Press Briefing Transcript

CDC Media Availability, Estimates of Deaths Associated with Seasonal Influenza — United States, 1976–2007

Thursday, August 26, 2010 – 12:00pm ET

• Audio recording (MP3, 4.55MB)

Operator: Welcome and thank you all for standing by. At this time, I will remind parties that your line is in a listen-only mode until the question-and-answer session, at which time you may press star 1 to ask a question. I will now turn the meeting over to Tom Skinner. You may begin.

Tom Skinner: Thank you, Rose, and thank you all for joining us today for this media availability with Dr. David Shay, s-h-a-y, who's a medical officer in CDC's Influenza Division. Dr. Shay is going to open up with a brief statement regarding an article that's published in today's Morbidity and Mortality Weekly Report on estimates of death associated with seasonal influenza. He'll provide an opening statement and then we will open it up for question-and-answer. Dr. Shay.

David Shay: Thank you. And good morning. I'm here today to share some updated information from my team here at CDC about influenza. As you all know, influenza has a significant health burden every year in the United States. It's associated with thousands of deaths, up to 200,000 hospitalizations annually and hundreds of thousands of illnesses that require doctors visits and missed days of school or work. But this report is only going to focus on influenza deaths. As you also know, quantifying the burden of influenza, particularly death, is challenging.

To get at the burden of influenza, we have to rely on statistical models, because many folks who die of influenza aren't ever tested for it or they die of subsequent complications of an initial influenza infection, like worsening heart disease or bacterial pneumonia. So, we look at death certificates here at CDC that go into at least two categories, pneumonia and influenza and respiratory and circulatory causes, and we use, again, a statistical modeling approach to estimate how many of those deaths might be influenza associated. I think it's important to keep in context, which we don't really describe in this article because of space, that there's at least four factors that affect sort of flu mortality in any particular year, and those four would be the specific strain or influenza strains that are in circulation, sort of the length of the season or how long influenza is circulating in the united states, how many people get sick, because of course, the more people get sick, there is more likely to be more serious outcomes, and finally, who gets sick.

The influenza epidemic in a particular year is hitting younger or older people differentially, because as we know, most influenza deaths, a majority, 90% occur in people aged 65 and older. So, what the report being issued in today's MMWR has done is really just update previous estimates last made in 2009 using the same methodology, incorporating an additional four years of data. So, this is an update that confirms really what we have seen previously in terms of flu. The updated estimates encompass data from 31 seasons from 1976, '77 influenza season to the 2006-'07 influenza season, which is the last date we have death certificate data that we can look at. And that range is from about 3,000 deaths in the 1986-'87 season to a maximum, if you will, of about 49,000 deaths that we saw in the 2003-'04 season. So, our major point today really is just to, again, point out the incredible variability of influenza seasons. Influenza's pretty unique in terms of vaccine-preventible diseases and sort of the wide range of outcomes that could happen in a particular season. So, that's the opening statement and we'd be happy to answer questions.

Tom Skinner: Rose, we'll take questions now, please.

Operator: Thank you. At this time, if you would like to ask a question, please press star 1 on your touch-tone phone. Press star 1 if you would like to ask a question. One moment for our first question. Our first question is from Daniel DeNoon from WebMD. Your line is open.

Dan DeNoon: Thanks for taking my question. Dr. Shay, I'm confused by the two different ways of getting numbers. I understand that the lower-bound number – well, can you talk about them, rather than me telling you what I understand, can you explain what the difference is between the lower and upper-bound number is? And I see there is a caution here that we can't use just one number, but of course, just one number is what people want, so can you help me understand what number we should be using when talking about the average flu deaths in a season and when we look at an upcoming flu season and what do we tell people?

David Shay: Thank you. That is an excellent question or excellent set of questions there. We have two categories that we look at, as you said. One is death certificates that have an underlying diagnosis of pneumonia or influenza. 99% of those deaths are actually coded as pneumonia. So, that's to make an estimate of deaths in a particular season from pneumonia that are associated with flu. And typically, that's about 8.5% of deaths over the time period that we looked at. Now, we know that influenza causes more deaths than those just associated with pneumonia. However, that metric or that outcome measure has been used in literature for over four decades. So, that's one thing that we always look at so we have an idea how things are going over time. The broader category of respiratory and circulatory deaths we think encompasses the full picture of influenza-associated deaths, including things such as people who might die because of worsening chronic obstructive pulmonary disease or worsening congestive heart failure that results in death after an infection. And we estimate that about 2% of that broader category in any typical year is associated with influenza. And your second question is what is the one number we want you to use. We understand that that is a difficult communications message. I think, from my perspective is that because we have this very wide range of deaths, from 3,000 to 49,000 in this three-decade time period that we looked at, it's relatively meaningless to say what happens in an average season, because when you look at sort of our data, there are very few of those 31 seasons where you would have an average outcome. So, I think that's one of the reasons that CDC is trying to move towards away from a single number, and rather, give a range over a particular period of time to give a better time text to what flu really means in terms of what it does to the community. And finally, we can't predict what's going to happen in any subsequent influenza season. I mean, that's – you can just look at the numbers and how they vary from season to season to come to that conclusion very rapidly.

Tom Skinner: Okay. Next question, Rose?

Operator: Thank you. The next is from Betsy McKay, Wall Street Journal, your line is open.

Betsy McKay: Hi, Dr. Shay and Tom. Thanks. I didn't hear all of Dan's question, so I hope I'm not asking the same exact question again. But you know, very often, we get this number, this estimate of 36,000 deaths annually from flu. So, I'm wondering if you can just walk us through what that number constitutes, and does this new range replace it? Also, you did give an annual U.S. average overall of deaths from underlying respiratory circulatory causes of 23,607. So, is that an average number we can use and does that compare to the 36,000?

David Shay: Thanks. Excellent question. The 36,000 number that's often used pertains to a very specific time period from 1990 to 1999. And in that decade, where we had prominent circulation of H3N2 viruses, they were prominent in eight of the nine seasons that are contained within the data that were used to make that estimate, and those are, as you know, typically more severe seasons. We had a high mortality for that nine-year period, 36,000. In that paper from which that 36,000 number came from, the overall mortality average for that time period was about 25,000. So, this average that we're coming up with now, 23,600, is consistent with what we've published in the past. And again, our feeling is that it's very difficult to summarize flu with, as we gain more and more data, with a simple average. There are very few average seasons, you know. We have one in the '86 to '87 season there was only about 3,000 deaths, and then we have another one, you know, several years later that's about 50,000, almost 50,000 deaths. So, a simple average, I think, fails to give the impact of flu in an understandable fashion. We could give a standard deviation or some other metric of the variants in our measures, but I don't think that that would be helpful to the public, so that's why we have decided to try to focus more on the range rather than on simple average. Does that answer your question?

Betsy McKay: I mean, your advice would be that instead of saying the flu kills an estimated 36,000 people a year, which is what I and others often do say, that we should say, you know, flu kills between 3,000 and 49,000 – or 3,300 and 49,000 people a year or has over the past, you know, in seasons over the past 30 years?

Tom Skinner: Betsy, this is Tom, and I don't – Dr. Shay will correct me if I'm wrong, but as you report that range, I think it's important to note that there is such a wide range and it depends on the various types of flu that are circulating. So, I think that going a step further and saying that that range is due the fact that the number depends on, you know, the strains that may circulate would help explain it a little bit more.

Tom Skinner: Agreed. Next question, Rose?

Operator: The next is from Julie Steenhuysen from Reuters. Your line is open.

Julie Steenhuysen: Thank you. Yes, I'm just wondering if you can sort of, you know, looking forward, explain the role of better diagnostics, better methods of tracking and what impact that might have? I mean, it looks like, you know, it may be, you know, even this range that you've given us is going to change quite a bit. Is that right? And you know, what are these numbers – how can these numbers be meaningful for us?

David Shay: Well, let me start with your last question first, because I could use some clarification on that. What do you mean by how could the numbers be meaningful to you?

Julie Steenhuysen: Well, it's just in terms of comparing year to year. I mean, we want to be able to give a number that, you know, is a good reflection of what people can expect in a flu season. You know, I'm just wondering if you can sort of – I think we're all sort of struggling with how do we describe flu deaths, because it looks like it's a moving target almost.

David Shay: Absolutely. No, that's absolutely correct. Flu deaths are a moving target. Again, for vaccine-preventible diseases, there isn't another one that sort of has this range of variability, certainly, and mortality each season, so that's one of the important messages we want to get out, if you will. I think another important point to make is that the numbers confirm that influenza has a substantial burden on mortality each year, but that that mortality really can be quite variable. And then, finally, around the question of better diagnostics, that's an excellent one. For an example, since 2004, pediatric influenza-associated deaths have been a nationally notifiable disease in the united states. That's the only category of deaths or the only age group for which flu deaths are nationally notifiable. And as a result of probably better diagnostics and more, better surveillance, we have seen over the first few years of that system an increase in the number of reports that was probably due to a combination of better diagnostics and people looking for flu more actively in younger children who were critically ill.

Tom Skinner: Next question, Rose?

Operator: Thank you. The next is from Lisa Schnirring with CIDRAP News. Your line is open.

Lisa Schnirring: Hi. Thanks so much for your availability today. I was wondering if, when you looked at the numbers and your different modeling with this year's analysis, if you saw anything interesting to you or different that popped out when you looked at the age, demographics and all of those issues?

David Shay: No, not really. The numbers are very consistent. I mean, again, flu was associated in this time period, as in other time periods and in other seasonal studies that have been done in other developed countries with approximately 90% of the deaths being in those 65 years of age and older. One of the things I do think, though, that we're going to do in the future to put these numbers, perhaps, into better context is to also not only report numbers of deaths but report years of life lost. Because, of course, an influenza-associated death in an 80-year-old has less impact on society and less impact on parents than an influenza-associated death in a 2-year-old. So, that's sort of future work that we're going to plan on doing.

Lisa Schnirring: An 80-year-old. You said an 8-year-old.

David Shay: 80-year-old, I'm sorry. It would have less impact on society in the death of an 80-year-old compared to a 2-year-old.

Tom Skinner: All right. Next question, Rose?

Operator: Thank you. The next is from Tom Maugh from Los Angeles Times, your line is open.

Tom Maugh: Two quick questions. First, is this the first time you've really broken down deaths by the strain of the flu virus? And second, how long do you have to go into a flu season before you recognize which is the predominant strain that's circulating?

David Shay: Two great questions. No, this is not the first time that we've done that. Painter that we published in the Journal of the American Medical Association in 2003 was first time we really broke deaths out by strain. And second is, that's a great question, sort of how many weeks into the season will we know which is the predominant or the dominant strain? And again, unfortunately, the answer is that depends. In some seasons, for example, H3N2 will start off big and will be the predominant strain throughout a particular season. In other seasons, we've seen, particularly recent seasons, 2006-'07 where we have some influenza, H1N1, some H3N2, and some b, and most of the b at the end of the season. So, it would have been very difficult to say, except at the end of the season, sort of what proportion of each of the three strains was responsible for illness.

Tom Skinner: Next question, Rose.

Operator: Thank you. The next is from Sabrina Rice from CNN. Your line is open.

Sabrina Rice: Thank you for taking our questions. For what you were just talking about, the differences between the influenza strains, the report says that the, I guess deaths from influenza were 2.7 times higher for the H3N2. I guess, what's the difference between that one and some of the other strains that are out there? And do we know which is out this season that, should people be more concerned or less concerned? And I actually have two questions. And the second part is, with the numbers varying so widely from 3,000 to 49,000, that might freak people out, because you just have no idea. So, what advice would you give to consumers?

David Shay: Well, I think it comes back to one of our messages that we frequently give, that flu really is unpredictable. And it's important to point out, at the beginning of the season, don't know what the impact of flu is going to be in a particular season. In terms of H3N2, what is it about the H3N2 virus, viruses, that causes more severe disease than, for example, H1N1 or b viruses? One of the factors may be that that virus has a tendency to genetically change more quickly than the other two viruses. So, it changes more rapidly, and therefore, even if you've been sick with it in the past, you're more likely, if it changes quicker to get a subsequent influenza infection. So that may be one of the factors that's involved, but we don't understand all of them.

Tom Skinner: Next question, Rose?

Operator: Thank you. The next is from Don Sapatkin from the Philadelphia Inquirer, your line is open.

Don Sapatkin: Hi. Thanks for taking my question. I'm still completely confused. I hope you'll bear with me. My understanding of what you're saying, you're trying to make, I think, essentially two points. One is that the range of numbers is important from year to year, much more important than an average, and you're also changing your estimates based on additional seasons. And so, one question I have is, the old 36,000 number was always based on a range. It wasn't often reported, but I believe the range was 17,000 to 52,000. That what was in the 2003 paper. So, is that essentially equivalent to the new range of 3,300 to 48,000? That's my first question.

David Shay: Let me recap your points. I mean, I think your first point was that because of the variability of flu, we're trying to move away from an average and into giving a range of flu-associated mortality, and that's correct. And your second point is, yes, for that time period of 1990 through 1999, the range of deaths there was what you just said, from approximately 17,000 to 50,000, with an average of 36,000 in that time period. But again, that was a rather of an outlier of a decade, if you will, because of the predominant circulation of H3N2 viruses. So, in that 2003 paper, we actually looked at quite a few seasons of data ended in 1999, which is essentially '76 through '99. That had a broad range of estimates as well, not quite as broad as the range that we're reporting this time, but there was a broad range of flu mortality over the data that were included in that study as well. Does that help?

Don Sapatkin: Sort of. But then where I get lost is when you're talking about 23,000 versus 25,000. In the old study, where did one of those – what was one of those numbers? I'm trying to figure out – I know you don't want us to use a range, but I'm trying –

David Shay: I do want you to use a range –

Don Sapatkin: I mean, average. But what does the 25,000 equate to in the study, even though you don't want us to use it? I'm trying to understand it.

David Shay: That what from – I may not be understanding your question, so I'm sorry about that. That was from 1990 to 1999. That 2003 paper actually looked at a broader range of seasons, from '76 through 1999. So, from that whole time period, the average was approximately 23,000-25,000 deaths, excuse me. And for this time period, this broader 31-season time period that we're looking at, the average is approximately 23,600. So, they're very similar.

Tom Skinner: Okay, Rose, we've got time for a few more questions.

Operator: Thank you. The next is from Joe Neel from NPR. Your line is open.

Joe Neel: Hi. Well, I also – I'm really scratching my head here wondering what I'm going to use, because we really don't have a lot of time in broadcast to present a lot of numbers, and I think in a sense to say that the range is 3,000 or 3,300 to 49,000 raises a lot of questions, and I think we don't have time to answer those questions in every report. And I also wonder if it's not a bit misleading to use 3,300 as the bottom number since it's been 20 years since it was that low, and even in the last 20 years, the mortality has never been much below 12,000. So, I'm wondering, if you can help me out here, could I possibly just use the average of 23,000, which covers respiratory and circulatory causes? I mean, wouldn't that be a fair number to use in a very condensed broadcast format?

Tom Skinner: Joe, this is Tom Skinner, and I – we certainly appreciate your desire to have one single number for an average, and we fully recognize how difficult it is on you all. I think at the end of the day, if you must use a single number, I would simply say that it's estimated that tens of thousands of people may die each year in an average flu season, and then leave it at that.

David Shay: I would say that – yeah, I would get away from average flu season, because that's one of the points I think we're trying to make, but I agree with Tom, we'd say in many years, influenza is associated with tens of thousands of deaths a year.

Tom Skinner: Yeah.

David Shay: I think that would be a very fair and a very quick way to summarize the data.

Tom Skinner: Yeah. Joe, is that fair?

Joe Neel: Well, I think that – yeah, I mean, I think some people might turn to these data and look at this table and say, kills up to 48,000 a year, which is certainly accurate, but I also think that –

David Shay: True.

Joe Neel: – that can also be misleading, because in 2006 they only had 15,000, so –

Tom Skinner: Correct. You know, again, Joe, I'd go back to what I said to Betsy McKay earlier in this call, is that whatever number you use, if you also include the caveat that the number depends on the strain, strains of influenza that may be predominating, that is a little bit – you know, that's a little bit more information to help put your number into context.

David Shay: And – this is David again. I would also add that when you do cite an average that you – I know time is precious, but give the seasons that that average pertains to. Because again, as you just pointed out, it can vary so dramatically from time period to time period.

Joe Neel: Yeah.

Tom Skinner: Next question, Rose?

Operator: Thank you. The next is from Cathy Clark from Radio Pennsylvania State Network. Your line is open.

Cathy Clark: Thank you very much for taking my question. Is it fair to say, then, that if one of the strains expected to circulate in a certain year is the H3N2, that that given flu season has a potential to be a deadlier flu season?

David Shay: We can't predict. It gets back to – to begin with, we can't predict even, you know, two weeks into the flu season what's the predominant strain is going to be. It is very fair to say that on average, H3N2 seasons are much more severe than h1 and b seasons. And again, 2.7 times more deaths associated with in this 31-year period when there was predominant or prominent—prominent is a better word—prominent circulation of H3N2. And that's another one of our key messages, again, that we're trying to – even though it's a complex message, it really depends quite a bit on the strain of influenza that's in circulation.

Tom Skinner: Last question, Rose?

Operator: The last is from Katherine Hobson from Wall Street Journal, your line is open.

Katherine Hobson: Hi, it's Katie Hobson with the Wall Street Journal health blog. On any one of the seasons, the low death rate in '86-'87 and the high one in '03-'04, was it primarily the strain that caused that? I know you mentioned those four factors, but if you give kind of a snapshot of why the death rates were so low and high respectively?

David Shay: That's an excellent question. And I can't completely answer it because we don't know many of the factors that lead to these dramatic differences. It's certainly true that the specific strain and circulation is very important and that H3N2, again, kills many more people than the other two circulating strains. It's also important, as I mentioned before, the length of the season can vary quite a bit, and a longer flu season, obviously, has potential to result in more people getting sick. And how many people get sick directly influences the number of people who get serious complications who end up in the hospital and who die. And then finally, again, it's really important to think about who's getting sick in a particular year. H3N2 viruses tend to make older people more ill. And if a lot of old people are getting ill in a particular flu season, because 90% of the deaths occur in people 65 and above, the mortality for that season is going to be much higher than a season when few elderly people become ill with flu.

Tom Skinner: Okay. Thank you, Rose. This concludes our call. Thanks to all who joined us. And the report that we're referring to today, you should have received via e-mail. It will be posted on the CDC website as well. So, thank you for joining us.

Operator: This concludes today's conference call. Thank you for joining. You may disconnect at this time.

###
U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES