a consumer approach?

In the medical approach, which is what we have now, physicians control dissemination. The device is deemed potentially dangerous. Thus, prescriptions are required. Reimbursement by insurance companies may or may not occur. Cost effectiveness studies demanded by insurance companies and HCFA are near impossible to do because of the ever shifting nature of indications for ICDs.

Manufacturer costs, sale priced to the patients remain high because of modest distribution and lack of competition. The net effect is limited dissemination in people's homes.

Contrast this approach, the medical approach, to the consumer approach. Because the device is considered safe and training is simple, it's my hope that this committee will recommend over-the-counter status. Economies of scale will lower the cost. Competition will increase, and the net effect will be a lower price.

Like any consumer choice, the consumer decides whether there is adequate value for his or her money. I suspect many older adults will consider $700 for a home AED as good an investment as optional side airbags, carbon monoxide monitors in their home, home security systems, and any other personal safety device.

Clearly, this argument is a simplification of a very complex subject. Nevertheless, the existing prescription based medical approach is leading to only a trickle of AEDs in the homes. A consumer approach with over-the-counter status is, I believe, the best means to achieve widespread dissemination in people's homes, and that can only result in more lives saved.

Thanks very much.

ACTING CHAIR LASKEY: Thank you, sir.

The next speaker on our roster is Kelly Harris. Ms. Harris. Is there a Harris in the group? Yes.

MS. HARRIS: Good morning. I'm Kelly Harris. I'm a sudden cardiac arrest survivor.

I was flown here by Philips who made the defibrillator that saved me a year a half ago.

I don't really have anything prepared, but what I want to do is just put another face to a survivor because I know a lot of people think that -- excuse me. I'm very nervous, too -- I know a lot of people think that sudden cardiac arrest only happens to maybe senior people or someone who is unfit or overweight, and as you can see, I'm quite the opposite of that.

This happened to me when I was only 27 years old. When I went to Philips about six or seven months ago to meet their team up in Seattle, they offered me my own home defibrillator for free, which I thought was great. It was an amazing offer.

And as I got to talk to them I said, "Well, I'm going to be living alone soon. I have my own implantable defibrillator. So I don't really need it for myself, but I want it to go to my family because, you know, whatever condition I may have might be hereditary, and I would feel much safer to have it around them."

And immediately I was told that that might not be possible, and that even for myself to get it, I would need a prescription from my cardiologist, and so that was surprising because I thought, well, I have already had my cardiac arrest. What more proof do you need that I need one?

So anyway, we went ahead and contacted my cardiologist, and he said that, first of all, it wasn't his top priority. So right there that was a lag in time for me to get one, but he said, again, he was happy to do it. He would just want to do more research first.

So from the time I contacted him to the time my sister and her family got the defibrillator I would say it was about six or seven weeks. And so that's a long time since sudden cardiac arrest is sudden, and it can happen at any time. And in that seven weeks, someone could have died in that time.

So anyway, it ended up happening, and they have it. It's around my family, which is all that I wanted, and I don't know. Basically I'm just here to say that I believe it's like having maybe a fire extinguisher in your house. You don't need one after your house burns down. You pretty much need one before that.

The same with this. It's going to be too late when someone has a cardiac arrest. That's not the time to go ahead and prescribe them or their family a defibrillator. It should be three in one who wants one. It can't hurt anybody because it only allows a shock if there's a shockable rhythm. So it reads the heart rhythm.

I couldn't put it on anyone conscious or unconscious that doesn't have a chaotic heart rhythm. So regardless if you push the button, it's not going to do anything. It can't hurt anybody. So if that's the concern with liability, that doesn't exist.

It also is helpful because it will walk you through the steps of CPR as well, and so it just speaks very clear English. As long as you understand English and can push a button, you'll be fine.

And last year I was flown to New York to do a video news release with Brandy Chastain, who is their spokesperson, and once that was complete, we actually attended a Girls Talker Camp, and Brandy was there to show all of the girls how to use the AED. She went through it one time and then had a young lady do it who was probably about, I'd say, 13 years old run through it, and she did it correctly the first time on the dummy.

So it's very easy. It's not just for adults to use or for medical personnel. Children can use it as well.

So I guess that's all I have to say, and I just wanted you to know that this can happen to anybody. It could be your child, your sister, niece, nephew, grandchild.

So all right. That's it. Thank you.

ACTING CHAIR LASKEY: Kelly, thanks for your time. You're becoming an excellent public speaker. It's nice to know our soccer team has great maturity as well.

All right. The next speaker is Dr. O'Connor.

DR. O'CONNOR: Good morning. Thank you for permitting us this opportunity to speak.

My name is Robert O'Connor. I'm the President-elect of the National Association of EMS Physicians.

The association is an organization of EMS medical directors, as well as other pre-hospital care professionals who are committed to excellence in pre-hospital care.

Regarding financial disclosures, I'm here on my own funding. In the past I've received indirect research support from AED manufacturers to support the salary of a research assistant.

Sudden cardiac death is one of the major public health problems. It has claimed as many as 350,000 lives per year. Many sufferers of cardiac arrest can be successfully resuscitated. This requires integration of 911 access, bystander CPR, prompt defibrillation, and pre-hospital advanced life support. It must be integrated. It must be accomplished in a very timely fashion.

Since AEDs have been developed, it has made early defibrillation feasible, first, by EMS responders, fire personnel, then by nontraditional police, security guards, et cetera, and finally by the lay public, as has been recently demonstrated.

Access to AEDs must not result in prolonged delays in activation. So we encourage the integration of a 911 response with the use of an AED.

Making AEDs available to non-traditional responders or minimally trained bystanders is an effective strategy for achieving early defibrillation in many communities. Regardless of the deployment strategy, we must insure that these AED programs are integrated into the local EMS system and included in their quality assurance programs.

Integration of AED programs into these systems is essential to insure the minimal delays take place during resuscitation.

So in summary, we would like to speak in favor of the removal of the prescription requirement for AEDs, with the understanding that either through labeling or intrinsic properties of the device itself, that it specify training and CPR as well as AED operation to anyone who is potentially going to use the device; that the device be located in an immediately recognizable and accessible location, recognizing that if this is in the home, the occupant of the home may be the person who suffers cardiac arrest and there may be a bystander not familiar with where they keep things in the house.

And then finally, that the requirement for integration with existing 911 systems, namely, through first and foremost 911 activation, be contained within the device as well.

Thank you.

ACTING CHAIR LASKEY: Thanks much.

I understand there's a different speaker this morning from Cardiac Science other than Kenneth Olson. So can the representative from Cardiac Science -- thank you.

MR. McKEAN: Good morning. My name is Matt McKean (phonetic), and I am the Director of Regulatory Affairs for Cardiac Science, speaking on behalf of Cardiac Science and Ken Olson.

Of course, as an employee of Cardiac Science, they paid for my travel and my salary.

My comments here today are I didn't know if I was going to be before or after. So I'll adopt this accordingly.

Based on the results of the PAD study and the clinical evidence that is now in place to support expanding the deployment of AEDs into the public domain, and since the majority of the SCA events occur in the home, granting easier access to AEDs will put more AEDs in homes and improve the survival rate of SCA victims, as has been discussed thus far.

Should the panel recommend over-the-counter for AEDs and FDA adopts this decision, Cardiac Science is calling for FDA to implement a least burdensome approach for all AED manufacturers to use to obtain rapid 510(k) clearance of qualified devices. This least burdensome approach could come in two forms or perhaps others, but two that come to mind are issuing a guidance document within 30 days to stakeholders including industry and FDA reviewers that allows simple modification of labeling to remove the on the order of physician or prescription requirement language from the IFU as a simple notification submission to FDA.

An alternative approach would be for the use of a 30-day special 510(k) vehicle currently in place to modify the labeling and present that to FDA.

Also, regarding classification of the device, currently classified as a Class III, FDA should consider down classifying the AED to a Class 2 for the following reasons. AEDs are cleared under the 510(k0 regulatory framework and do not require PMA application. Clinical studies have been conducted and published to support the safety and efficacy of AEDs when used within labeling.

And, second, in the event of AEDs become over-the-counter approved, the integrity of the regulatory classification scheme for Class 3 devices would be compromises or, i.e., contraindicated.

So those are my two comments that I'd like to present both to the panel and to FDA.

ACTING CHAIR LASKEY: Thank you.

Next up is Richard Lazar.

MR. LAZAR: Good morning. My name is Richard Lazar. I'm the CEO of the early defibrillation Law and Policy Center, which should tell you I'm a recovering lawyer.

I am here on my own dime and on behalf of EDLPC. I am not here at the request of or on the dime of any of the manufacturers. In the interest of disclosure, I have in the past done consulting work for two of the major manufacturers and occasionally I'm invited to speak at conferences sponsored by manufacturers.

I have provided the panel with written submission which goes into some detail in terms of my views of the current prescription model and how it operates in the real world of public access defibrillation, and I won't reiterate those comments here. It does go into the whole issue of supervision under the direction of a practitioner and adequate indications for use.

My comments this morning are directed really and admittedly at a high level on the issue of public health policy, and here's what I think we know in that regard.

We know that SCA strikes somewhere between 250,000 and 450,000 people annually in the United States. We know that most of those events occur in public places or the home. We know that the frequency of sudden cardiac arrest in particular venues is unpredictable and perhaps unknowable. We know that most SCA events are caused by ventricular fibrillation.

We know that currently the survival rates for sudden cardiac arrest are somewhere on the order of five percent on an annualized basis -- I'm sorry -- on a nationalized basis, and what that means in real terms is somewhere between 240,000 and 430,000 people die from this condition, and only about 12,500 to 22,500 survive.

We know that rapid defibrillation with AEDs is a safe and effective therapy capable of successfully treating VF induced SCA. We know that based on the current design and usability characteristics of AEDs that the devices are being promptly and properly used by both trained and untrained users in a variety of venues. We know those things.

The conclusion that I draw from those facts is that widespread deployment of AEDs us a public health solution that will, indeed save thousands of lives because AED coverage areas in terms of geography is limited, that is, how long it takes someone to retrieve and use the device.

We know that really the solution is to have AEDs deployed throughout places of daily life. We know those things to be true. At least I believe them to be true.

With regard to the prescription model, which is really the issue before the panel this morning, the prescription requirement currently in place today adds an unnecessary layer to the purchasing process for those that want to buy and deploy AEDs, and it creates a perception that AEDs are difficult to use and are not designed for use by lay people, when in fact the data suggests otherwise.

The perceived benefits of the prescription model which we derive from sort of the drug prescription relationship between a physician and a patient doesn't transfer well to the public access defibrillation environment, and reasons are described in my written submission, but basically a drug prescription model is a one on one relationship whereas an AED prescription model really involves the doctor and three potential persons, the buyer, the user, and the patient with regard to the AED.

So the notion of a consultative or instructive interaction between a physician and a patient doesn't occur in the public access model nor, frankly, could it based on how the system works, and the notion of shared information between a physician and a patient about risks and benefits in those sorts of things can't take place.

So, again, my judgment the prescription model simply doesn't transfer well to public access defibrillation.

Finally, from a public health perspective, the question I pose to myself is what would change if over-the-counter status were granted for AEDs? It would be easier for people and organizations and corporations to buy and deploy AEDs. It would create a perception in the mind of potential purchasers that AEDs are, in fact, easy to use and intended for use by lay people

From a risk standpoint, which is certainly an FDA mandate, would more people die from sudden cardiac arrest? The answer, of course is no. Most people are dying already.

Would more people survive who suffer sudden cardiac arrest? And I think the answer from the data we have today is absolutely yes.

So unlike drug interactions and issues relating to the taking of drugs, the issue with sudden cardiac arrest is very binary from a public health perspective. People either live or people die, and the only variable that we can impact here is the promptness with which defibrillation occurs, and that means to continue this effort we have already undertaken over the last decade, for the first time in the health care system industry, continuing to put these therapeutic medical devices in the hands of lay people.

And by the way, just a final note not related to anything else, this is a policy change that wouldn't cost the government any money. People who buy and deploy AEDs pay for them themselves. Whether the insurance industry ultimately does will remain to be seen, but by and large this is not a government funded effort.

So that's the conclusion of my comments. I'm happy to take any questions the panel might have.

ACTING CHAIR LASKEY: Thank you.

MR. LAZAR: Thank you.

ACTING CHAIR LASKEY: And our last scheduled speaker is Dr. Gordon from the Red Cross.

MS. WOOD: Actually Dr. Gordon submitted a statement to be read into the record.

This is the American Red Cross position statement regarding over-the-counter automated external defibrillators.

"Sudden cardiac arrest can happen any time and anywhere, and it claims the lives of more than 680 Americans each day. The American Red Cross believes that this is a tragedy that can and should be prevented. We believe the introduction of an over-the-counter AED would be a positive step toward insuring that properly trained citizens are better able to respond to an unexpected cardiac emergency event.

"The Red Cross continues to champion community access to defibrillation as part of an ongoing commitment to save more lives. As a supporter of public access to defibrillation since 1998, the organizations vision is to have at least one person in every household trained in life saving first aid, CPR, and AED use.

"Because the Red Cross reaches over five million people annually with our first aid, CPR, and AED programs, we know that the availability of a properly trained person and an AED is key to providing the best care to a cardiac arrest victim until emergency medical personnel arrive.

"The Red Cross currently provides defibrillation information in all CPR courses and encourages the public to make defibrillation a part of their emergency preparedness plans at home, at school, at work, and at other public places. The inclusion of defibrillation in preparedness plans and greater access to AEDs can enhance preparedness efforts, help reduce the public's vulnerability and enable citizens to respond to cardiac emergencies.

"The American Red Cross mission is to help people prevent, prepare for, and respond to emergencies. We believe that removing barriers to public access to AEDs and training more people could result in more of the American public responding to an unexpected cardiac event. If the removal of this barrier results in even a five percent decrease in the number of lives lost each year, this positive step would result in approximately 25,000 lives saved annually.

"Please join the American Red Cross in helping citizens save more lives."

ACTING CHAIR LASKEY: Thanks Geretta.

Is there anyone else who wishes to address the panel today on the topic?

If not, then -- yes, sir.

MR. POLLEOTHICO: Good morning. My name is Frank Polleothico (phonetic). I'm a registered nurse. I'm the executive director of the AED Instructor Foundation. We are a 501(c)(3) nonprofit corporation funded manufactured AEDs and also funded by programs that we conduct.

I'm here to speak on behalf of the recommendation to remove prescription, and we fully support that. However, I want to emphasize the fact that AEDs, despite all of the wonderful things we've heard this morning, and I fully believe, do not save lives. AED programs save lives, and AEDs that are not instituted, AED programs that are not implemented and are not part of an on-site emergency preparedness plan, that involve training and leadership and guidance and some oversight, not needlessly bureaucratic, but within the context of an on-site program, be it in a home or a small business, are not going to work.

There were 15 million people trained in CPR in this country. Yet paramedics and EMTs report that less than five percent of the time that they respond to a cardiac arrest emergency is somebody doing CPR. We must not just give people AEDs. We must do it in the context of them being prepared to use them.

The growing number of horror stories of AEDs being on site and not utilized scares me. I just heard of another one last night. I probably hear about two a week.

So clearly AEDs are marvelous. I have personally used them. I've used them successfully, and I've used them where the patient didn't survive, and I know the benefit they provide. In my experience as an emergency nurse and as a paramedic, as the former Director of EMS for the City of New York, I can only speak to the missing link that AEDs help to fill in the wonderful system of emergency medical services that has been developed in this country in the last 30 years.

But AEDs must be operated in context, and while I think the prescription does nothing to help that, and as I agree with all of the speakers so far, I won't even reiterate that; the prescription does nothing to help the AED program issue.

However, guidance and control, involvement directly from the EMS system is essential or AEDs will not fulfill the promise of reducing the tens of thousands of needless premature deaths that occur in this country every year.

I thank you.

ACTING CHAIR LASKEY: Thank you.

All right. Last call. Does anybody else wish to come forth this morning?

If not, then I will close the open public hearing portion and would like to proceed with the sponsor's presentation, and if I could just have Geretta.

MS. WOOD: I would just like to remind the speakers to introduce yourself and state your relationship to the company and any other conflict of interest you might have.

MR. MORGAN: Can we have about 60 seconds to set up a demonstration?

MS. WOOD: Sure.

MR. MORGAN: All right. Good morning. I'm Carl Morgan, one of the founders of HeartScreen, which is now a part of Philips Medical Systems. I'm an employee of that organization.

As you've heard, we are here today because we propose to remove the prescription requirement for the Philips HeartStart home defibrillator. The prescription requirement reads as follows on our device: "caution. Federal law restricts this device to sale by or on the order of a physician."

Our organization was formed 12 years ago to prevent hundreds of thousands of unnecessary deaths due to sudden cardiac arrest. We believe that our focus during that entire 12 years has been towards providing small, easy to use, automatic external defibrillators specifically designed for people that do not have defibrillation in their job description, that is, AEDs that can be used by virtually anyone to help save a life in a moment of need.

A lot has gone on in this 12 years, but I think of note this morning is that for the last five years our organization has been in discussion with FDA on removing the prescription requirement from defibrillators.

In November 2002, we launched the HeartStart Home Defibrillator, which we believe is an idea design for prescription removal. We filed our 510(k) in 2004, and we're here today to present.

During our discussions with FDA, we found that the law requires that medical device labeling must bear adequate directions for use. FDA regulations further define that to mean directions under which the layman can use a device safely and for the purposes for which it's intended.

In addition, certain devices, including historic defibrillators, must bear this prescription caution, and this comes under the conditions if a device is not safe, except under the supervision of a practitioner licensed by law to direct the use of such a device, and because of that perceived lack of safety, adequate directions for use cannot be prepared.

This description suggests a basis for removing the prescription requirement for the HeartStart Home Defibrillator. That is, today we hope to demonstrate for you that the technology has an established history of safe use.

Further, we hope to demonstrate for you that the Heart Start Home Defibrillator can be used safely and for its intended purpose based upon its labeling alone.

Your presenters today include David Snyder, our Director of Research at Philips Medical Systems in Seattle; Dr. Lance Becker, a noted resuscitation researcher, Professor of Medicine and Director of the Emergency Resuscitation Center at the University of Chicago.

Dr. Becker will present the results of some of our studies and provide some perspective on the need for early defibrillation.

Dr. Jeremy Ruskin, a noted researcher in the management of cardiac arrhythmias. Dr. Ruskin is the founder and Director of the Cardiac Arrhythmia Service and Clinical Electrophysiology Laboratory at Massachusetts General Hospital. Dr. Ruskin will provide today some clinical perspective on the management of sudden death by providing early defibrillation capability.

I'd now like to introduce David Snyder, our Director of Research at Philips Medical Systems in Seattle.

DR. SNYDER: Thank you, Mr. Morgan.

Good morning, members of the panel, Food and Drug Administration, the public. It is, indeed, a pleasure to be speaking to you today.

I will be speaking to you twice. First I'll be introducing you to the product, and then later I'll be standing before the podium to present some study results to you.

With that I'd like to proceed to an introduction to the product. As you've seen a picture, this is the Philips HeartStart Home Defibrillator. It did receive clearance from the Food and Drug Administration in November of 2002 specifically for use in the home by lay responders.

The indications for use of this product are for application to an unresponsive or not breathing victim or not breathing normally victim, and I want to draw particular attention to the second indication, which is "if in doubt, apply pads." These indications were designed to prevent the necessity for adequate assessment skills on the part of the responder in case of a sudden cardiac arrest. It is not essential for proper use of this product for a person to be able to properly assess whether a patient is in cardiac arrest. Again, if in doubt, apply the device.

The safety and effectiveness of the technologies employed in this product have been established over a long history of AED products, and I'll give you some more background on that in a few minutes.

But at this point, I would like to do a demonstration for you. The operation of the device is very simple. Activate it by pulling the handle. Place the pads per voice instructions, and press the shock button.

I do want to say you just got an inadvertent demonstration of one of the product's features. The beeping was because we have practice pads installed in this device. After a certain amount of time with practice pads, which cannot deliver therapy, the device will start alerting you telling you that it's not ready for use. So it was not happy being not ready for use.

Okay. This is the HeartStart Home Defibrillator in its case. The first thing I want to draw your attention to is it does have a first aid reminder to activate EMS. We recognize that rapid defibrillation is only one element in the important chain of survival in order to assure survival from sudden cardiac arrest. So "call 911" is right on the front.

We also recommend placement in a visible place adjacent to a telephone so that can be done properly.

There's also a place as you put the device in service to add your own address information. One of the comments that was made this morning is correct. If you have an arrest in your home, it may not be one of the family members that is called upon to respond to the emergency and use the device. It may be a visitor to your home, and they not even know your address. So that information is right on the front with the EMS reminder card.

When you open the device, you're presented with a pair of scissors for cutting away clothing, a quick reference card that can be used, and the device itself. What I'm going to do now is walk through a mock cardiac arrest scenario with our manikin here so that you can understand and see how the voice prompts interact with the user.

So you begin, and again, these are practice pads. It's a safe device. It can't deliver therapy. You begin by activating the device.

(The following is a transcript of the recording played by the defibrillator while being demonstrated.)

DEFIBRILLATOR: Begin by removing all clothing from the patient's chest. Cut clothing if needed. When patient's chest is there, remove protective cover and take out white adhesive pads. Look carefully at the pictures on the white adhesive pads. Peel one pad from the yellow plastic liner. Place pad exactly as shown in the picture. Press firmly to patient's bare skin.

When the first pad is in place, look carefully at the picture on the second pad. Peel the second pad from the yellow plastic line. Place pad exactly as shown in the picture. Press firmly to -- no one should touch the patient.

Analyzing.

No one should touch the patient.

Analyzing.

Shock advised. Stay clear of patient. Press the flashing orange button now. Shock delivered. No one should touch the patient.

Analyzing.

Shock not advised. Be sure emergency medical services have been called. It is safe to touch the patient. Check airway. Check breathing. Check circulation. If needed, begin CPR. For help with CPR, press the flashing blue button.

Pinch nose, tilt head, and give two full breaths. Breathe. Breathe. Place the heel of one hand in the center of the check between the nipples. Place your other hand on top of the first. Push the chest down firmly two inches. Keep time with the beat.

Pinch nose, tilt head, and give to full breaths. Breathe. Breathe. Continue with compressions.

Pinch nose, tilt head --

(End of defibrillator audio demonstration.)

DR. SNYDER: And that's really all there is to it.

I would like to draw your attention to a few aspects of the scenario you just saw. First is that there was as second level of EMS reminder. The first level again is labeling right on the exterior of the device to activate EMS. Should for any reason that not happen, there is a second vocal reminder after first shocks are delivered. Make sure emergency medical services have been activated.

The second thing I want to mention is that once you complete your initial sequence of CPR, you are instructed to stop CPR, at which time a reanalysis of the heart is performed. If shocks are indicated, you're given the same guidance on delivering shocks. If not, you proceed directly to the opportunity to get CPR coaching again. If you're firm in your CPR skills, you can elect not to get the voice prompts and just proceed without the prompts.

These CPR prompts are intended to reinforce. They're not intended to teach CPR. We found that was impractical to do, but we also found that even people who have had regular CPR training don't remember the protocols well. They don't remember placement of hands. They don't remember depth of compression. They don't remember how many compressions, how many ventilations. So the voice coaching is really designed to reinforce those skills.

Another thing you may have noticed was the pacing of the prompts was very methodical. It was not rapid. We also found in user testing, which we'll be talking a little bit later about, that if you got ahead of certain responders, not everybody responds at the same rate, but if you got ahead of people and issued instruction before they had completed the previous task, they tended to completely fail. That is, if we went too fast, we could get a little bit faster for shock, but a certain percentage of people couldn't do it at all. So the pacing is very methodical to insure that the vast majority of people are able to complete these tasks successfully.

Now, should you be secure in your skills, you know what you're doing. You move quickly. The prompts will actually catch up with you. The device will detect where you are and the stage of applying these pads and delivering a shock and jump forward to catch up with you.

And I want to give you another demonstration right now so that you can see how that works. What I'd like you to pay attention to this time, you heard a lot of very detailed prompts. Now, this time we're going to go much more quickly, and you'll see that all of those detailed prompts do not appear in this scenario because I'm completing tasks.

I'm also going to apply the device to myself. This is now a live defibrillator, and if it detects V up (phonetic), it will deliver a shock. I'm doing this to demonstrate my confidence in the specificity of this product, and I'm going to apply it, lead two. It's not a good defibrillation vector, but it is a representative vector for the ECG that's observed by a defibrillator

DEFIBRILLATOR: Begin by removing all clothing from the patient's chest. Cut the -- place pad exactly as shown in the picture. Press pads firmly to patient's bare skin. No one should touch the patient.

DR. SNYDER: Okay. I can press this button as many times --

DEFIBRILLATOR: Analyzing.

DR. SNYDER: -- as I want. It's not going to deliver a shock.

DEFIBRILLATOR: No one should touch the patient

Shock not advised.

DR. SNYDER: I'll press the button again. No shocks.

DEFIBRILLATOR: -- emergency medical services --

DR. SNYDER: It simply will not do it.

DEFIBRILLATOR: -- have been called. It is safe to touch the patient.

DR. SNYDER: So that's a quick run-through of how the product works. You can see that for people not secure in their skills it gives them very detailed instructions. I'll talk about how we derived those instructions a little bit later in the presentation.

If you're secure in your skills from your AED training, you can proceed very rapidly.

So with that introduction, I'd like to now mention a concept which one of the introductory speakers this morning actually talked about, and that is the notion of a defibrillator as a piece of safety equipment as opposed to a piece of medical equipment prescribed for a particular patient at risk. And this product was specifically designed as a piece of safety equipment.

In fact, the labeling in our pre-sales materials on the outside of the retail box and in the owner's manual contains the statement in the first bullet, which is: "if you have concerns about your health or an existing medical condition, talk to your doctor. A defibrillator is not a replacement for seeking medical care."

Again, this device is designed as safety equipment. It's to address this problem of the large cohort of patients for whom symptoms have simply not presented, and they are not a part of the medical system at elevated risk for sudden cardiac arrest.

This is a product designed because we don't know who might need it or when. It's equipment that's intended to be used perhaps once in a lifetime. Best case is it will never be used, but perhaps once or twice in a lifetime, and because of this use model, we have identified some characteristics that are important to this kind of product.

First of all, it must be safe for any user. The second is it absolutely has to be ready to use when needed. Mostly it's going to sit on a shelf and gather dust, but the device has to be able to assure its readiness when the emergency occurs.

And finally, it has to be very easy to use in the moment. We're not talking about EMTs that do this several times a year. These are people, again, who may do this once in a lifetime.

Now, we're going to be presenting you some data on the reliability and safety history of these products, a point I need to establish right at the beginning, is that the HeartStart Home Defibrillator, which is shown in the upper right of this slide, has core technologies that are actually common to all of the defibrillators that Philips Medical Systems has produced.

The ForeRunner defibrillator was introduced in 1996. That was followed by the FR-2 in 2000, and the HeartStart Home Defibrillator in 2002.

The ECG analysis system you just saw demonstrated is common to all of these. There has been no change in what we call the life threatening arrythmia detector across this base. So any results derived from uses of the earlier devices are also applicable to the HeartStart Home.

You also heard mention in Dr. Tovar's presentation -- I think it was actually the Q&A -- about ECG validity measures, the ability to detect artifact within the ECG. That is, signals that are introduced mechanically or electrically that are not of cardiac origin, and this is a key aspect of safety in these products. It's essential that the device understand whether it's truly analyzing cardiac signal or artifacts, and we do have an ECG validity system that, again, is common across all three of these products and has been very effective, and with the advent of the HeartStart Home, we have actually reinforced this capability with a second way to determine whether artifact is present.

Now, this is not a feature that's common to all defibrillators. Some have similar capabilities and some have no capability in this regard.

The therapy that's delivered by this device is a 150 Joule impedance compensating biphasic waveform. Again, it's common across the entire product line.

For the FR-2, we introduced pediatric attenuation capability or pediatric treatment capability, the ability to deliver 50 Joules for a pediatric patient, and this technology has also been taken forward to the HeartStart Home Defibrillator.

And finally, key pieces of the user interface, the core interface, that is, pieces that measure things like pad connection, adequacy of pad connection, various pieces of safety prompting and the general protocol management technologies are common across all three of these products.

Now, there is a distinct difference between the earlier products and the HeartStart Home in that the earlier products in the bottom left hand of the slide had ECG displays as well as manual override capability should a trained medical professional have a disagreement with the advisory system.

These were deemed inappropriate for a lay market because of no training in ECG interpretation, and they were removed specifically because of the lay use model, and that feature has actually been replaced by enhanced prompting on proper pads placement which I also demonstrated for you. This was an area that we found that lay responders had a great deal of difficulty with.

And we have also added the CPR coaching, again, to reinforce skills that have already been obtained.

So the sensitivity and specificity of the ECG analysis system is due to a fairly sophisticated design. We actually take multiple looks at the ECG. the first is rapidity of the signal conduction. It's really a mathematical measure of the electrical health of the myocardium. We also look at the amplitude of the ECG; a measure that we call stability, which is the repeatability of the morphology of the ECG complexes.

Normal sinus rhythms and organized rhythms are highly repeatable. VF is very unrepeatable, and it's a strong predictor. We also look at heart rate. Now, the important point of this slide is that no single one of these indicators is capable of making the machine advise a shock. It actually takes the concurrence of all four of these measures before the machine will advise a shock.

I'd like to demonstrate the sensitivity and specificity of this algorithm by presenting a few studies that have been published. The first bullet was the post market study that Philips undertook with the introduction of the biphasic wave form into the marketplace. This was done with our first generation, AED, the ForeRunner, and we reported on the first 100 consecutive applications to VF patients. This involved a total cohort of 286 out of hospital patients. That is, 186 were in nontreatable rhythms.

And in this study, the authors reported 100 percent sensitivity to treatable rhythms and 100 percent specificity to rhythms that should not receive a shock.

The second study I would like to draw your attention to was published by American Airlines in the New England Journal of Medicine. They reported on, again, our first generation AED, the ForeRunner, which they had equipped their airplane fleet with and flight attendants to use. They reported on the first 200 consecutive uses of this product, 15 VF patients in all. These were applications both in flight and in terminal. In some cases the device was retrieved from the airplane to treat a sudden cardiac arrest in terminal.

In particular, I want to draw your attention to the specificity number here. Again, it's 100 percent specificity to these 185 applications in untreatable rhythms, 100 percent sensitivity to the VF patients.

This device was often used as a cardiac monitor in flight. If there was a physician in attendance that wanted to look at the cardiac rhythm, the airline did apply the device as a monitor so that the physician could watch the ECG screen.

Now, the authors didn't specifically report on the amount of time spent in doing cardiac monitoring, but our shock advisory system, our arrythmia detector, is active throughout monitoring. That is, we are continuously looking. So by a very conservative estimate we're guessing that at least tens of thousands of analyses were performed on these patients in nonlethal arrhythmias without a single false positive advisory.

Now, I do want to tell you that we know of one inappropriate shock that has been delivered by these devices. It was a case of successful defibrillation. The patient presented in a course VF. It was recognized. The device was charged. VF was terminated, and the resulting rhythm following shock was actually low amplitude atrial fibrillation with no ventricular activity. It was truly indistinguishable from ECG alone from a very fine VF. The device did analyze that, recommend and deliver a shock based on this atrial fibrillation with no ventricular activity, but there was no negative outcome. The patient survived neurologically intact.

Now, I want to present you with some use estimates, and I'm going to put a big caveat on this. In fact, I'm going to jump to the bottom bullet first. These projections are based on a non-random sampling based on ForeRunner and some FR-2 AEDs. As devices are returned to us for service, they have an internal memory that we're able to examine to see how many times the device has been used, how many shocks have been delivered, and so forth, and from these we have been able to extrapolate out to our installed base of over 150,000 AEDs since 1996, and actually this is over 170,000 today.

And what the data tells us is that probably greater than one million total applications to patients have been performed with this line of defibrillators, and of those, approximately 200,000 patients required shocks, and approximately 800,000 patients did not require shocks.

So when you consider that one inappropriate known shock, it has to be considered in the context of approximately a million patient applications.

I also want to report that we have seen six confirmed AED emergency use failures across this installed basis and over a million applications. Four of those had no patient impact. One, the patient impact was indeterminate, and in one case there was possible patient impact. This was actually an event that occurred subsequent to our filing in the 510(k) for this product, and it has been filed as an MDR.

During all of this experience since 1996, we have had no complaints about the effectiveness of the biphasic therapy.

Now, you heard a lot about MDRs, medical device reports, this morning. So we have summarized the top three causes of MDRs from this line of products, and I want to reinforce the fact that in spite of the large number, 7,000 MDRs, over 7,000 MDRs, over the time period that these products have been on the market, Forst & Sullivan 2003 estimated that Philips Medical Systems has about in excess of 40 percent of the U.S. AED market, and in spite of that 40 percent market share, we are responsible for fewer than one percent of the filed MDRs during that period.

This is a summary of our first top three causes of MDRs. The first is a failure of the voice prompting system due to a speaker failure. Basically one of the wires to the speaker can break, in which the device does not present voice prompts.

There was no patient involvement in any of these cases, and we filed 35 MDRs on this particular failure.

The next most frequent cause has been poor patient pads connection. We have filed 11 MDRs on this subject. Patient impact in these cases has been indeterminate, and we suspect pads damaged as the root cause of this problem, although we've been unable to confirm because pads are one of the first things that's discarded after execution of a code, and we have been unsuccessful by and large at retrieving these pads following failure.

But, again, you must put the 11 failures in the context of an estimated one million patient applications.

And the third most frequent cause of MDRs has to do with algorithm sensitivity. Patient impact in this case was indeterminate. These were basically very long down time, low amplitude, low rate VF, right on the borderline of being classified as an asystole. So patient impact is truly indeterminate in this case. The ECGs simply did not meet the rate criteria of our algorithm. It was not a product malfunction

Now, this slide illustrates steps we have taken in the HeartStart Home Defibrillator, which you are considering today. We use our MDRs as a learning experience, ways to improve our products and to address issues that we see through the MDR reporting system.

So I've taken these same three top issues from the ForeRunner in FR-2 experience and show you the actions that we have taken in the design of the HeartStart Home Defibrillator.

The first problem is no voice prompts. Now, that's mitigated in the ForeRunner and the FR-2 because there are instructions written on a screen. Because there is no screen here, we have been able to develop a speaker self-test. So as part of the self-test, which you heard a little bit about this morning; you'll hear more in the remainder of the presentation, once a week we turn the device on, and it's able to determine whether the speaker is functional or not.

With regard to the second issue, the poor patient pads connection, sometimes this is due to skin condition of the patient, but it may also be attributable to pads. So in the HeartStart Home Defibrillator, we've enclosed the pads in a rigid plastic container, but more importantly, we now have a pad self-test. Every 24 hours the device powers on. It checks for electrical continuity and presence of the pads, but it is also able to do an electrical determination of the condition of the electrogel, the adhesive aspect of the pads, to determine if any drying has happened.

And if the pads get to a state, before they get to a state actually, where they wouldn't be usable for defibrillation, we can alert the owner of the product that the pads are drying out. They need to be replaced. They call customer service and get information on how to replace the cartridge.

I want to move on now to reliability. We have made a great effort to improve the reliability of these products because basic design and fundamental product reliability is really essential to safe use and lack of failures, and for context, we've presented data on first year, all causes failure rates for three different products.

The first is code master manual defibrillator, which those of you in EMS and hospitals may be familiar with. It was introduced in 1991, and during the first year of service of that product we experienced a seven percent all causes failure rate.

Now, these aren't emergency use failures. These are failures from all causes.

With the introduction of the ForeRunner AED in 1996, we were able to improve that first year all causes failure rate by nearly an order of magnitude to 1.3 percent of devices shipped, and with the advent of the HS-1 -- and I want to mention what the HS-1 is for just a moment -- we're considering the HeartStart Home Defibrillator.

It also has a sister product which is identical. AED is the same product, but the labeling and accessories provided with it are really adapted to a commercial market to a corporate market.

So from a standpoint of the use of the product, reliability of the product, the functioning of the product, they're identical. So for some of this data, we've put statistics from both of these products together and they're identified as the HS-1.

So over that entire installed base, the first year annualized failure rate of the HS-1 class of products was again reduced by an order of magnitude to .04 percent all causes failure. That's one failure out of 2,500 devices shipped in the first year.

I want to go into some more detail now about the self-test. This is a diagram that shows you the testing that we do. It's done on daily, weekly, and monthly basis.

On a daily basis we power up the device. It checks for the readiness and condition of the pads to make sure the electrogel is suitable for use. It also checks the functioning of the two computers that are included in the device. It checks all of the circuitry, references. You no longer have to calibrate this device. It has internal references, and it checks for functionality of the therapy circuitry.

On a weekly basis we do all of those tests. Plus we add a full calibration of the ECG front end and a test of the audio system to make sure that the speaker is functional.

On a monthly basis, again, we do all of those. Plus we add a full high voltage charge of the defibrillator capacitor and discharge of 150 Joules into an internal load that's contained in the device. You do not have to attach an external load.

So on a monthly basis, absolutely everything in this box is tested, and we have a very comprehensive test that's performed on a daily basis.

Now, I mentioned our design process of learning from simulated use, and up here I've identified some problems. The way we do this is we take our best cut at what's a good product. What's going to interact well with the user?

And then we go out and we seek volunteers, and we have primarily sought untrained volunteers in environments such as shopping malls and senior citizens centers. And we'll provide them with this product and a manikin and ask them if they could try to save the manikin, and we watch how they interact with the device. We see where people succeed well. We see where they have difficulties, and where they have difficulties, we take the device design back. We come up with new ideas. We implement them. Then we go back out and we do it over again, and we continue with this iterative process until we've satisfied ourselves that all of the common issues are addressed.

Five of them are listed here. I'm only going to discuss two of them because they're of key importance. The first is that lay responders don't understand electricity or defibrillation, and a lot of the examples they see on television are not very appropriate for actual use. And we found that if people aren't given explicit instructions, they will do things that are really nonsensical from a defibrillation standpoint.

A common mistake is to put pads on top of clothing. As you saw in the demonstration, what we've done is two things. We've added emergency scissors, and we've added an explicit voice instruction to begin by removing all clothing from the patient's chest, cut clothing if needed.

Now, the scissors and the cut clothing as needed is another cue that time is of the essence here. You don't have to unbutton the shirt. Destroying clothing is fine. We give you permission to do it, and we give you the tool to do it with.

The second problem we identified that I want to highlight is poor pad positioning. There's a lot of research being published even today. There were two more just this month published.

On the problem of properly positioning the defibrillation pads not only among lay users, but among medical professionals, it's a very challenging things for people that don't understand what needs to be done.

Because of that, we've added explicit graphics on the pad. You saw those when I was holding the pads up. They show proper placement. But we also found beyond that, to get people to pay attention to the graphics, we had to add a voice prompt that said look carefully at the picture and place it exactly as shown.

By taking these kind of steps, we've been able to achieve a very high success rate in the ability of lay responders to apply this product to a patient and successfully deliver defibrillation shock, and you'll see some study data on that in just a few minutes.

Just a little bit more about the product that's being considered. Most of what you're going to see and hear about today is the device itself, what you saw on the table, what you saw me holding and demonstrating. There's actually much more to the total product than just that box.

In particular, I want to draw attention to the purchase aspect of the ownership life cycle for this product. We have specific information on the problem of sudden cardiac arrest, what a defibrillator does, et cetera, et cetera. It's contained in the product packaging itself. It's also available in pre-sales materials.

We maintain a product Web site that also has resources and links to resuscitation organizations and frequently asked questions.

We also have the customer service available that can provide product information, training resources, and in particular, if you purchase the product we offer grief counseling following use, and we also offer throughout this life cycle access to physicians should you have a question you would like to address to a physician.

Another aspect of the product that may not be obvious has to do with set-up and maintenance, and that is there are voice prompts in this device that were not demonstrated that help you actually set the device up. When you install the battery, it tells you it's not ready for use. Install the pads cartridge. If you remove the cartridge, it will tell you to reinsert it. It wants to be ready for use, and it's very unhappy if it's not ready for use, and it will start giving you instructions and warnings to make sure that it is ready for use.

This also is important in case of any failure that may be detected by the self-test. Again, the device will start chirping, and that blue information button that I used for CPR information will start flashing. If you press the button, it will give you information on what needs to be corrected. You can contact Philips customer service and get help in correcting the issue.

So that wraps up my introduction to the HeartStart Home Defibrillator. Again, it was designed specifically as safety equipment, really a paradigm shift in the way we're using these products.

The technology is using the device that has a history of safety and readiness, and it was specifically designed for ease of use in the hands of the lay responder.

With that I'd like to introduce Dr. Lance Becker, Professor of Medicine, University of Chicago.

DR. BECKER: Thank you very much.

My name is Lance Becker. I'd like to thank the panel and the public for coming. This is quite a honor to be here.

I'd like to disclose that I'm a paid consultant to Philips this morning. In addition, I have no equity or stock in that company nor any of the other manufacturers.

I have been a consultant to several of the other defibrillator manufacturers over the years. I have intellectual property, some patents in resuscitation that involve cooling induction, and I compete for grants for the University of Chicago.

In my real life, I'm an emergency medicine doctor on the South Side of Chicago, and I have studied cardiac resuscitation for many years. It's my pleasure to do a clinical overview of the problem of cardiac arrest and our therapies, although had I known that Dr. Eisenberg was going to be here beforehand, since he's the world's expert on that, he could have probably done it for me.

I will then go over some of our safety and usability data in our simulated study.

In the United States, we have about two million people a year that die. Those are the number of deaths each year in our country, and most experts estimate that about half of those deaths, about a million deaths a year occur from some form of cardiovascular disease.

You've heard a lot of different numbers on how many sudden cardiac arrests there are, and most experts agree that it's somewhere in the vicinity of a quarter of a million deaths per year from cardiac arrest.

Now, while there's some dispute on what the actual numbers are, no one disputes that this is not a major public health concern, and I think none of us in the room would have to go too far before we could think of an individual who we know who has died from sudden cardiac arrest.

Now, I want to make two points about the epidemiology of this disease. The first is that because 80 percent of these occur in the home and 50 percent of these are witnessed in the home, the point I want to make is we do not right now have a national idea, a strategy for how to deal with home cardiac arrest, other than call for the EMS system, and we know that the survival rate in the home is far worse than the survival rate in public places.

So lacking a home strategy, I just want to suggest how important the potential of a home device may actually be.

The second point that I want to make is that according to the American Heart Association and many other investigators, the majority of victims of sudden cardiac arrest have no prior symptoms. That is to say they are not high risk patients who have been identified as having had a mild cardioinfarction or anything.

They are individuals who are first getting their wake-up call to cardiac disease when they sudden collapse and die. And it's for this reason and for that segment of our population that the notion of moving from a prescription to a public access to public availability is so important.

Now, a few words about how we treat cardiac arrest. This is from the American Heart Association, and it says that the way that we try to save victims of cardiac arrest is that we want them to receive early access, early CPR, early defibrillation, and then early advanced care.

And note that the American Heart Association suggests a goal of receiving defibrillation in less than five minutes, and I'd like to spend just a few more minutes on why time is so very critical during cardiac arrest.

You can see here in the shaded area the mortality curve that we see with each passing minute of cardiac arrest, and you can see on this axis is the probability of survival. Minutes go across the bottom, and with each passing moment somewhere between seven, ten percent of the potential individuals who could be rescued are lost.

If we just start to think, well, how long does it take to defibrillate someone, we learn some very important information. We know that in most arrests, it takes somewhere around four minutes before the collapse is recognized and EMS is called and an ambulance is dispatched.

Well, time unfortunately does not stop for that. It then takes a certain amount of time for the ambulance to get to the address. In very good systems, six minutes is considered a very good response for the EMS system.

Time continues to pass. It takes a certain amount of time to get to the patient, to apply the pads, and so in very good locations, significant numbers of our current survivors are defibrillated at approximately 12 minutes after their collapse, and what we know is that that is a very prolonged time, and it's not surprising that most studies have survival rates in the vicinity of five percent or less.

Now, the notion has been for many years that we want to move victims of cardiac arrest up that curve so that defibrillation can be done at a much earlier point, and you can see that in theory as you move up that curve, you'll have much higher rates of survival.

And what I want to do now is share with you some of the data that lets us know that that is, indeed, true, that one can, indeed, move up that curve.

These are three studies all published in the New England Journal of Medicine that I'd like to just talk about for a minute. The first is the American Airlines program. In that study, you'll note that they had 15 patients up in the air with ventricular fibrillation, a survivor rate of 40 percent, and no adverse outcomes.

In our own study in Chicago O'Hare and Midway Airports, we basically placed AEDs about a minute walking through the terminals of O'Hare and many airports have followed suit. The result that we had was a nearly 61 percent survival rate, 11 of 18. I might add that ten of those individuals were alive a year later.

Many of these defibrillations were done by lay people. Half were done by people simply walking through the terminal. There were no adverse events. In the Las Vegas study, in the casinos, you'll see again a 60 percent survival to discharge, and again, no adverse events.

Now, I want to point out that the airline study and the casino study are sort of individuals using the AED who have a duty to use the AED. It's part of their job. In the O'Hare experience, ours was more mixed, particularly with people simply walking through the terminal, but the study that I think has really given us a great deal of information is the recent public access defibrillation trial.

I have to highlight that the principal leader of that study, Dr. Joe Ornato, is here on the panel and can probably speak to it in more detail than I will, but the basic question here was would AEDs plus CPR improve survival compared to CPR alone, and this was a large NHLBI sponsored study, 20,000 lay responders trained, 23 cities, 1,000 sites, and the results were that in the arm that had the AED, survival was doubled compared to CPR alone.

Importantly, there were no serious adverse events associated with AED use, and there were no recorded instances of EMS not being called or a failure, a failure of EMS to be notified.

Individuals retained their skills during this study, and that was measured, and the conclusions really were -- and I think it's very important for this panel -- that layperson can use AED safely to provide early defibrillation.

So if we try to summarize those, it's clear that rapid response is what saves lives, that the time is critical, and that early defibrillation is highly effective and we want to do it as early as possible, and so a primary question for this panel is: can the HeartStart Home Defibrillator be used safely by lay people.

And I'd now like to share the simulated data that we generated in Chicago in an attempt to provide some information on that question.

Our study had two primary hypotheses. The first was that the HeartStart Home Defibrillator and the FR-2 are safe and safe even in the absence of training. We were designing essentially a worse case scenario, someone who had never seen the device and who did no training whatsoever.

Our second hypothesis was the HeartStart Home and the FR-2 would have high usability when used with primary labeling components, including the voice prompts that you've heard David demonstrate and watching the trainee video.

The methodology that we used is a mock cardiac arrest scenario with a fully clothed manikin and an AED, and the details of this are in your packet. It's much like the demonstration actually that David performed up here earlier.

I'd like to share with you the enrollment. You can see that for the two devices, there were both naive and video trained individuals. Of note, for the FR-2 here, that videotape was an eight minute videotape that those individuals watched, and then all volunteers went to a simulated use.

For the HeartStart Home, a three minute videotape was watched by the volunteers, and they then went to the simulated test.

Our primary endpoints, first they were safety and successful use, and I want to define those. Safety meant that there was no touching of a patient in a manner that it could result in a shock across the rescuer's chest.

Successful use meant that they had to complete a whole series of adequate deliverables, and those including power the unit on, attaching the pads with the appropriate placement, allowing the device to analyze, if necessary plugging in the pads, and then defibrillating within five minutes.

Our secondary endpoints that I'll be sharing include the time that it took from when the volunteers stepped into the room, from when pads were placed successfully, and how long it took for them to then deliver the shock.

Results of the study, for the FR-2 are on this slide, and I first want to highlight that you'll note that in terms of safety, both the naive users and the video trained users were completely safe, and as you'd expect, when we look at successful use, you can see here that about 50 percent of the naive users were able to actually operate the FR-2 without any instructions or any training by listening to voice prompts.

After watching the videotape, you'll notice that this improves to 86 percent, and this was not at all surprising to me.

If we look at the data on pad attachment, we can see that, again, there's a beneficial effect to watching the videotape with the pads being attached faster and the shock delivered in a more timely fashion.

Now, as we look at the HS-1, the Home Start, we see that again the safety was complete for both devices, rather, for both naive users and video trained users, and what I want to highlight in this successful use is that we see what to me appear a difference in the naive users. Eighty-seven percent of the naive users, that is, individuals who had never seen the device before, were able to successfully use this device in the five-minute period. When they watched the videotape it only increased to 89 percent, which is not statistically different.

And I just want to add that it was this observation that made me first aware that not all of the devices would function the same in terms of human use, and it was for this reason that it is the HeartStart Home that is being brought forward today as an acceptable over-the-counter device.

When we look at the effect of the video on improving pad placement and the time to shock, again we see that while there was no difference in successful use, there indeed was some improvement in the amount of time that it takes, and of course, it is everyone's intention and suggestion that we hope that individuals who use these devices will be totally trained and fully trained with a formal course.

We, indeed, got these results under what we would consider to be fairly adverse conditions.

Now, like all studies, there are limitations to this, and I want to be forthright about those. The simulated use cannot simulate everything in a real use. The chaos and noise and uncontrolled setting of someone falling down and dying in front of you, and I might add that you can't take volunteers and subject them to that. It is not ethical.

So the ability to do a simulated test where you would actually have a real person with a naive -- you know, it would just be totally unethical. So there are limitations to simulations.

Our demographics were those that were due to convenience. We picked to very different places, an urban emergency department and then a parochial school after school program, and those were the people who we studied. And of course, another limitation was that real human anatomy is far more varied than one will find on a manikin, and that's another limitation.

But what I want to tell you all is that despite these limitations, this was and is the state of the art of doing simulation studies, and I think there are some important things that we can learn from this.

The first is that in terms of safety we can see that both devices were used safely in all cases. We note the significant improvement that took place with the FR-2 when individuals watched that videotape. They improve significantly in their ability to successfully use the device.

But importantly, the HeartStart Home Defibrillator was successfully used by both naive and video trained volunteers at a rate of 87 and 89 percent, very high. And I want to try to draw one conclusion from this, which is that most of our data right now on the effectiveness of defibrillators is based on the FR-2 device, and as we see, the superior human characteristics, the use characteristics in the HeartStart Home, it seems to me that we have very good data to suggest that the experience with the HeartStart Home will be certainly as good and very likely better than our experience has been to date with the FR-2, which has been highly successful in previous studies.

I thank you, and would now like to have David continue with the labeling evaluation and the other surveys.

DR. SNYDER: Thank you, Dr. Becker.

There are three studies that I would like to present to you, actually two studies and a continuation of an existing study. The first is a labeling evaluation and simulated use.

We did undertake some validation studies to gauge the ability of people to read and comprehend additional labeling components that were developed for the product, and then we continued on to simulated use very much like Dr. Becker just described.

We also have conducted a survey of likely lay users of our earlier products, the ForeRunner and the FR-2. Now, it's important to understand that these products from the very beginning have been sold to lay operators and even into homes. There's been no prohibition on that, and physicians have certainly prescribed them for that application.

So we made an attempt to contact as many of those people as possible and find out what issues they may or may not have had with the use of their product.

And finally I want to describe a post market study which was already underway. It was launched with the FDA clearance of the HeartStart Home Defibrillator for home use to look for specific issues that may come up in home use.

To begin with the labeling evaluation, the purpose of the study was to test the comprehension ability of the public basically to read these additional labeling materials, comprehend them. We gave them a written examination to see how well they understood the materials.

These materials specifically are the owner's manual, the quick reference card which you saw in the case of the device when I demonstrated it, a training video which ships with the product and the quick start poster which is a guide to assembling the device and putting it into service.

In addition, for two of those pieces we proceeded on to a simulated use in order to demonstrate safe and successful use after reviewing only one aspect of this additional labeling, and these are the pieces of labeling that are actually included in the defibrillation case and would be accessible in the event of an emergency use.

I want to show you what these look like. the owner's manual is right here. It has the same form factor as the quick reference card that you saw in the demonstration, and it's inserted in a pocket behind the quick reference card and is actually not visible when you first open the case.

This is the training video, standard VHS, and the quick start guide, and as you can see in some of the graphics, it demonstrates how to open the battery in the pads' cartridge and assemble them into the device.

A methodology involved recruitment of volunteers in three geographically diverse shopping malls. We sought people with no medical or defibrillator training and having received no CPR training within the prior two years.

This was a precaution against, again, accidentally enrolling people that had been exposed to AEDs because so many CPR courses within the last two years had begun teaching AED skills as well.

We saw a broad age range in this enrollment and were successful in recruiting volunteers over the range of 21 years to 74 years of age. Forty percent of our volunteers were in the 45 to 65 age group, and 20 percent approximately were in the 66 to 74 percent age group -- excuse me -- 66 to 74 year age group.

The hypothesis for the comprehension test -- this was a little tricky. So I want to be careful with it -- was that the labeling materials were well understood, and what we sought to show was that 90 percent of these people could achieve a passing grade. If you refer back to your grade school experience, a passing grade was 70 percent of questions answered correctly, or a C, and you'll see that in the slides. So we wanted 90 percent of these people to be able to get a passing grade on the written exam, and we powered the study to establish a lower confidence limit on those results of 80 percent, given the presumption of a 90 percent passing grade.

The hypothesis for simulated use was that the HeartStart Home Defibrillator is safe and that the HeartStart Home Defibrillator can be successfully used by lay persons to deliver defibrillation shock, and this was powered for a noninferiority test against a presumed success rate of 90 percent, with a ten percent detection margin.

I want to say that this is an arbitrary goal. There are no gold standards for either of these tests, either the reading comprehension or performance and simulated use. We tried to set a high bar, and I'll point to the public access defibrillation trial. They recorded in abstract that three months following initial training approximately 89 percent of the people still had adequate AED skills.

So we used that as kind of our benchmark for what we wanted people to be able to achieve in the simulated use.

This diagrams the enrollment randomization for the study. the study was first stratified according to how much time the volunteer had to spend with us. If they had 15 minutes available, they were randomized into the reading comprehension only arms and either viewed the training video or reviewed the quick start poster and then took a written comprehension exam.

If they had 30 minutes to spend with us, they were stratified into a randomization between either owner's manual or quick reference. Again, the materials that are in the case, should you use the device in an emergency, they took the written comprehension test and then proceeded to a simulated use test.

Primary endpoints were essentially the same as those reported by Dr. Becker: safe, meaning no touching of the patient in a manner that could result in a shock delivered across the rescuer's chest; and successful, meaning shock delivered with pads positioned in a manner likely to defibrillate, and the activities necessary to do that were turning the device on, deploying the pads in proper position, allowing the device to analyze without interruption, and delivering the defibrillation shock.

Secondary endpoints as for the Chicago study were time to pads on and time to shock.

The written comprehension test, first of all, the labeling materials themselves are written to a sixth grade reading level or lower according to a Flesch-Kincaid score. Topics for the test included what is the definition of a sudden cardiac arrest. Is it different from a myocardial infarction, stroke, so forth?

Questions regarding set-up of the device, the importance of training, how is the device properly stored; rescue steps; post shock care and so forth.

These are the results of the reading comprehension test. A C grade is shown here. This was our goal. We wanted 90 percent of people to get a C grade or better, and what you'll see is the median scores on all four of these pieces of material were actually in the 90 percents.

Now, this slide does contain some information which has been provided to the FDA, but not reviewed, and I want to draw attention to that. We have shown interquartile ranges, the 25 and 75 percentiles. That was not originally submitted to the FDA in our 510(k), and we've also shown the 90 percentiles which are these -- excuse me -- ten percentiles, which are these dots.

And you can see that in all cases 90 percent of the people achieved a C grade or better, and in fact, the medians were in the A grade range.

These are the results of the simulated use test. Our lower confidence limit goal is shown here, the 80 percent. And in the case of the owner's manual we did not meet our predefined goal. You can see that the lower confidence limit actually extends somewhat below the 80 percent lower confidence limit goal.

In the case of the quick reference card, however, we actually achieved an observed successful use rate of 97 percent among these volunteers with a lower confidence limit that easily exceeded and a highly significant non-inferiority result.

Median time to show is similar as that reported in the Chicago study as 104 seconds following owner's manual, 89 seconds following quick reference. This is from entry to room now. This is not once you've turned the device on. This is entry to room to deliver of shock.

Limitations to the study are clearly the same as associated with the simulated use study performed at the university of Chicago. Given those limitations we conclude from the study that all of the labeling was well understood. At least 90 percent received a passing grade, 70 percent correct or better, and the lower confidence limit on passing grades was actually greater than or equal to 88 percent in all cases.

The defibrillator was used safely in all cases. That was 178 mock arrests, and successful use was achieved by 97 percent of the individuals with a lower confidence limit of 92, following review of the quick reference card, which is the recommended labeling for emergency use.

Now, we did take some follow-up actions as a result of this study. I want to show them to you here. First of all, we added information to the training video and the quick start poster regarding the intended use of the various labeling materials, and in particular, we modified the cover of the owner's manual to clarify its purpose as a guide to set up maintenance and accessories as opposed to an emergency use piece of labeling.

I'd now like to briefly describe the results of the lay user survey I mentioned. The purpose of this study was to determine if lay use of Philips AEDs resulted in any previously unreported or not understood problems.

The method was to establish a contact list of all people who had owned an AED at least a year and in an environment in which a lay use may have occurred. We specifically excluded all medical professionals.

We were able to identify 145 homes and almost 2,700 businesses and public facilities where a lay use may have occurred. We then attempted contact via a phone center and seven attempts were made for each person or installation that had been identified.

A brief interview was conducted, and in particular, we asked if the defibrillator had been used by a layperson. If the answer to that question was yes, we requested a detailed interview between a medical professional and the lay user of that device.

The results of that survey were that we were able to contact 78 homes and 1,645 businesses. No problems were reported. Two hundred and nine of the businesses or 13 percent had used their AED at least once to respond to a suspected cardiac arrest. This doesn't mean that the paths were necessarily attached to the patient or that shocks were delivered, but a patient collapsed. It was suspected to be cardiac arrest and a person responded.

We also identified nine uses in home or home offices. We were able to conduct 11 detailed interviews regarding incidents in which paths had been applied to unresponsive patients by lay responders. EMS was called in every case. Three patients appropriately received no shocks. They were not in VF. Eight patients did receive shocks, of which six survived a hospital admission and four received shocks solely by lay responders, and all four of those survived to hospital admission.

Now, these are anecdotal data. I don't want to draw too much attention. The real purpose of the study was the first sub-bullet up here, which was that no problems with use of the product were reported.

In conclusion, a limitation to the study are obviously that it was a survey interview. Participation is voluntary and bias is, therefore, possible, but we conclude from what we observe that no harm or injury to users by standards or patients was reported. No malfunctions or problems were reported. All users expressed willingness to use a defibrillator again should they suspect a sudden cardiac arrest, and finally no safety or effectiveness issues were reported.

To wrap up, I'd like to talk a little bit about the post market study that's currently underway for the HeartStart Home Defibrillator, and we do propose an extension to that study should over-the-counter clearance be granted for this product.

The purpose of the study is to evaluate lay uses of the HeartStart Home Defibrillator for safe and appropriate application. The methodology is similar to the phone survey you just heard about. Philips is generating contact lists following one year of ownership for using the one year in order to allow some opportunity for an event to have occurred. There's not point in calling people one month after they purchased their defibrillator because it's very unlikely it will have been used.

We're also enrolling people in the study should they contact Philips for accessory replacement. Certainly if a customer calls us for a new set of pads there's a possibility that those pads had been used in a rescue. Through both of these avenues, we then asked was the AED used. If the answer is yes, similar to the other survey, we request a detailed interview with a medical professional.

We propose to extend the study by an additional four years should we be granted OTC clearance for the HeartStart Home Defibrillator, that is, we will continue this process for four years following receipt of the clearance or terminate it should we record 200 home uses of the HeartStart Home Defibrillator, results to be reviewed by a data safety and monitoring board and reported annually to the Food and Drug Administration.

I'd now like to introduce Dr. Jeremy Ruskin, Director of Cardiac Arrythmia Service, Massachusetts General Hospital.

DR. RUSKIN: Thank you.

Good morning. It's a privilege to participate in this important discussion. I'm Jeremy Ruskin. I'm here as a paid consultant to Philips Medical. I own no equity.

I also consult for Medtronic on implantable arrhythmia control devices, and I've been asked to make a few comments about the HeartStart AED and the problem of sudden cardiac arrest from a clinical perspective.

Just by way of introduction, I want to underscore what you heard from Dr. Becker, but from a slightly different perspective. When I began my academic career in the 1970s, my colleagues and I focused on evaluating the electrophysiologic and anatomic substrate of survivors of out of hospital cardiac arrest.

And we understood from risk stratification schema that patients who were at highest risk for cardiac arrest were those with prior myocardial infarction, left ventricular dysfunction, heart failure, and spontaneous ventricular arrhythmias.

What we learned from our own observations and also from those of Drs. Eisenberg and Cobb and his colleagues in Seattle, that in fact the profile of survivors was somewhat different from that of the high risk subset that we just described, and that is that prior infarction was present in only about 50 percent of survivors. More than a third had normal or near normal left ventricular function. Fifty percent had no clinical or prior historical evidence of congestive heart failure, and about half had no spontaneous arrhythmias.

And the question then is what explains this seeming paradox, and it is explained by what you've heard from Dr. Becker, and that is that many cardiac arrest survivors, in fact, are drawn from an undetected high risk patient pool whoa re not definable prior to the event. These are patients in whom sudden cardiac arrest is the first manifestation of their underlying heart disease, and this represents a failure of our risk stratification schema.

This subset is now at least beginning to be addressed from a primary prevention standpoint with implantable cardioverter defibrillators. But from an epidemiologic and preventive medicine standpoint, we have no approach to this subset at the present time, and it is this group that is targeted by more widespread availability of AEDs.

This slide depicts for you the dissemination of defibrillator technology from physician delivered, hospital based defibrillation in the 1960s to emergency medical systems in the 1970s and '80s, to public access defibrillation in the 1990s, and most recently to the concept of defibrillation as a lay procedure in the early 2000s.

With the evolution of this technology and thinking, it is perhaps reasonable to think about the device under discussion today as a piece of time critical safety equipment, much in the way that we think of air bags fire extinguishers, smoke alarms, and seat belts.

All safety equipment has benefits and limitations. We know that smoke alarms reduce the risk of dying in a fire by as much as 40 percent, but that, in fact, alarms may be deactivated because of nuisance factors.

Seatbelts save approximately 11,000 lives per year and reduce the risk of death in car accidents by as much as 45 percent, but we also know that they are not used as much as a third of the time.

Air bags reduce the risk of a fatal event in a head-on collision by as much as 30 percent and have saved large numbers of individuals since their widespread implementation.

But we also know that they may cause injury or even death in children and small adults.

This slide depicts for you the absolute numbers of deaths attributable to motor vehicle injuries in the green line and sudden cardiac arrest in the red line for the year 1999 as a function of age group. And as you can see, even in the mid-'40s, the risk or the numbers that toll from sudden cardiac arrest begin to exceed those from accidents, and in fact, you see the curve rise very steeply into the mid-'60s. Still a relatively young group of individuals, many of them not known to be at risk, and experiencing cardiac arrest as the first manifestation of their underlying heart disease.

This slide depicts the ten-year probability of experiencing any number of emergency events, including reportable fires, air bag deployment, sudden cardiac arrest in all households, and sudden cardiac arrest in households over the age of 45, and you can see that the ten-year probability of a household over the age of 45, experiencing a sudden cardiac arrest is three times that of a reportable fire, and substantially higher than that of air bag deployment in a motor vehicle.

The benefits of removing the prescription requirement would include broader access to a safe and effective technology that constitutes the only definitive treatment for a sudden cardiac arrest and thereby provide an opportunity to save some of the lives that would otherwise be lost to this public health problem.

The question that arises in proposing this kind of a change in thinking and change in practice is whether or not the AED can, in fact, cause harm. In addressing this question, it's important to underscore that the HeartStart AED is intended for the same user and patient population as the currently cleared prescription device.

And as you've also heard, this device is designed with safety first and foremost in mind, and some of the critical safety features include an extremely sensitive and specific ECG analysis algorithm, an extremely robust artifact detection algorithm, and the absence of a manual override to preclude the delivery of inappropriate or incorrect shocks.

The safety of this device of the Philips AED line is supported by the use experience. These are estimates of use. You've heard of these from David Snyder, including an estimated more than one million patient applications, with an estimated approximately 200,000 individuals requiring shocks, with only one known inappropriate shock in this vast experience and no complaints about shock effectiveness.

This overriding commitment to safety is verified by an established history of safe use and the demonstrated fact that the HeartStart AED can be used safely for its intended purpose based upon its labeling alone by lay responders.

A number of theoretical risks have also been raised with regard to the OTC AED concept. One is whether or not an OTC AED would interfere with appropriate medical care.

Clearly, an OTD AED is not a substitute for any kind of medical care. Risk factors continue to need to be addressed. Care and prescribed therapies for preexisting conditions need to continue. Physicians will retain the option to prescribe AEDs in cases of medical necessity. And finally, the target populations, particularly those in comparison with those who require ICDs, are entirely different, as Dr. Becker emphasized and I did earlier.

Another issue that's been raised is whether or not the OTC defibrillator will interfere with the EMS response. We know that the sudden cardiac arrest survival rate is abysmally low, less than five percent, precisely because defibrillators logistically cannot be delivered in time to victims in a vast majority of instances.

Philips supports calling EMS in its labeling, as you've heard, and in verbal prompts and recognizes that early defibrillation is only one part, but a critical part of the emergency response.

Finally there's no evidence that a prescription requirement would in any way enhance an EMS response.

In approaching an issue of an OTC home defibrillator, it's important to consider this with realistic expectations. Sudden cardiac arrest is an epidemic and major public health problem of enormous proportion. It is the most common cause of death in adults.

We know also, as you've heard innumerable times this morning, that current survival rates are unacceptably low, in the range of five percent.

It's also important to recognize that an AED is not considered by anyone to be a cure for the problem of sudden cardiac arrest. As many as 40 percent of arrests are likely to be unwitnessed. Devices will in some instances be used incorrectly, and there will be some device failures.

There will also be human and logistical factors that will interfere with effective use.

Those limitations notwithstanding, OTC defibrillators represent a paradigm shift and a step towards wider access that will provide the potential to save some lives that will otherwise be lost. It's important to consider that long term with widespread dissemination, even a small impact, perhaps salvage of as few as five in 100 cardiac arrests would double current survival rates.

Ultimately the prevention of sudden cardiac arrest lies in the prevention of coronary artery disease, but until this goal is achieved, current strategies for addressing this public health scourge include risk factor modification by life style changes and appropriate medications, revascularization and implantable defibrillators in the highest risk patient subsets, and widespread availability of on site rapid defibrillation.

Defibrillation is one of the most important therapeutic advances in the history of medicine. Over the last five decades we've come to take this therapy for granted in a medical environment. We now have a technology that is designed specifically and cleared for use by lay responders.

Removal of the prescription requirement will provide broader access to a safe and usable technology and thereby provide the opportunity to salvage some lives that in the absence of prompt defibrillation will otherwise be lost.

Thank you.

DR. SNYDER: And that is the completion of our presentation.

ACTING CHAIR LASKEY: Great. Thank you.

I would just ask the panel members if there are any queries at this point.

Dr. Kato.

DR. KATO: One other reason that the sponsor is going to eliminate the or desire to eliminate the prescription requirement is to broaden the availability of these devices to the public. Do you have an estimate for what you think the annual growth would be in terms of AED deployment sine you're already seeing a 20 percent growth per year? At least that's what you presented.

You know, on a cursory review last night, it seems that you can get these things over the Internet now.

DR. SNYDER: It's a good question, and we don't have specific data on any anticipations of growth rate, given removal of the prescription requirement.

We have, however, in conversations with people who have attempted to purchase home defibrillators via our phone contact system, we've discussed with them the problems that they've had in obtaining prescriptions, and we do have data on that, and it has not been submitted as part of our 510(k).

If you would like to view some of that, I could show you the data.

I would like to also mention that the internet sites that you're referring to are actually doing it under prescription. It's not a methodology that we really endorse because no benefit is being provided by that prescription. If an individual contacts Philips to purchase a home defibrillator, we actually walk them through the process of learning about the product, and we do ask them to visit their own physician, set up an appointment, and we perform follow-up after that appointment to see if they were able to get a prescription or not, and again, we do have the data for those that were refused prescriptions to tell you the reasons that they were refused.

And I'd be happy to share that with the permission of the Executive Secretary if the panel would like to see it.

DR. KATO: So currently you're involved in a fairly elaborate system of education of these patients as they obtain these AEDs from your company, right?

DR. SNYDER: Correct.

DR. KATO: So what is the advantage of just being able to purchase one of these at Save-on (phonetic) without that educational benefit?

DR. SNYDER: Well, first of all, over the counter doesn't necessarily imply Save-on, but we would continue to sell the product through the channels we're using now. It's certainly an eventuality, but the educational materials provided with the product do educate the purchaser about the problem and the possibilities.

In addition, if someone is to purchase a HeartStart over-the-counter defibrillator, they do have a 30-day, fully money back, no questions asked guarantee.

So even if they purchase it without studying, after reviewing the materials if they decide it wasn't a purchase they desire to make, they can get a full refund on that.

But the issue we're really attempting to address is the fact that there is very significant fallout in people how have made committed decisions to purchase a defibrillator when they contact their physician and attempt to get a description, a large number are denied those prescriptions.

DR. KATO: So you have data that physicians are actually denying these prescriptions now?

DR. SNYDER: That's correct, yes.

DR. KATO: What happens if this device is used at a pool or at the beach?

DR. SNYDER: It's an excellent question, and it's a problem we have had concerns about. We actually performed a study on this, which was published last year. What we found was that even defibrillation and wet environments on wet surfaces was entirely safe to the responder. We actually instrumented a surrogate. It was actually a turkey, and we attached defibrillation pads and set out a grid so that we could do electrical measurements and only very small voltages were present.

And I can tell you that I personally have defibrillated a patient in the pouring rain with no adverse consequences. It's actually a very safe thing to do.

DR. KATO: Was the turkey alive or dead? Is there a difference with --

(Laughter.)

DR. KATO: Well, I mean, is there a difference in resistances?

DR. SNYDER: No. The turkey was dead, but it was selected to present a reasonable approximation of a human impedance.

ACTING CHAIR LASKEY: It was fresh killed, Norm.

DR. SNYDER: It was one of the more embarrassing studies we've ever published.

DR. RINGEL: I have a number of questions regarding the pediatric usage.

DR. SNYDER: Yes.

DR. RINGEL: I would like to point out that even though you have suggested that this will be labeled for pediatric usage, none of the speakers talked at all about any of the pediatric issues, the algorithms, the pads or anything like that. So you'll have to bear with me because there are a lot of questions to follow.

We can start off with the question as to why you are suggesting the system with two sets of pads. I think it has potential for a great confusion, and you don't explain how these pads would interact. For instance, I don't think your tests tested removing an adult cassette, put it in a pediatric cassette and then using the pediatric cassette. I'm not sure you give guidelines to tell people how to figure out if someone is 55 pounds or less.

I'm confused by one of the statements in your brochure when you say, "Note. If you're not sure about the child's exact age or weight or if infant child pads are not available, do not delay treatment. Use the adult pads, but place them on the child's chest and back as shown in using the HeartStart."

So if that's the case and you can use the adult pad effectively, why not just have one set of pads and avoid the confusion?

I also have concerns about the fact that most small child and infant arrests are respiratory arrests. They are not fibrillation or cardiac arrests, and there's no mention about or study about how this delay in fiddling or fumbling with the AED might delay appropriate respiratory care.

There are more, but you can start at any time.

DR. SNYDER: All right. I'll attempt to answer all of them in order. If I miss one, please let me know at the end.

DR. RINGEL: Sure.

DR. SNYDER: First of all, with regard to cartridge exchange, we did perform a design validation on people's ability to do this.

Slide up, please.

We performed a design validation with ten volunteers, and I want to admit right up front that clearly this aspect of the product has not been validated to the same standards as the primary use of the product, which is adult defibrillation.

But we have performed validations. We sought ten volunteers. These were people who had previous adult AED training, but no training on this particular product. So they had never seen this product before. Nor had they received any prior training in pediatrics.

The sample size was based on an FDA guidance document called "Do It by Design" from '96.

All volunteer viewed a training video, which included a pediatric chapter on pediatric application of the device prior to the test. The scenario was that upon entering the room they were presented with a toddler manikin, and an AED with an adult cartridge installed. A pediatric cartridge was available in the AED case.

Next slide, please.

The results of the validation are tha t all ten of the volunteers were able to recognize the need for exchanging the cartridge. Ten out of ten Retreat the pads cartridge or knew where it was located in the case and seven out of ten of those were able to exchange the cartridge, remove the adult cartridge and insert the pediatric cartridge with an average time to do so of 40 seconds.

And of the seven that successfully exchanged the cartridge, all seven were able to deliver defibrillation shock with adequate pads placement.

Now, I want to emphasize that these people had never had an opportunity to operate the release latch of the cartridge. So, again, they had never seen the product. They had never inserted or removed a cartridge, yet 70 percent were able to succeed. That's an admittedly small sample size, and it's obviously a limitation that the panel needs to take into consideration.

With regard to the next question, guidance on 55 pounds or less, I'll show you the labeling we have in the quick reference guide. Slide up, please.

The statement is zero to eight years, less than 55 pounds use the pediatric pads, and it chose adult pad placement adjacent.

Now, the guidance we give in the owner's manual is if uncertain about whether they're 55 pounds or greater, to go ahead and use the adult cartridge. We believe this to be an entirely safe does, and it's consistent with current AHA recommendations on pediatric defibrillation if pediatric defibrillation is unavailable, go ahead and use an adult does with the biphasic therapies.

We have selected the anterior posterior positioning because we also did a validation on pad placement. I don't have specific data with me, but we found that in very small children, ability to place anterior-anterior pads was quite poor because people go the pads adjacent, not providing a good defibrillation vector. But we found very high success rates in placing anterior-posterior pads. So that is the recommended placement, and for simplicity, we wanted to keep the pad placement consistent for all infant child application even if you were using the adult pads.

The next question was with regard to respiratory arrest. Your statement is correct that there is no discussion of respiratory arrest, and I think that this is an issue of some controversy in the resuscitation community today as to what the proper sequence of rescue is, given perhaps lack of information about the nature of arrest.

So for that we really have to defer to our policy and recommendation body, such as the American Heart Association. What we're trying to do here is to provide the opportunity to defibrillate, but I think it would be out of place for a manufacturer to make some specific recommendations in that regard.

Did I address all of your questions?

DR. RINGEL: Many of them. I think that some of the others we can discuss this afternoon.

DR. SNYDER: Okay.

DR. RINGEL: That will be fine. Thanks.

ACTING CHAIR LASKEY: Dr. Somberg.

DR. SOMBERG: Is there a population the manufacturer has identified that would not be appropriate for this device? For instance, and I'm not suggesting it is an inappropriate population, but I'm just using it as an example patients diagnose and with current epilepsy, for instance, who are prone to seizure disorder.

is there any population that has come to the fore from your consumer queries, et cetera, that suggest caution should be identified in that group?

DR. SNYDER: If you'll give me a moment, I'll consult with my team and give you an answer on that.

DR. SOMBERG: Or maybe for the sake of time you might want to come back at a later point. Is that what our chairman would like?

ACTING CHAIR LASKEY: He may have the answer.

DR. SNYDER: I have the answer. The only group that we've really identified as inappropriate is that for people with certain medical conditions, this is out of our owner's manual, who are likely to suffer SCA, an implantable cardioverter defibrillator may be advised. The HeartStart is not intended as a substitute for an ICD.

Now, the other group I would say that could not really derive benefit from a homodefibrillator or would have minimal benefit, it's not necessarily an inappropriate market, would be people in single persons households. Obviously if they live alone, the only time another responder would be available in the event of sudden cardiac arrest is if they were a guest in the home.

DR. SOMBERG: Where is the concern about having an ICD in place? Located that's in the manual you said?

DR. SNYDER: That's correct. It's page 2 of the owner's manual.

DR. SOMBERG: So that's just probable the only instance where a physician's place in the chain might be a benefit to discern whether someone has an ICD or not?

DR. SNYDER: Well, remember there's specific labeling on the outside of the carton, as well as in the owner manual and the pre-sales material saying if you have concerns about your health or a medical problem, that this is not a substitute for medical care and that you need to contact your physician.

On the specific language, if you have concerns about your health or an existing medical condition, talk to your doctor. A defibrillator is not a replacement for seeking medical care. This message is repeated throughout our labeling materials.

ACTING CHAIR LASKEY: Dr. Krucoff.

DR. KRUCOFF: I'd just like to follow the quick question on not so much the indication for an ICD, but in patients who collapse in public settings, is there any instruction to the responder in how to identify that the patient who is on the floor has an ICD, or do you have any experience with if you defibrillate both externally and internally simultaneously how the two devices are likely to interact?

DR. SNYDER: We do not have specific instructions, and I would like to defer to Dr. Becker to address the question of interaction between internal and external.

DR. BECKER: We don't have a lot of data on that. I mean, not quantitative data that's really going to answer the question. What we do know is that numerous individuals have had the device applied to them who have internal devices, and the device is operated normally.

The specific scenario that you're suggesting, the possibility that the internal device is about to shock and a shock is delivered is one that I have yet to see report as any kind of adversity.

DR. KRUCOFF: I guess as Jeremy pointed out, on the one hand, you know, a third of the folks who are sudden death survivors don't have low EFs and indications, but I think we have to recognize that there is a skyrocketing number of individuals who are now candidates for permanently implantable defibrillators. So that population is going to be more and more a part of our population.

We can talk a little bit more about that this afternoon.

MR. MORGAN: I had one other just quick question as to whether --

DR. SNYDER: Can I add just one comment on that? I think it's important to point out that in the presence of an AED or ICD by the time you could

retrieve an AED and get it applied and get to the point of a shock, there's multiple opportunities for internal defibrillation shocks and the possibility certainly exists that the ICD has failed or is not defibrillating. In that case shock by an AED would be an appropriate activity.

DR. KRUCOFF: Yeah, I guess it's a crossover. If while you're running to get the AED the patient on the floor has their internal defibrillator fire but they don't convert, I mean, a lot of the safety margin of your device is based on its not shocking rhythms that are inappropriate to shock.

DR. SNYDER: That's correct.

DR. KRUCOFF: The population I'm thinking of are patients who actually would potentially have a shockable rhythm, and without a layperson's appreciation, we now have the potential for two devices, each appropriately going after that same rhythm.

So this is one of the places that I just think we can talk about more later this afternoon.

I had one quick question just on your anticipation of what the removal of the prescription feature would do. I guess if it's not so much Dr. Kato's answerable rate of growth, most of the data that we've seen today have by a power of ten or more public access type, you know, shopping malls, airports, airline systems.

Do you all have a sense or have you examined the ramifications of without prescription how much more a private home presence is likely to eventuate or whether you think the proportion that we're seeing in the data currently reported would be maintained?

DR. SNYDER: The proportion of home deployment versus public deployment?

DR. KRUCOFF: Right.

DR. SNYDER: No, I'm afraid we don't have any projections. Really what we're operating on is I'm going to fall back on the mission of our company, which is to make defibrillation available to as many people as possible as quickly as possible, and we have identified the home as a way to do this, and we have identified the prescription requirement as a barrier to the purchase of these products.

We really are unable to extrapolate beyond that because we don't know how well accepted this will be accepted by the public, and it's certainly a sensitive --

DR. KRUCOFF: Well, that's the point, isn't it? If you remove the barrier, what happens next? If it's different from the basis of what we're thinking about as a good reason to remove the barrier, what happens next is different than how do we know what happens next.

DR. SNYDER: Well, it will only be different potentially in the extent of the deployment. Is there a specific concern that you have that I might be able to address?

DR. KRUCOFF: Well, for instance, in a public access area, if there are 100 people standing around in the middle of an airport and somebody collapses, there's likely to be an EMT or a nurse or somebody who steps forward in the use in addition to other laypeople or a calmer lay person as opposed to ten people who panic. Because, as you say, simulating these events in the types of simulations any of us could do is simply not possible.

In a home there may be a single individual. It may be a teenager. It may be a spouse, and I think some of the performance characteristics when you have people step forward in a lay environment out of a public may, in fact, be quite different than if you have one individual in the house, and at least the data that I'm looking at and aware of, that we have very little knowledge about a lot of home use. Even though we keep calling this a home use, the vast majority is really public access use.

DR. SNYDER: Yes. I do have a small amount of data I can share with you. You're correct there's not very much.

Slide G-11, please.

MR. MORTON: Dr. Laskey, Dr. Laskey. Let me ask the sponsor. Let us not lead you into economic and business results that you might present. We've got some legitimate questions, but you need to protect that information for yourself.

DR. SNYDER: Yes, thank you.

DR. KRUCOFF: I'm sorry. I had no intention of going there. In fact, I'm not interested in the business. To me the real ramification is the user.

MR. MORTON: It's just the penetration.

DR. SNYDER: Yes, I understand the question.

If we can have the slide up, please.

There are a limited number of studies that have looked at this. I can draw your attention to these three. They're all home use studies. In 1987, Chadda and Cammerer reported on experience with AEDs in the home. In the study they saw five arrests. Two of five were converted to sinus rhythm. One use was by a spouse, and there were no adverse events reports.

They concluded that AED use was feasible by family members and lay persons.

Another study by Swenson trained 48 families of sudden cardiac arrest survivors. In their study they observed five arrests, four uses by spouses. Three of the individuals were resuscitated. These are very admirable survival rates, by the way.

Again, no adverse events reported, and concluded that AED was far easier to learn and maintain than CPR.

And finally, Dr. Eisenberg who spoke this morning has done a study in which they observed two VF arrests in home. One patient was resuscitated and no adverse events were reported.

I would also on this question like to defer to Dr. Jeremy Ruskin.

DR. RUSKIN: May I first offer a response to something that Dr. Krucoff raised earlier about ICDs? Since I'm standing here, I'll take the opportunity.

It's unusual for a patient with an ICD to lose consciousness these days. The detection and charge times are so fast that for a patient to actually go down and stay down long enough for somebody to think about getting an AED would suggest that something is wrong.

And in that circumstance, at least in an in hospital environment, we routinely recommend that standard external defibrillation be performed without any hesitation or question, and I'd be interested in Bill Maisel's thoughts about that since he's got plenty of experience as well.

DR. MAISEL: My only other comment is that we routinely defibrillate patients in the EP lab during ICD testing. If they fail internal defibrillation, we routinely defibrillate patients for atrial fibrillation who have ICDs, and the general recommendation is simply not to put the pads directly over the device.

I noted in the user manual that that's mentioned. If that person has a known pacemaker or defibrillator, not to put it over the device.

DR. KRUCOFF: Actually, Bill, I'd like to come back to that this afternoon because, you know, that to me would be the key piece, is just to make people aware that if there's a lump, you know, under the skin somewhere, put the pad somewhere else.

DR. RUSKIN: That said, and that is absolutely in agreement, I think, with everybody's current procedure. There's precious little evidence that you can actually do damage to these devices. It's not impossible, but it's difficult to do.

With regard to your second question, I'm going to interpret it in a very broad sense, and that is the issue of whether or not there is effectiveness data on home use of the type that we'd all like to see as academics and clinical investigators.

And I think the answer is no. We don't have clinical trial data to direct to the question that you've asked, and it's a critical question. And I've struggled with this, and the conclusion that I've come to in trying to construct in my own mind some sort of benefit risk assessment is really to raise the question of what the stakes are and then what the risks are.

And on the risk side, I think you've heard very solid data from a randomized control trial to three other uncontrolled trials, to field experience, to simulated use data, that with this device at least there are simply no safety issues that have been raised. I mean there just is no evidence of harm to an individual victim or to a user, and there are no theoretical concerns right now for which there are signals in the experiential database with regard to safety. That's my own personal interpretation.

I think the safety is on very solid ground. You look as if you disagree with that.

DR. KRUCOFF: Well, my only real concern because the issue we're here to discuss is if you pulled a physician prescription out of the loop, then the 40 to 60 patient type family cohorts that we have data on, these are all selected patients and presumably I would think a physician was probably in that loop of identifying them as good candidates for inclusion in these studies, for instance, as one.

And what I'm literally sitting here and I think we're probably all going to have to balance is what is the burden of requiring the prescription and having the physician in the loop versus what happens if relative to the 45 well selected homes, 45,000 families who can go to Target, you know, and buy one of these things go out and start buying these things. What happens next?

DR. RUSKIN: Well, I think that's a critical question, and my own bias is that, one, no harm will come to anyone because I can't find in any of the data from all the trials and the field experience and the simulated use studies evidence of a safety concern, which for me is the most important question if you're going to expand access.

then the next question is what can you expect on the benefit side, and the answer is we don't know with certainty, but if you take a fraction, just a tiny fraction of current survival rates and you extrapolate that to a broader application of this technology, there will be some lives saved, and the question is really what's the benefit or -- excuse me -- what's the risk. What's the target, which is a 95 percent lethal condition and what is the experience to date outside the home?

All of those, I think, add up to an overwhelming benefit risk assessment, but we can't put a number to it with regard to the home application for precisely the reason that you've raised. We don't have that experience in the home.

The question is: is that achievable? And personally I don't think it is. I don't think you can do a study that's large enough with an event rate so low in the general population to answer the question. It's not dissimilar to the QT issue that keeps coming up on the drug side. It's not answerable in a clinical trial.

So where are you left? You're left with trying to construct the best benefit risk ratio that you can from the available data. It's not a perfect situation, but my own bias is that with a situation in which we're losing 95 out of 100 victims and we know that defibrillation in every other scenario it has been tested works, even if the success rate is a tiny fraction of what it is at O'Hare Airport, close to 40 percent and 60 percent in other studies, take that down to five percent or take it down to one percent. I think you're still going to save some lives.

DR. KRUCOFF: Yeah, I don't think there's going to be any argument about that from anybody, Jeremy. I think the real question is what's the burden of the physician in the loop versus what's the concerns about taking the physician out of the loop.

ACTING CHAIR LASKEY: We're going to come back to this subject, this language, I'm sure. I would suggest we take a break so we can at least hear the FDA presentations before lunch.

But I have just about noon. If we could take a ten minute break and resume at 12:10.

Thank you.

(Whereupon, the foregoing matter went off the record at 11:58 a.m. and went back on the record at 12:13 p.m.)

ACTING CHAIR LASKEY: The plan here is to have the FDA presentation followed by some short questioning, hopefully very short, and then we'll break for lunch.

DR. TOVAR: Well, during this presentation, I am going to show a brief summary of the regulatory history of the HeartStart Home Defibrillator. Next I will present the regulatory context of a prescription device followed by a summary of the review. The panel questions are going to be presented in the afternoon.

At this time I would like to present the members of the review team. Dr. Lesley Ewing was the primary clinical reviewer. Michael Mendelson, peter Carstensen reviewed human factors. Gay Kamer is statistics. Dr. Victor Krauthamer, risk analysis, and Ms. Beverly Gallauresi, post market.

And I will use some of the clinical aspects of this submission as well. I was the lead reviewer.

The purpose of this submission is to remove the prescription requirement from the Philips HeartStart Home Defibrillator prescription label. This is a summary of the regulatory history for the HeartStart Home Defibrillator that the sponsor presented previously. In September of 1996, the FDA cleared the ForeRunner AED. In May 2001, the FDA cleared the FR-2 defibrillator with attenuated defibrillation paths for pediatric use with prescription caution.

In November, 2002, the FDA cleared the HeartStart Home Defibrillator designed for home use. All the three previous were always with the prescription caution.

In April 2004, Philips submitted a 510(k) proposing removal of the prescription caution from the labeling.

As I mentioned before, this device has had a prescription caution that reads, "Caution. Federal law restricts this device to sale by or on the order of a physician."

The Code of Federal Regulations defines a prescription device as a device which, because of any potentiality for harmful effect or the methods of its use or the collateral measures necessary to its use is not safe, except under the supervision of a practitioner licensed by law to direct the use of such device and hence for which adequate directions for use cannot be prepared.

The present regulations do not describe requirements for removing the prescription language. However, the definition of prescription device suggests that two elements must be met to remove the prescription language.

One, the device is safe to use without supervision of a licensed practitioner, and adequate directions for use can be created.

Therefore, the FDA's review has focused primarily on the device labeling and human factors evaluation, and those are the summaries that I'm going to present.

First, I would like to talk about the device itself, the automatic defibrillator. Philips Medical Systems states that the HeartStart Home Defibrillator and its accessories, including this 510(k), are the same as the HeartStart Home prescription defibrillator. The previous was the over the counter.

And these similarities are in the way from therapy, the sign, functionality, technology, software, manufacturer processes, acceptance criteria, and packaging with some modification to the text of the written materials. The target populations is going to be the same, adults and children.

In summary, this device is the same device already used and comes with a prescription.

These are the indications for use of the device. The HeartStart Home Defibrillator is intended to be used to treat someone who the rescuer thinks may be a victim of sudden cardiac arrest. A person in sudden cardiac arrest does not respond when shaken and he's not breathing normally. If in doubt, apply the pads.

For children eight years or older or who weighed 55 ponds or more, use the defibrillator with the adult pads that come with it. For younger children or those who weigh less than 55 pounds, the special infant child pads should be used if available.

Next I'm going to present the summaries of our review.

Around 80 percent of sudden cardiac arrest occur in homes. However, the greater experience with 80 is in public places, like for example the PAD trial in the Chicago area airport, the casinos study, et cetera.

Cardiac function, the ventricular fibrillation deteriorates rapidly with time. This fibrillation in function is associated with rapid decrease of survival of sudden cardiac arrest produced by ventricular defibrillation.

A victim of cardiac arrest loses approximately ten percent of the probability of survival, which means that state in VF has been previously presented.

Therefore, rapid access to a defibrillator addresses this problem, and in some instances might be sufficient. A recent report presented previously by Dr. Baker proposed that if the duration of ventricular defibrillation is less than four minutes, the electrical phase, a defibrillation shock could be sufficient to compare VF to a normal sinus (phonetic) rhythm. Between four and ten minutes are the circulatory phase. CPR should proceed or you might need it for defibrillation.

Beyond ten minutes of ventricular defibrillation of the metabolic phase, other additional measures like control reperfusion, cooling, et cetera are required.

In an effort to increase survival from sudden cardiac arrest, the American Heart developed the chain of survival. If a person is un responsive to links and the chain of survival are early access or call 911 or the EMS because other conditions can cause unresponse in victim, for example, stroke, choking, et cetera.

We have a question for the panel on this issue. Early CPR is the next link, pump and blow, and the defibrillation. Use the AED followed by early advanced care or care provided by EMS personnel.

The timing of CPR relative to defibrillation is one concern for this type of device. CPR should precede defibrillation in the chain of survival according to the American Heart, according to the guidelines from 2000.

However, we are aware of the dynamics of this sequence. When the chain of survival was established, the arrival of the defibrillator to the victim would take some time, and research suggests that even changes can be done to CPR. That means the weight of CPR is done. We are aware of all these changes, and we want to recognize that.

But so far this is the sequence. However, the design of the device allows that if the victim collapse is witnessed, there is a reasonable probability of a prone application of the AED and shock delivery within four minutes and probably no need for CPR.

If the victim collapses, is a witness, there is a high probability of longer duration of defibrillation and pressing of non-shockable rhythm, and in this situation the device should not deliver a shock and even call CPR.

Of course, with this examples, I am not trying to cover every single situation during a rescue attempt. I am trying to put the extremes of the spectrum, and we have also a question for the panel on this issue.

The next studies were presented by Philips previously, and I'm not going to repeat all of the results, and the next studies that I'm going to just mention were simulated tests with the manikin or comprehension tests to study appropriate and timely use, speed of application of the elctropads of the AED, as well as safety and effectiveness of the defibrillator, and this were the three studies, the safety and the civility study, the Liberty label (phonetic) user survey and the signed validation.

The Sarapo Infant/Child Cartridge was a study to show the ten users who had viewed the training video could recognize the need to change from adult pads to infant/child pads and successfully change the cartridges. This was one of our concerns and is one question for the panel.

The labeling, label evaluation was a written test. The purpose was to assess the participants' comprehension of one of four labeling items. There was a simulation part with 190 subjects were tested on how to set up and place pads based on their understanding of the labeling. We have, again, another question for the panel. And there was a marketing survey.

The comments or concerns that we have are -- actually our main concern on the previous test was that testing was preceded by a script read to the subject informing them that the manikin was a simulated victim of sudden cardiac arrest; that the victim required immediate help, and then that an AED might be used to help the victim, and also the EMS has been already called.

Survival rates and adverse events with home use of AEDs have not been directly evaluated.

The full performance summary, Philips reported 150,000 units distributed with approximately 200,000 uses and reported 59 devices among functions that are represented here, and at the time of our review there were four AEDs, and specifically the FR-2s that fail in a use related event. That means that you compromise delivery of therapy.

The sponsor dates are different from ours because they are including the reports up to date. That's why they presented six. One case was a case that it was believed it was a simulated case, and it turns out to be that a patient was involved.

Human factors. Human factors examines all aspects of assistance interface that are necessary for safe and effective use. The definition of use includes the installation, calibration, operation, maintenance, repair and ultimately the disposal of the system or its components.

Training and labeling are part of the use interface of a device. In this context, delivery of medical care is only one part of a device's use. In devices for emergency use, as this one, concerned with preparing a device at maintained readiness are extremely important. For the HeartStart Defibrillator, it appears that the user must accomplish the following basic activities: set up training, operation, storage, and maintenance of the device.

The actions needed for the operation of this device includes determine that a patient is unresponsive. Notify 911. Turn on the device. Expose the patient's chest. Place electrodes. Respond to the device's instructions, possibly pressing a shock button and possibly perform CPR. And attend to the patient until arrival of emergency personnel.

Philips' testing covers some aspects of device wrap (phonetic) and device operation. However, Philips' testing did not cover training, storage, or maintenance.

And we have two questions for the panel on this issue. FDA considered this a track device requiring the sponsor to have processes in place to promptly identify users in the event of a recall. The sponsor proposes the use of a registration card to build a database of shipment records.

However, in the case of a recall, multiple methods of identification will be used, like Web sites, collaboration with consumer organizations, public warnings. We have a question for the panel regarding this issue.

For the post market surveillance, Philips is planning to continue with the NDR reporting as for the prescription device and also has a follow-up survey of post market study that consists of a follow-up survey of U.S. customer after one year of after-years of the device, whichever come first. Purpose: assess safety and effectiveness following use of the nonprescription device and also a strategy that ties the direct reordering of pads.

Our final comments are uncertainty remains regarding the public's ability to safely use AEDs given the increased access, and we have conveyed this concern in some of the questions. The sponsor has presented data that characterizes the human factors attributed to the device and labeling.

However, survival rates and adverse events with home use of AEDs have not been directly evaluated.

Thank you.

ACTING CHAIR LASKEY: Let's see. Is there anybody else from the agency who's presenting? Is that it?

Okay. Anybody have a comment? Yes, John.

DR. SOMBERG: I'm going to ask the same question to the FDA group here as I did to the sponsor, and that is in your review have you identified a population, a subpopulation that you think would be placed in some degree of jeopardy by the use of this device?

And if you have identified that population, would the physician link in the prescribing obviate that problem?

DR. TOVAR: No, we didn't identify that population

ACTING CHAIR LASKEY: Yes, Dr. Vetrovec.

DR. VETROVEC: A somewhat different question, but in the experience of the agency with other devices, I recall that there has been approval of the Heart Card for monitoring heart rhythm that patients can elect to do. That does require at least not a prescription but physician input, but is there any other model that the agency has dealt with in which prescriptive authority has been removed from a medical device? And what has been the experience with that?

DR. TOVAR: As far as my knowledge, I don't recall anything on that, but probably Megan could help me with this.

MS. MOYNAHAN: Megan Moynahan, FDA.

Our primary experience with over-the-counter products has been with over-the-counter diagnostics, blood tests, urine tests for pregnancy, ovulation, glucose monitoring, and that. I think there is something called a liquid bandage, which is considered a therapeutic device, but besides that, our primary experience has been with over-the-counter diagnostics.

ACTING CHAIR LASKEY: I thought I saw your hand up earlier, Mitch.

DR. KRUCOFF: No.

ACTING CHAIR LASKEY: No? Okay.

Well, then I apologize for our premature break. I didn't realize we'd be so brief.

That brings us to the lunch hour, which we've been smelling for the last hour.

(Laughter.)

ACTING CHAIR LASKEY: We've been trying to do something about that with management, but I have 12:30. If we can adjourn and break for lunch and meet back here at 1:30, I think we'll still stay to schedule.

Thank you much.

(Whereupon, at 12:34 p.m., the meeting was recessed for lunch, to reconvene at 1:30 p.m.)

A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N

(1:37 p.m.)

ACTING CHAIR LASKEY: Okay. Welcome back again. Thank you for your promptness. Hopefully this will get us through the remainder of the day on schedule. And I'd like to resume this afternoon and open by having our lead reviewer, Dr. Maisel, give his review and/or query the sponsor.

So Bill.

DR. MAISEL: Good afternoon. I don't think there's any debate that AEDs effectively defibrillate the heart and can restore the rhythm back to normal, and there have been some very impressive results that have been cited and discussed this morning regarding public access defibrillation: American Airlines, Chicago Airport, casinos, the public access defibrillation trial. I think public access defibrillation is a good thing, but I don't think that's what we're debating today.

I think you could make the argument that with prescriptions public access defibrillation could go on as it is now with AEDs wherever they needed to be.

So what I'd like to try to focus my remarks on are the specific implications of removing the prescription from this device and the first topic I'd like to try to cover is the intended patient population, and it has been stated by both the sponsor and the FDA that there is no change in the intended patient population. I'm not sure that I completely agree with that statement.

Something is happening when a patient goes in to see a physician and because not every one of those patients walks out with an AED, and so I'd like to ask the sponsor if they could be a little bit more specific in both their presentation and in the packet they presented some anecdotal comments that physicians made or that patients made stating what their physicians said regarding the reasons for refusing to give them a prescription or dissuading them from getting a prescription.

So I wonder if you have any data from physicians or from patients reporting about their physicians, about the reasons why physicians turn down patients for an AED.

DR. SNYDER: We do have some data, if I can have Slide G-33 up, please.

This is information gleaned from our telephone fulfillment center. Now, this is the point of contact for a customer who has either gained information from a Web site, a print advertisement, seen a spot that we've run on television, has interest in a home defibrillator. This is the number they call to get information from Philips.

First, I want to draw to your attention the fact that in this three-month period or -- excuse me -- four-month period we had contacts from a little over 5,850 people interested in a home defibrillator.

Now, of those, based on the information and materials provided by Philips through the phone contact, 90 percent of those did decide that this was not the purchase for them and declined to pursue the purchase further.

So the first point I want to make is this has not been a hard sell, and in fact, Philips has successfully dissuaded 90 percent of our potential customers.

But of those that have decided to proceed, the 9.6 percent -- that's 619 individuals -- these are people who have researched the issue, made the decision that they are willing to put down a credit card in order to secure a purchase. We then instruct them in the need to get a prescription from their physician and assist them in making that appointment so that they can talk to their physician about a prescription.

Next slide, please.

Of those people that go to their physician to get a prescription, 71 percent were refused a prescription. Twenty-nine percent did receive a prescription from their physicians.

Now, we've done a further breakdown of this data. It has not been submitted to the FDA. We did categorize the reasons given for refusal of prescription. With the permission of the Secretary, I'd be happy to share that data.

MS. WOOD: That would be fine.

DR. SNYDER: Next slide, plesae.

This is a breakdown of all the reasons for refusal cited by the physicians when these people who had made a purchase commitment decided to attempt to get a prescription. Now, you'll notice that there are more reasons cited than there were refusals. This is because sometimes multiple reasons were given. So this is categorized as a percentage of all reasons given.

The most common reason for not prescribing a defibrillator, 48 percent, was that it doesn't meet or doesn't treat your condition or you don't need it, and I think this is really the point that we're trying to make in our presentations here today, is that what we're trying to address is the population that doesn't have an identified medical issue.

The second most common reason for refusal is simply that the physician was not comfortable with prescripting the device or no reason at all is given.

The remainders fall in the four percent or less reasons for refusal, and they include things like some physicians offered to implant an ICD instead. The cost was cited, suggesting to use EMS instead, and so forth.

Slide down, please.

ACTING CHAIR LASKEY: I'm sorry. These are per patient report or you went to the physicians and got their --

DR. SNYDER: No, this is patient reported reason that they were given by their physician for refusing the prescription.

DR. MAISEL: Okay. Thank you.

I find that very interesting and useful information.

Next, I just want to clarify a couple of things about the device description. You did a very thorough job of explaining the device and what it does, and I understand that the device is already approved and, therefore, there's been no new arrythmia testing, but it would help me to understand a little bit more regarding the algorithm for detecting ventricular fibrillation, exactly what it's doing. You put up that slide with a variety of generic statements, and my main issue is just trying to understand from a safety standpoint what sort of circumstances might make the device misinterpret a rhythm.

DR. SNYDER: Slide up, please.

This is a graphic depiction of basically how the algorithm works. I talked about the four different mathematical measures we take of the ECG, and what I've done in this graph is I've plotted three of those. The first is the morphological stability of the complex. Again, how repeatable is the complex, not the heart of our intervals, but the shape of the complex. How repeatable is it from one to another.

And as it's scored from the low end here, you can consider this to be a zero. It's a high stability you can consider to be a one.

Similarly, the conduction properties, the health of the myocardium, this is basically measured by rapidity of edges in the complexes, sharp transitions in the complexes and so forth, and that's plotted along this axis from lowest conduction, which would be a very sinusoidal in nature, ventricular tachycardia, for example, up to the very highest conduction values, which would be typical of a normal QRS complex or a rhythm of super ventricular origin.

And then on the vertical axis we have the heart rate, which in this graph goes from zero complexes per second asystole up to 600 per minute -- excuse me -- not per second. What I've plotted here are the results of analyzing three different rhythms, and this was our validation database and our original development exercise of this algorithm.

The red dots represent rhythms for which our over reading physicians -- we sought consensus from three professionals, a cardiologist, an electrophysiologist, and emergency physician. The red dots are the rhythms that they judged would benefit from a shock.

The green rhythms that you see down in the lower corner are the rhythms for which no shock should be advised. Those include super ventricular tachycardias, normal sinus rhythm, bradycardias.

And what you'll see is characteristic, for example, of normal sinus rhythm, is it has very high stability, very high conduction, at a modestly low rate. So it falls down in this corner.

In order for us to advise a shock, when we plot those three numbers, it has to fall on the other side of the gray surface. For a no shock or for a rhythm requiring shock, it has to be above the surface. If it falls below, that results in a no shock advice.

So that you see, ventricular fibrillation, for example, resides in this part of this graph. It has low stability. It has moderate conduction, and it has moderate to high rates. So this is the cloud representing ventricular fibrillation.

With tachycardias, you often get higher stability, higher conduction properties, and they reside in this part of the space. Normal sinus rhythms, bradycardias, asystoles, they reside way down here.

Now, of course, there are some indeterminate rhythms, rhythms for which for fusion status it's really uncertain from the ECG alone. We have taken a conservative approach to these arrhythmias, and those are represented by the small number of red dots for which the majority of our over eaters would have preferred the shock, but our algorithm represents a no shock, and this reflects the conservative nature that's responsible for the truly high specificity we've achieved with this arrythmia system.

DR. MAISEL: Is that latter statement then consistent with what I read, that -- and I'm quoting the most common reported concern, I guess, from users -- was that the device did not recommend or deliver a shock for a rhythm that was considered shockable? Can you comment on that statement?

DR. SNYDER: Yes. That statement would be in the case of a very fine VF, either had a very low rate of complexes, typically below about 130, or it was a very low amplitude which actually was right on the edge of whether we considered it an asystole or actually a fine VF.

So these are patients from which outcome from shock, certainly the arrythmia could be terminated, but it actually may be more appropriate and beneficial to do CPR and see if the nature of the arrythmia could be improved.

DR. MAISEL: Another question I had regarding the device description was there was a mention of the event review data management software. In other words, I'm trying to understand if a patient does survive their experience, is there a rhythm strip that some medical personnel can review, and if so, how do they obtain that strip?

DR. SNYDER: Yes, there is. I'm looking for the proper slide if you give me a moment. There is a number of pieces of information we record within the device. Here we go. Slide up, please.

Internal to the device is a memory, and we actually performed three types of data collection. First of all, there's data which is useful in handoff to a second tier responder, and it's available audibly through the voice prompts of the device. And this information is accessed by pressing the I button. There are instructions in the owner's manual so that purchaser knows how to do this for EMS when they arrived. Certainly EMS organizations will become aware of this as well.

It will report the number of shocks that have been delivered to the patient and the minutes of use. So that information is available immediately to second tier response.

The device also records information for the purpose you referred to, review through our event review software. It records event data for 15 minutes, including internal storage of the electrocardiogram in events, all events. That's shock advisories, shocks delivered, and so forth that occurred during those first 15 minutes of patient use.

And then finally there is some more internal data that's recorded, including device manufacturing identification, records of self-test performance and so forth. So we have a complete record of the devices, the diagnostics on a day-to-day basis.

DR. MAISEL: I was a little confused. I believe I read that it said the event review data management software continued to require a prescription. Is that --

DR. SNYDER: That's correct. It does. The event review software is really not appropriate for a lay user. I can't envision why a lay user would have need to evaluate the ECG. It's really appropriate for an EMS reviewer, physician review. And it's the same software we use for all of our other defibrillator products. So any system that has event review software has full access to the ECG information that's contained in this device.

DR. MAISEL: So I'm a little bit confused. Does the product that we would be giving OTC status to have -- it has the capability to store the ECGs --

DR. SNYDER: That's correct.

DR. MAISEL: -- but only an appropriate medical person can --

DR. SNYDER: Can actually gain access to that information.

DR. MAISEL: -- get it out of that device?

DR. SNYDER: Yes.

DR. MAISEL: Why is that? Why can't -- I just see that potentially as a barrier to the patient getting appropriate care once they get taken off to the hospital.

DR. SNYDER: Can you explain in more detail your concern?

DR. MAISEL: A patient has a cardiac arrest. They're resuscitated. EMS comes and they would like to immediately take the patient to the hospital.

DR. SNYDER: Yes.

DR. MAISEL: There is a delay in trying to retrieve from the device the rhythm strips that will be helpful to the patient's subsequent care.

DR. SNYDER: That's certainly a limitation of the product. Unless the hospital had the specific event review software, they would not be able to gain access to that. The information they would have is the length of time the product was attached and how many shocks had been delivered.

DR. MAISEL: I see that as a little bit of a limitation. It's, you know, going back to ICDs in the early 1990s when people got shocked and we didn't know why. Certainly you've demonstrated relatively convincing sensitivity and specificity of the device, but that concerns me a little bit.

DR. SNYDER: I will add that this capability is similar to all other AEDs that I'm aware of in that some additional facility is required to extract the data from them. I'm unaware of any AEDs for even public access or EMS response that have built in printers. With that kind of functionality you typically go up to a manual defibrillator.

DR. MAISEL: All of those AEDs require a prescription.

DR. SNYDER: That's correct.

DR. MAISEL: Two main safety issues that I wanted to discuss and mainly trying to get at the point of whether the AED is truly helpful in all cases or whether there are circumstances under which it could actually be harmful.

The first is I'd like to discuss the potential delay in notifying EMS, and I'm a little bit concerned that the only initial notification warning on your device is simply a sticker on the outside of the device, and maybe you can discuss why you chose not to have the very first thing that the device says when you turn it on be, "Notify EMS," or, "Call 911."

That seems like a very simple thing to do. Why was that not done.

DR. SNYDER: Certainly. I'll address that question first, and then I'd like to ask Lance Becker to come up to add a little more perspective on the history of notification of EMS.

But the reason we didn't make the initial voice prompt, "Call EMS," because imagine the scenario. The device is placed in the home according to recommendations adjacent to a phone. The arrest occurs in a remote part of the house. The event is witnessed. You go to retrieve the AED. You then walk back to the patient. You open the device, and then you get the instruction to activate EMS at which point you have to go back to the telephone, dial EMS, discuss the problem with EMS, then go back to the defibrillator and actually deploy.

If we get past the initial labeling, and I want to emphasize that our experience with lay people using AEDs is that they don't fail to activate EMS, and in fact, the studies you heard about this morning were using products that had no labeling about calling EMS.

So we've taken that successful EMS notification history and we've now added two more layers of labeling to remind people to call EMS. So we're quite confident in that regard.

DR. MAISEL: So are you suggesting that if a user does not immediately notify EMS and they bring the AED to the side of the patient they should not go back and notify EMS? I mean what you just said suggests that you're suggesting something other than the chain of survival.

DR. SNYDER: the protocol that has been implemented is if you fail to activate EMS per protocol, take the device to the site of the patient in VF. The protocol that's implemented is to deliver shocks and then issue a reminder.

DR. MAISEL: Okay. I guess I also have an issue with that, I think, for a number of reasons. You've cited many examples of when the device is used or put on a patient and it does not do anything for the patient. There are times when the shocks are delivered, but EMS subsequently comes and delivers additional shocks.

I think notification of EMS before use of the device remains a critical component, and the scenario you describe worries me a little bit, and I am concerned that the label alone is not adequate.

DR. SNYDER: Yes. I think Dr. Becker has some information he might like to add to my comments.

DR. BECKER: The concern that you raised is one that has been sort of peer reviewed, expert reviewed by a number of very large panels, and it is with some difficulty that everyone agrees that you may not be able to come up with something that is perfect in every case.

What the American Heart Association determined as they thought about this subject is what do you do -- it's sort of facetiously called the loan rescuer issue, and the issue is if you are the lone rescuer, do you apply a defibrillator or do you call EMS first.

And in fact, the decision was that in the adult, you should apply the defibrillator first and see if that was a shockable rhythm and then go to the telephone, and that is the American Heart Association standard, and that's an international standard as well.

So one can imagine why that could be problematic in a very few cases. I think the thing to remember with this device is in approximately 60 seconds you will then be prompted to call EMS.

So those are the current ways that we do it.

DR. MAISEL: The other potential safety issue I wanted to discuss was the timing of CPR, and certainly there is some data, and we can debate the merits of the data that in patients who have been down for longer than four or five minutes, that CPR first rather than immediate defibrillation might be of some benefit. And I certainly recognize the trouble in making things more complicated than they need to be.

Are there instances in your mind when CPR should be administered prior to placing the AED on the patient? For example, in a patient who is found down and either the time is known to be down for longer than four or five minutes or it's not known how long the patient has been down?

DR. SNYDER: I'd like to first review what I believe is the study you referred to, and just review the data quickly so that we can consider it, and then again, I'm going to ask Dr. Becker to answer the more general question.

Slide up, please.

I believe you're referring to the study that was done in Oslo, Norway conducted by Lars Wik, and this study examined 200 VF sudden cardiac arrests over a period of 36 months. Patients were randomized to either immediate defibrillation. There was 96 in this arm or three minutes of CPR prior to defibrillation, and there were 104 patients in this arm.

Now, in the system during the duration of this study, the mean collapse to arrival time -- so that's collapse of patient estimated until ambulance arrives at curbside -- was 12 minutes.

Next slide, please.

Now, the general hypothesis was that three minutes of CPR performed prior to shocks would result in survival benefit for the overall population, and this hypothesis was not sustained in this study. There was a trend towards that result.

If you look, return of spontaneous circulation had a trend towards better results with the CPR first arm, the same for survival to admission and survival to one year or survival for one year following arrest, though none of these differences were significant.

Next slide, please.

Now, what was interesting in this study was they examined the subset of patients for whom not the down time, but the ambulance response time was greater than five minutes, and in this cohort of patients they discovered that there were significant improvements in return of spontaneous circulation, survival to hospital admission, and one year survival.

But if we add up the numbers that Dr. Becker provided to us this morning, that is, approximately four minutes typically from collapse to activation of EMS, in this case five minutes is the cutoff for ambulance response. That's nine; another two to deploy and deliver defibrillation. We're really talking about a ten or 11 minute down time population in this study.

And if we go to the next slide, which represents ambulance response times of less than five minutes, we see that the trend is actually reversed. Again, none of these differences are statistically significant, but observed rates of ROSC was actually higher in shocks where survival was higher and one year survival was also higher, again, not statistically significant, but the trend was in favor of shocks first for, and when we're speaking of down time, approximately ten minutes or less.

And certainly with this device we're hoping to achieve response times much less than that. This is the response that's appropriate to EMS arrival.

Dr. Becker, would you like to add further comments?

DR. BECKER: This is an issue that a lot of us are very interested in, and I think the point that I would like to make is that what we have to go on today is exactly what the device, in fact, advises.

And if I could have the slide up, please.

This is the AED treatment algorithm that is currently circulated all over the world, and what it essentially says is that as soon as you have an AED available to you, you attach the AED and you defibrillate if that rhythm is ventricular fibrillation.

So I just want to highlight -- slide down -- that our current state of international and national recommendations is when you see ventricular fibrillation, to defibrillate that rhythm.

Now, there is a body of data that is emerging to suggest that there may be a time where it would be better to receive CPR first, and I personally believe that that's probably going to be true, but as of yet, there is not a national organization that has made such a recommendation. There's not an international organization, and there are really fine training organizations that will be examining this, and that may be something that will change in the future.

DR. MAISEL: What would be the recourse for -- "recourse" is probably the wrong word -- but a person purchases the device and there are significant changes in the recommendations regarding defibrillation or CPR over the course of the next two or three or five or ten years, and you know, there are hundreds of thousands of devices out there that are not doing the recommended algorithm. How would we respond to that situation?

DR. SNYDER: Well, to begin with the device is actually configurable. Again, you use the event review software so that it's not configurable by the lay responder. But the protocol can be changed. It supports one, two, or three shocks after every interval of CPR, should they be indicated.

It also has various CPR compression intervals that can be programmed, anything from a half minute to three minutes at a time. So there is a good deal of flexibility built into the device.

Beyond that, if protocols are adopted that the product is not currently capable of supporting, the software can be upgraded in the device, and this would most likely be handled by a male end upgrade for the product.

DR. MAISEL: Finally, I wanted to touch on some of the safety issues that have been discussed this morning both by the FDA and Philips. From your submission it appears that there's about 167,000 devices that are in service and close to 450,000 device service years, and so I think the relatively small number of adverse events that have been reported in the context of that is not that concerning to me.

What is of more concern is that there do continue to be a number of recalls related to these products. I know Philips themselves had a recall related to their AED, and I'd like to have you walk through how a patient notification system would work if a patient needed to be notified or a buyer needed to be notified that their device was under advisory or recall.

DR. SNYDER: Certainly. First, I'd like to bring up Slide S-77, please.

I would like to review our recall history for the panel's benefit because certainly there have been many recalls of defibrillators.

Slide up.

The recall history, this product line of defibrillators, the ForeRunner, the FR-2, and the On-Site, is we have had one Philips AED recalled. That's not one recall. It is one recall, but it's one device.

The FR-2 has been built with components that might cause -- excuse me? Oh, this was in a device -- excuse me -- that had been built with a component that could lead to an early self-test failure. It was not a device that we understood to be disabled. It was actually a functional defibrillator, but we did recall the device. The customer was contacted the next day, and the device was retrieved.

So our recall history is very solid, partly because of the comprehensiveness of the self-test. We call it our recall prohibitor.