AMCHP 2005 ANNUAL CONFERENCE
DELIVERING RESULTS, IMPROVING PREGNANCY & BIRTH
February 19-23, 2005
HAROLD WEISS: Thank you very much, Justine and thank you all for being here today. When I first was thinking about submitting an abstract, I wondered how much interest and how much audience energy I might be able to garner and I had long discussions about whether to come here or not, and I'm pleased you are here. I think it fulfills some assurances I got from the planners that yes, this is an open minded group and we'll find some new areas to talk about interesting and something that they will probably take home.
So let's start with a quick pop quiz. The leading cause of pregnancy related serious injury is? How many say falls? How many say violence? How many say motor vehicle crashes? Poisonings? Parachute jumps? Here is the parachute jump issue. I swear this is true. Well, it's on the internet, who knows if it's true. But a pregnant Russian women went into labor in the middle of the a parachute jump and gave birth to a baby girl, healthy as far as we can tell after landing from the jump in Moscow. But to tell you the truth, parachute related injuries are not on my radar screen; however, violence sometimes is on the radar screen. How many saw the Washington Post series recently? About half. All of you should have. It was okay from a science perspective, although I think they missed a few points and we'll talk about that. But from a personal perspective, it was quite a good, rich article. It talks about the influence on the surviving families which what we sometimes forget after we look at the numbers. But violence is not the leading cause of death in all places for women of reproductive age.
Here are some of the numbers comparing violence to other injuries. In the District of Columbia where that article took place, look at homicide compared to unintentional injuries, just under HIV is the leading cause of death among women of reproductive age.
In Washington, violence against pregnant women is probably then a very important issue and may be indeed the leading injury issue.
But look in Utah where I also have some data. There instead of excess of homicides, there you have a tremendous excess of unintentional injuries. So the take home message is what's the leading depends on where you are and what the individual factors are in that community.
For the US as a whole, though, it's more like Utah than it is the District of Columbia.
So although we haven't said motor vehicle injuries are the leading cause, I think you can begin to suspect that on a national basis. But I think we have to be careful in saying, what are the leading causes? We have to say, what are the leading causes for this particular population? So the answer is that the leading cause of pregnancy related serious injury is motor vehicle crashes and I'll go in the data that supports that a little bit.
But if you want to see that for yourself, Google pregnancy in crashes some day. This is three days of Google news about a week ago. And you can see the kinds of cases that continually pop up at least in Google and covered by the news and the media.
Unfortunately, as you'll see, many of these report don't end up in vital statistics coded as due to motor vehicle injuries. And that's why it's often called the hidden epidemic.
So what I want to do in the next 20 minutes is give you
an overview of maternal injury. The nature and importance of maternal injuries to the fetus and the infant. The deficiencies that exist in data systems of why it remains a hidden epidemic. A little focus on motor vehicle injuries because you don't often hear about that, and I want to discuss and list some ways to better understand and address the problem.
But I think any talk in this area has to start first with what is the impact on the mother? And the way epidemiologists like to go about this for a new audience is to talk about the deaths in terms of a severity pyramid. We'll talk about maternal infant deaths, maternal injury hospitalizations, and maternal injury ED visits.
So let's first look at the deaths. Many of you may know that actually ascertaining among women who die of reproductive age who's pregnant and who's not is a big challenge. We do have a new item on the recommended death certificate that asks whether the mother was pregnant, but it has not been adopted by all states or it's newly adopted by all states. And if history is any guide, New York has had such a check box for quite awhile and it's not filled out very well. As you can imagine, in order to get that data accurately, what you really need to do is an autopsy and a test on every single woman.
So what the reporter did in the Washington Post story was she went to coroners offices and looked at some of the papers that had been done in Washington D.C. And in Baltimore and in Maryland, and I think got a picture of what was happening here, but not a picture of what was happening across the US and missed some of the issues relating to autopsy bias.
And in a nutshell, the autopsy bias is that almost every case of homicide does get autopsied. But if it's a car crash, it doesn't necessarily lead to an autopsy. One, only the driver is likely to get an autopsy. So if the pregnant women is next to the driver, they are not going to be autopsied and you may not know they were pregnant. And if it was the driver and they weren't killed right away but were in the hospital for several days, there usually is no autopsy because the cause of death is very clear from several days in the hospital. So the autopsy selection bias then occurs. If you rely on coroner data to understand the big picture of injury mortality and pregnancy you may be led astray.
Well, since there isn't very good data out there, what we have to do I think to get the big picture is to look more at the leading causes of death among maternal age and assume at least for the time being, and it's an assumption that has it's problems, that the rate of injury death among pregnant women is similar to the rate of injury death among non-pregnant women. That's probably not quite true, but since we don't have good data, this is a reasonable place to start from.
And when do you that, you can see that in the United States as a whole, unintentional injury in this reproductive age group, 15 to 34, that's the leading cause. Malignant neoplasms is Number 2, here comes homicide and suicide is right next to it. In the end what you end up with is a total reproductive age death in the US of 21,267, half of which are due to injuries.
And as you see, about half of those injuries are due to motor vehicle crashes. But, maternal deaths as defined by the World Health Organization is the death of a women while pregnant or within 42 days of termination of pregnancy irrespective of the duration or the kind of the pregnancy from any cause related to or aggravated by the pregnancy or its management, but not from accidental or incidental causes. This is the world health organization definition. This the definition adopted by the CDC and the Center for Health Statistics. So when you look at the statistics for maternal mortality, you don't see injury and violence, they have been excluded out. Maybe there was a time when this definition made sense, when the lack of medical care and the lack of antibiotic treatment led to a tremendous number of maternal deaths due to those conditions and they had to slice the pie somewhere. But in these days when there's so few maternal deaths, I think we need to take a look at that definition. This just shows the figures you've seen probably a hundred times, the decrease in the maternal mortality ratio in the US it's now down at a very low level. And while that definition is down to a low level, remember if we include other causes, it's really up here and I think this number is so low now we need to expand that definition.
So actually, if the number of US maternal mortality deaths in 2002 as reported by NCHS is reported as 357, I would estimate that there are approximately 770 injury related deaths just among pregnant women. And that doesn't include the longer definition of women who were injured the 42 days after birth.
So you can see, there are several times probably the number of injury deaths during pregnancy as there are maternal deaths.
So clearly this is a conclusion that one can make. There are more pregnant women dying from intentional than unintentional injury than all maternal mortality related conditions combined.
Let's move to maternal injury hospitalizations. How many of you collect maternal injury hospitalization data? Isn't that amazing? We've got the center of the healthcare system in the United States and nobody is accessing it to look at why pregnant women are going into the hospital or as we see later going into the emergency department. That should not be happening. But there are ways which on an individual basis, and I think on an institutional basis we can begin to collect that data.
This is one -- these are the results from one study I did looking at 19 states hospital discharge data using the ICD 9 diagnosis and searching for pregnancy related codes among the other diagnoses codes and making some assumptions from there that in fact those women probably were pregnant. Not a perfect way of trying to ascertain who's pregnant while hospitalized. You'll miss all the very early cases, but you'll get a great deal of the later cases, and I think it is 85 to 90 percent sensitive. And you can see from the 19 state data which covered over half the US hospital discharge population that the leading cause of pregnancy associated injury hospital discharges was motor vehicle occupant injuries followed by falls, followed by poisonings, and followed by struck by. Again, some of the violence events can be found here. And I'll be the first one to point out that many unintentional falls maybe domestic violence conditions that were reported honestly or appropriately.
So, let's go at the bottom of the injury pyramid and look at maternal ED visits. You're one of the first audiences to see this data. To my knowledge, this is the first time anyone has ever linked emergency department data with birth certificates and fetal death certificates to ascertain which women were pregnant when they made a visit to the emergency department and what was the distribution of those injuries. This is data from Utah and some colleagues that I'm working with there. Once again, motor vehicle crashes are the leading mechanism of pregnancy related ED visits followed by falls, cutting and piercing strap, overexertion assaults and poisoning. But a full quarter of all the events are motor vehicle related.
So what do we have? And I through this one in so you could see the intent versus the unintentional injuries looking at it a little different way. This is Utah, this is not Maryland, this is not Washington D.C. And you have to keep in mind those regional differences.
So the leading cause of maternal injury, back to our quiz for deaths, it's crashes; for hospitalizations, it's crashes; for ED visits, it's crashes. So let's look at that for a minute. I don't think this is a phenomenon that's been around all that long. When you like at the total miles driven by women of reproductive age from 1969 to 1990, you see this huge increase in mileage driven and this translates to risk. You know, if the rate of driving has gone up two or three fold, the risk to the mother and to the fetus has also gone up two or three fold and this rise in risk has really gone unnoticed by the maternal and child health and until recently, the injury community.
So that increase risk translates into an impact on the fetus. I'm going to talk about that area a little bit. Of course, the worst outcome would be fetal death. And I have some data to show on that. But there are many other conditions that may be coded in the birth certificate and in other statistics as prematurity, low birth weight and obstetric complication when in fact the underlying cause, the underlying event that triggered that outcome was indeed the injury. But it's not coded that way on the birth certificate or on the death certificate. But you can have all of these complications and risks to the fetus. You can also have threats to the baby. You can have injury in utero. The baby can be born live and then die. Or the baby might suffer damage either directly or indirectly in utero leading to utero or other organ or structural damage and that can lead to a host of problems, some of which are suspected, some which can be confirmed. It all depends on what region is injured in any individual situation.
Well, I mentioned fetal injury mortality. We can tell a little bit about that simply by looking at the fetal death certificates. Most of you are probably aware of that system in your state. Usually 20 weeks or more gestation, a fetal death certificate must be filled out and there was a code in ICD 9, 760.5, the fetus or newborn effected by maternal classifications, maternal conditions classifiable to the injury range. So that's nice and at least we can know from looking at the fetal death certificates how many fetuses or newborns were effected by maternal injuries. But, how many are familiar with E codes? The external classification of injury codes that we have with all the other injuries? We don't have those with maternal injuries because of the definition in the ICD 9 classification that says E codes should only be used as the primary code when the morbid condition is classifiable to injury. In this case, the morbid condition is 760.5 which is not an injury. It's in the perinatal causes of death and birth related. So there are no E codes on the fetal death certificate. It's like going back 20 years when we didn't have E codes when you try to look at fetal deaths and try to ascertain the causes.
And unfortunately it's gotten worse with ICD 10. With ICD 10, the new definition, the closest one I can find is POO point five. Instead of fetus or newborn, now it's newborn suspected to be effected by maternal injury. I don't know what you're supposed to code if it's the fetus, and I don't know what the fetal death certificates are going to look like. And this other code is newborn effected by maternal death, but it doesn't say whether the maternal death was an injury. So instead of taking a step forward, ICD 10 I think has taken a step backward and certainly the data is not going to get better in terms of fetal death.
Clearly as James Marks said in a commentary on motherhood, what gets measured gets done. And in the maternal and fetal injury area, unfortunately we are not measuring it very well, and I don't think therefore, much is getting done.
In all of these reports, the Injury Fact Book, Healthy United States, from the National Highway Traffic Safety Administration on highway safety fact and healthy People 2010, there is nothing about the injury during pregnancy issue because the data is not there. What's even scarier as we go on is we all know about the future is an issue and how science based it's going to be, what if the science isn't there? What if the numbers aren't there? We are going to continue to overlook this major problem.
But there are ways that we can do it and these are things that you can do in your state. This I did as a Ph.D. Student on a budget of about $10,000 and I did it for -- I think it was -- I forget now, it was too long ago, something like five states or ten states. But I looked at the fetal death registries and got copies of all of those that seemed to have a pregnancy injury related code and looked at them by hand to see what was on the fetal death certificate in terms of the causes of injury. You can't use the computer except to do the first selection. You have to look at the fetal death certificates by hand. Back to the gum shoe epidemiology. But when do you that, you do find a richness of written mechanisms of injury on the fetal death certificates. And in the selected states I looked in for this period 1995 to 1997, this is what I saw. This is, once again you would see and you would expect this. If the leading cause of maternal injury death and maternal serious injury is motor vehicles, why didn't it the leading cause of traumatic fetal deaths?
And if you take these numbers that I did from the -- several states and inflated those up to a national level and tried to adjust a little bit for the missing data that's even in that data system, you come up with these surprising, some say even shocking, conclusions. Comparing fetal motor vehicle crash deaths alone to other common causes of childhood injury. And you really have to look at this to understand the tremendous difference. My estimates are somewhere between 350 and 700 fetal motor vehicle deaths a year. Look at that compared to the number of bicyclists killed every year, ages 0 to 15. A hugely increased population, but much fewer cases. Firearm injuries ages 0 to 9, and motor vehicle occupant ages 0 to 1 compared to what I think is actually happening in the fetal motor vehicle area.
And unfortunately, as I listed before, the effects are not just death, but there can be other impacts on the baby. There are a few studies out there that have begun to look at this. One of them was done in 2002 by Shift that looked at all trauma in pregnancy and the outcomes of birth. But there are no studies that I'm aware of that looked at outcomes to the children impacted by maternal injury one or more years after birth. That's unbelievable to me. But here's some of the data from the Shift studies. This is all injuries, mostly motor vehicle, but still, this is all injury data, and these are the adjusted risk ratios for things like preterm labor, placental abruption, labor induction and Cesarean delivery. It's just amazing to me that all of the discussions on preventing premature labor, none of them pay attention to this problem. And why? Because the data isn't there to tell them it is a problem.
Here are some other risk ratios for preterm delivery, low birth rate, fetal distress and of course a large correlation with fetal death.
Well, if a picture is worth a thousand words, how much is a video worth? Let me show you what happens when a pregnant woman driving let's say at 22 miles an hour is in a car crash. Obviously this is not a picture of a real woman in a car crash, but what we do have in the injury area is biomechanical computer simulation of what happens. And what you see here are a few from Virginia Tech. How would you like to be that fetus? So that's an unrestrained crash. And it shows you the danger inherent in an unrestrained crash.
Let's look at a 34 mile an hour crash with seatbelts. A lot better. Not perfect, but a lot better. But still, some interaction between the steering column and the uterus that you wish weren't there. And that's with proper seatbelt use.
Let's look at front passenger at 22 miles an hour with improper belt use over the mid section. So belt use is not a panacea, just like the child passenger safety. Proper belt use is very important.
And then finally let's look at one more front passenger with the air bag, proper belt use. Some interaction between the air bag and the uterus. Maybe not too bad. Certainly, you know, notice the head protection. The mother is going to be much better off that way and we certainly continue to recommend that if women do drive during pregnancy, they should not get rid of the air bag.
But how many cars do you think are designed for pregnant women? And how many -- do you think that NTSA requires the testing of pregnant dummies in their regulations? No. They do have a regulation that the car company have to test infants 0 to 1, so there's an infant test dummy. But a problem that's demonstrateably five to seven times worse doesn't get looked at by the car companies on a regular basis. Some do it, some are looking at it, but they are not all required to do so. So those are great pictures.
There are some other population based fetal motor vehicle trauma outcome studies. Wolf did one, it was the first one in 1993 looking at non seatbelt risk and showed clearly a non seatbelt risk as did a Utah study that I was involved with in 2003 looking at motor vehicle crashes on birth and fetal outcomes. But again, there is no child outcome study of a year or more looking at children who are involved in a motor vehicle crash in utero.
Just some of the major highlights of the Utah study, we showed that three percent of all births linked to a driver motor vehicle police reported crash during pregnancy. That's up there with some of the overall prevalence of binge drinking and cigarette smoking. It's not that different. That's just the driver. Many women are not the driver too.
If you extrapolate this number to the 4 million annual births in the US, you get about 160,000 pregnant crashes per year, seven times the number of infants in crashes.
In our Utah study, we showed that pregnant women not wearing a seatbelt were 1.3 times more likely to have low birth weight babies compared to pregnant women not involved in a motor vehicle crash and nearly three times more likely to experience fetal death compared to pregnant women with seatbelts.
Well, let's look again at the injury pyramid. This time I think I inverted the pyramid. But this compares to the pregnancy related motor vehicle crashes to the infant motor vehicle crashes. Here, if we extrapolate to the US we have about 120,000 crashes involving pregnant women compared to 24,000 for infants. With an ED visits about 40,000 a year compared to 7,000 for infants. For hospitalizations, about 4,000 a year compared to 400. And for fetal deaths, about 700 versus 120. This of course is not to say we ought not to be doing something about infants and injuries, but it says we must be doing something about pregnant women and injuries.
So in summary, injury in motor vehicle injuries is a surprisingly common occurrence during pregnancy. Motor vehicle trauma to pregnant women has probably increased substantially over the last 20 years. There are critical gaps in reporting and surveillance of pregnancy related injury that hide the problem from the decision makers. Fetal injury mortality represents a large proportion of childhood infant mortality because these are babies that were going to be born healthy. We are just beginning to understand and measure adverse birth outcomes due to trauma.
And here were some of the recommendations. Maternal fetal injury issues should be incorporated into national prevention and research objective. How many times do you think the word pregnant or pregnancy is listed in the injury priorities for the CDC for the nation? Zero.
ICD coding guidelines need to be changed to allow coding external cause of injury of maternal injury in the vital statistics. Maternal injuries should be included in an expanded definition of maternal mortality, and that's something you can do in your state. .
Pregnancy status and fetal outcomes need to be added to injury surveillance systems. And I think the best way we can begin to do that is that all states with access to hospital discharge data or emergency department data should continuously link to their hospital discharge or ED data, the birth data and the fetal death data to create in reality a pretty darn good pregnancy morbidity system. That can be done. That can be done at 20, - 30,000 dollars a year assuming you can get over the data access issues. I know that's difficult for some. But for many health departments, you own that data and you should be able to do it. And I think the data linkage techniques are much better than asking the cop to make a little check box off on an 18-year old girl in the middle of a crash whether she was pregnant or not, or to ask the EMS or the ED doctor to accurately do that on the medical records. I think the data linkage is the way to go.
CDC should incorporate better injury experience in the pregnancy risk assessment, PRAMS, there are some injury questions in there, but it's collected in a disorganized manner. Not necessarily looking at what the major causes are and much of the data is hard to tease apart. CDC should improve maternal injury details in their birth defects surveillance symptoms. If a baby comes out with a neurological acquired injury, it's a birth defect, but it's never listed as due to injury. And the NICHD should examine ways to study developmental outcomes to children exposed to in utero trauma. These are large expensive studies to follow children over a long period of time and those are the kind of things NIH should be doing.
Specifically to motor vehicles, states should continue to link birth and crash data. The Federal Highway Administration should add pregnancy status to driver behavior surveys so we can better understand why pregnant women are driving.
NTHSA should mandate pregnant crash dummy test to understand crash dynamics on women and fetuses. States should enhance education and belt laws to improve the use and proper use of seat belts by pregnant women and research needs to be done to explore ways to reduce pregnancy related driving, motor vehicle travel, crash risk and occupant protection. I'm going to leave you -- do I have three minutes? Okay. I'm going to leave you with this collage of maternal-related crashes I put together based on actual crash investigations.