Transcript for CDC Telebriefing: 2018 Mortality Data

Press Briefing Transcript

Thursday, January 30, 2020

Please Note: This transcript is not edited and may contain errors.

Telebriefing Audio

>>> Welcome and thank you for standing by.  At this time all participants are in a listen only mode until the question and answer session of today’s conference.  At that time press star one on your phone to ask the question.  Today’s conference is being recorded.  If you have any objections, you may disconnect at this time.  I would like to turn the conference over to Mr. Jeff Lancashire

>> Thank you, Sheila and thank you, everybody, for joining us for the release of today’s final 2018 mortality data for the U.S.  With are joined by dr. Jennifer Madans, the acting director of the CDC national center for health statistics, as well as Dr. Robert Anderson, the chief of the mortality statistics branch of NCHS.  The data we are discussing today are featured in several new NCHS reports which are now publicly available on the NCHS website.  Without further ado, I will now turn the call over to dr.  Madans.

>> Thanks, Jeff.  Let me thank you all for joining us this afternoon so we can discuss some of the results of the 2018 final mortality release, as Jeff said, the final file was posted on our website today and along with it six reports.  Three of those reports will address maternal mortality and changes in the coding and classification in maternal mortality.  One of the reports talks about the general mortality findings and general trends in the top tier mortality indicators.  Another looks at drug overdose and a third looks at changes in life expectancy.  All of these reports are using the final 2018 data.  We have released some more recent data through our provisional quarterly releases, that covers many of the topics that we’ll be talking about today, but our provisional data on overall mortality and also by leading causes, so and also drug overdose and also monthly counts of drug overdoses so I encourage to you check our website for those, for more recent data.  All of the reports we’re talking about today are on the website and I just wanted to also mention that this is our first release for 2018, the first set of reports, but we will be releasing additional reports based on the 2018 data in the month to come, and as I know many of you know, there is a schedule release so you can look and see what’s coming up when you get a four-week notice when we’re going to be releasing data.  I’m going to turn this over to Bob.

>> All right.  Thanks, Jennifer.  So, I’m just going to talk a little bit about general mortality and leading causes of death.  We’ll move to drug overdoses for a few minutes and talk about life expectancy, and then then we’ll talk about maternal mortality a bit.  So before I start with the day and talking about the data, I wanted to mention that the data that we’re using here, this mortality data from the national vital statistics system, it comes from death certificates that are filed in state, I also want to mention that vital registration is a state function, not a federal function, so it is governed by state laws and regulations, and we have a federal/state collaborative relationship with them, where we provide some funding and coordination of standards, and the states  maintain autonomy in their operations but they collect data according to standard specs and agreed-upon time lines.  So, with regard to overall mortality, overall mortality declined by 1.1% from 2017 to 2018.  We also saw declines in the major demographic groups.  We saw declines for non-Hispanic white males and white females, non-Hispanic black females and Hispanic females.  The rates for non-Hispanic black males and Hispanic males did not change significantly.  Ten leading causes of death remain the same as in 2017.  The ten leading causes overall accounted for about 74% of all deaths in the united states.  We saw declines in six of the ten leading causes of death and increases for two.  The others didn’t change significantly.  The notable ones here are unintentional injuries which declined after increasing for almost a decade.  The increases and the decrease that we saw were largely driven by increases and a decrease in drug overdose mortality, which we’ll talk about in a minute, and suicide mortality continued to increase.  It’s been increasing since the year 2000 and what we see today at 14.2 deaths per 100,000 population is the largest death rate since 1941.  Drug overdose death rates have been increasing over the last several years.  It would be hard to miss that headline.  In 2018, there were 67,367 drug overdose deaths in the united states.  That’s 4.1% fewer than in 2017, when it was 70,237.  So, in terms of the rate, the drug overdose death rate at 20.7 was 4.6% lower than the rate in 2017.  With regard to specific drugs, including drugs like oxy codone and hydrocodone, we saw declines for heroin and methadone, and we saw continued increases for the synthetic opioids other than methadone, which is mostly fentanyl and we also saw increases in deaths involving cocaine and category called psychostimulants with abuse potential which is mostly methamphetamine.  We saw also increases in the suicide rate, as i mentioned before, it’s been increasing since 2000.  We saw increases for males and females and by various demographic groups and those increases are continuing into 2018.  With regard to life expectancy, life expectancy at birth increased from 2017 to 2018 by a tenth of the year from 78.6 to 78.7, and to put this in a little further context, if we go back to 2010, we see the life expectancy in that year was also 78.7.  It rose by two-tenths of the year to 78.9 in 2014 and then we saw a period of declining life expectancy from 2014 to 2017.  From 78.9 to 78.1 and 0.1 of a year increase from 2017 to 2018.  Just to focus in a little bit on the 14 to 17 decline, largely due to increases in unintentional injury mortality, mostly drug overdose and also suicide.  The one-tenth of a year increase in life expectancy from 2017 to 2018 was largely driven by declines in the cancer mortality, and in unintentional injury, unintentional injuries shifted back to a decline and that again was mostly a decline in drug overdose deaths.  So now Jennifer did mention that we have some provisional data as well that allows us to reach in to 2019 up through June of 2019, and so I’d encourage you to go on to our website and check out that information.  It’s very interesting particularly with regard to the drug overdose deaths.  One thing that we’re seeing is the drug overdose deaths don’t appear to be — the decline doesn’t appear to be continuing in 2019.  It appears rather flat and may be actually increasing a little bit.  With regard to maternal mortality, just by way of background, I do want to mention we suspended publication of maternal mortality rates in 2007, and we’re resuming now after more than a decade, and the main reason why we had suspended publication in 2007 was that we had some serious concerns about the data quality, and how all of that was working.  I also want to mention that we use at NCHS the WHO  definition of a maternal death, which focuses in on the period during pregnancy and up to 42 days outside of pregnancy.  You know, that said, we do compile additional information, we go beyond the 42 days up to one year.  One thing that we noticed is that before 2003, the research showed underreporting of maternal death in vital statistics and as a result sop states introduced pregnancy check box stats to capture a recent pregnancy but there was a lack of standardization among the death certificates.  It was decided when we revised the U.S. death certificate the revision scheduled for 2003, we’d add a standard pregnancy check box item and recommend that for use in all of the states.  The cause of death section on the death certificate, the cause of death is basically determined by what is written in the cause of death field.  For maternal death we take into account not just what’s written in the cause of death section, but we also take into account the check box.  So, the check box item is really designed to help us identify additional deaths involving pregnancy and also to give us a sense for what the timing was post pregnancy.  We have this new standard certificate with a standard pregnancy check box item.  It was supposed to be implemented in all states in 2003.  That was the original plan.  Unfortunately, not all states implemented in 2003.  Actually, only five did in 2003, and those were implemented gradually over time.  The final state implemented the new standard certificate and the pregnancy check box in the year 2017 and so now in 2018 we have all states that have a check box item and so here we are prepared to report maternal mortality rate again.  But you know, before I mention, before we talk about the data itself, I want to mention that we spend a fair amount of time evaluating the effect of the pregnancy check box and that evaluation showed a fairly large increase in the number of maternal deaths we were able to identify, and this actually very quite largely by age, the older ages we identified even more than we might have otherwise.  We also, as we were working on the evaluation, identified some research that had been done on the pregnancy check box that showed that there were errors in the check box.  Just to give you an idea, in 2013 we found 147 cases of female deaths ages 85 and over that were flagged using the pregnancy check box at the time of death or recently pregnant which of course is completely unrealistic.  So, we also had to take into account not only the increasing maternal deaths that were being identified over the period but also this issue of errors in check box.  So, at the older ages because of errors in check box, we didn’t feel like we could really, that we could rely solely on the check box.  So, in the absence of explicit evidence of pregnancy or pregnancy related condition, in the cause of death section of the death certificate, we’ve not used the check box for decedents aged 55 and older.  The results of our evaluation showed the errors are fairly prevalent in even younger ages, and so we decided we needed to restrict the application check box further, so we’ve done that for decedents 45 and over.  So, with our new coding method and the new data we have, we are making this additional restriction with regard to the check box.  We’re restricting the ages from 10 to 44, where we apply the check box information, and that said, we do count maternal deaths at those older ages, so if there is a mention of an obstetric condition or pregnancy complication cause of death section, we count it as maternal death regardless of age.  Where the check box is the only evidence, we don’t use it for ages 45 and over.  So, let me just walk you quickly through the major findings with regard to maternal mortality.  Our new method results in a rate of 17.4 deaths per 100,000 live births.  This varies quite a lot by race and Hispanic origin.  Non-Hispanic black women are two and a half times more likely to die than non-Hispanic white women.  Hispanic women have about 80% of the white race.  In terms of age, at ages 40 and over, the rate is five times that for those in the age range 25 to 39.  Those under 25 have about two-thirds the risk of maternal mortality.  And we do have some difficulty assessing trends.  We can’t really work backwards with the new method because of the incremental implementation of the check box.  Even with the new method, we can’t correct for the fact that states are as they implement the check box item, that they’re getting more deaths.  So, we have a real problem with that.  So the best we can do really is to compare rates, assuming that the check box wasn’t there at all, and those rates are much lower, but they’re at least comparable across time, and so when we compare 2002 before implementation of the check box and 2018, assuming that the check box didn’t exist in 2002, the rate was 8.9 per 100,000 live births and 2018 is 8.7 per 100,000 live births.  It’s really very little change there at all.  Actually, I think the change is not statistically significant at all.  Now I do want to contrast that 8.7 rate with the 17.4 rate using the new method.  So, using the new method, the risk of dying is actually about double what it was without accounting for the check box.  One more thing you wanted to mention is that we are planning to continue our work in this area, to continue to try to better understand why these errors are occurring, and to figure out some ways to sort of solve those errors.  We have some ideas on how to do that.  We’re also looking to work with the state vital records office, state maternal and child health agencies in order to better identify these deaths and determine whether they’re pregnancy related.  Obviously, we need to make sure all this information gets back into the vital statistics system so we can accurately count the deaths.  And I will — oh, one other thing that I do want to mention is that we do have a website devoted to maternal mortality reports and products, and there all of these issues are summarized.  There are links to all the reports.  There’s frequently asked questions, a whole lot of great information.  So, we’ll stop there.

>> Okay.  Thank you, Dr. Anderson.  Sheila, I believe we are ready for questions, please.

>> Thank you.  We will now begin the question and answer session.  If you would like to ask a question, please press star one, unmute your phone and record your name clearly.  If you need to withdraw your question, press star two.  Again, to ask a question, please press star one.  Our first question will come from Dan Vergano with BuzzFeed news, your line is open.

>> Thanks very much.  I have two questions.  The first is in your discussion about the preliminary data and what it says about drug overdoses, can you characterize what it’s telling you are the increases in stimulant deaths overpowering the drop in heroin and pain killers or is fentanyl just rocketing upwards?  The second question is, will this data go into the wonder database?  It will it be updated to 2018?

>> Let me address the second question first.  Yes, it will.  We had hoped to have that ready today.  Unfortunately, due to some staffing issues, it’s not ready, and I’m not entirely sure when it will be ready.

>> It would be great to know when that happens.

>> Right.  Well, we’ll, we’re monitoring the situation and like i said we hope to have it up as quickly as possible.  With regard to the other question, it’s really difficult to tell exactly what’s driving the pattern into 2019.  We have about half of 2019 available at this point.  We have data through June 2019, and we do have information on the specific drugs.  One of the issues that we struggle with a bit is that often — well, there are actually more of these drug deaths that involve multiple drugs than there are that involve a single drug.  So, you may have really common combination we see in the data is fentanyl mixed with cocaine or fentanyl mixed with methamphetamine.  It makes it difficult to decide what is exactly driving the pattern.  The synthetic opioids are continuing — deaths due to synthetic opioids like fentanyl are continuing to increase into 2019, and we’re seeing increases similarly with cocaine and the psychostimulants with abuse potential, the methamphetamine deaths.

>> Thanks very much.

>> Thank you.  As a reminder, if you would like to ask a question at this time, you can press star one on your phone and record your name when prompted.  Our next question comes from Beth Mole with Ars Technica.

>> Thanks for holding the conference and taking my question.  I’m a little confused about the overdose data from 2017 and 2018 and what’s driving the decrease.  So, it looks like from the data charts that there’s about 2,870 fewer deaths in 2018 from 2017 but those don’t seem to be explained by the drops in the opioid deaths and the increases by, in cocaine and psychostimulant deaths.  There’s only a decrease of 798 opioid deaths and increase of 724 cocaine deaths and 2,343 psychostimulant deaths.  Is there another category of drug overdoses that are driving the overall decline?

>> Part of the problem trying to interpret the data that way is the issue of multidrug toxicity.  The categories aren’t additive.  So, there’s a lot of overlap between these categories, so a death may be actually counted in multiple categories, two or more in many instances.  So, it makes it really difficult to partition the decline in sort of using an additive method.  We really don’t have a good handle on how best to do that.

>> Okay, so there aren’t any other categories of drug overdoses that are also seeing significant declines?

>> No, not really.  I mean, the other categories are categories in which there are very, very few deaths overall.

>> You would really have to look at all the combinations of the deaths to parse out where the declines and the increases went, and you know, that’s impossible to do, it’s a lot of categories.

>> Thank you so much.

>> We are showing no further questions at this time.  But again, if you would like to ask a question, you can press star one on your phone and record your name when prompted.  One moment, please, for any additional questions.

>> Okay, without any further questions then, thank you Dr.  Madans and Dr. Anderson for joining us today, as well as reporters and guests.  For any follow-up questions, you can call the NCHS press office at 301-458-4800, or else better still send an email to us at  Thank you for joining us.  This concludes our call.

>> Thank you.  That does conclude today’s conference.  Thank you again for your participation.  You may disconnect at this time.


CDC works 24/7 protecting America’s health, safety and security. Whether disease start at home or abroad, are curable or preventable, chronic or acute, or from human activity or deliberate attack, CDC responds to America’s most pressing health threats. CDC is headquartered in Atlanta and has experts located throughout the United States and the world.

Page last reviewed: February 3, 2020