AMCHP 2005 ANNUAL CONFERENCE
DELIVERING RESULTS, IMPROVING PREGNANCY & BIRTH
February 19-23, 2005
DAN THOMPSON: Okay. I'm going to talk about infant and fetal mortality, maternal age and multiple births in Florida. And this is a comparative analysis of births and the authors are listed there. One of the authors is with us today, Cheryl Clark. She's sitting right in front, so if there's any questions, you can address them to Cheryl.
Okay. We are going to look at two hypothesis concerning the recent increase in the infant death rate in Florida. And the first one is increasing rates of births to older mothers and mothers with multiple birth are associating with the increase in the infant death rate. And the second one is a shift may be occurring between fetal deaths and infant deaths and the neonatal deaths in particular.
Okay. So the reason we became concerned about this is because Florida's infant death rate reached an all time low at 7.0 per 100 live births in the year 2000. This increased to 7.3 in 2001 and went up again to 7.5 in 2002 and was 7.5 again in 2003. So it went from 7.0 up to 7.5 in 2003. And for many years, 20 or 30 years, it's always been going down. So we are real concerned about this.
Nationally, the infant death rate has also increased from 6.9 in 2000 to 6.8 in 2001 and the preliminary data we have for 2002 is 7.0 for 2002. That's for the US.
Here's a graph of the infant death rate per thousand live births for Florida versus the US, and you can see at the very end there we have reason for concern in Florida because the Florida rate seems to be branching off of the national rate and going up a little higher than the national rate. So again, we are real concerned about this in Florida.
Okay. The proportion of multiple births increased from 2.96 percent in 2000 to 3.01 percent in 2001. So multiple births have increased. A proportion of multiple births have increased. And the proportion of births to mothers age 36 to 50 also increase from 10.7 percent in 2000 to 11.20 percent in 2001.
Okay. Multiple births and older maternal age are both risk factors for infant death. So it's logical that increases and the prevalence of these factors would influence the infant death rate. So we generated this table based on birth data linked to infant death data and the infant death rate is higher for every category in 2001 compared to 2000, except for the 13 to 19 singleton category. That doesn't make a lot of sense to you before you see the table, so here is the table.
You can see in this table what we've done is we've taken maternal age and plurality and categorized them into ix categories. So these six categories are mutually exclusive and they cover every combination of maternal age and plurality. We have three maternal ages 13 to 19, 20 to 35 and 36 to 50. And for each of those ages, you have singleton births and you have plural births. So you put that altogether and you have six categories.
So you can see within each category, the infant death rate from 2000 compared to 2001 which is the next to last column there, the two columns on the end there, only in the first category, the 1 to 19 singleton births, did the rate go down. It went from 10.10 down to 9.61 from 2000 and 2001, and that's the only category. All the other categories went up. And you can see that in the rate ratio in the last column. It's always above one except for the first category.
Okay. The increase in infant death for the category for maternal age 20 to 35, singleton births, is particularly relevant since multiple in older maternal age are not factors in that category. So we are directly controlling for maternal age and singleton births and plurality. We are confining the analysis in that category to just singleton births and only mothers who are age 20 to 35. So older mom's and multiple births would not influence the rates in that category.
So this is a graph that shows you that 74 percent of the births are in the singleton births mothers age 20 to 35 category, so that's a big proportion of the births.
In that category, maternal age 20 to 35 singleton births, the infant death rate increased from 5.26 per 1000 births in 2000 up to 5.43 in 2001. So it went up in that category. Those are the two numbers from the table before.
Okay. So if the infant death rate in the maternal age, 20 to 35, single birth category had been the same in 2000 as it was -- in 2001 as it was in 2000, there would have been an expected 802 infant deaths in the category in 2001. That would be holding the rate constant with the same number of births. But the rate wasn't constant, it increased. And so because it increased, we had 829 actual infant deaths in that category instead of the 802 we would have had if the rate stated the same it was in 2000.
So there were 27 infant deaths more than expected in 2001 in the maternal age 20 to 35, singleton birth category. That's the difference between the two numbers. In comparison for women with multiple births and birth to women age 36 to 50 accounted for six infant deaths more than expected 2001. I didn't show the calculations on that one, but you can take my word for it now.
If all the infant death rates in all of the categories of women had been the same in 2001 as they were in 2000, there would have been an expected 1,390 infant deaths. But actually the number was 1,441, so that's 51 infant deaths more than expected.
So in summary, the expected differences between 2000 and 2001 in deaths, there were six additional deaths associated with mothers age 36 to 50 and mothers with multiple births. 27 additional deaths were associated with mothers age 20 to 35 singleton births. And overall total of 51 infant death deaths for all categories.
Okay. You really didn't have to follow those numbers at all because this graph sums up the whole thing. The lower line there shows that the rates are going up for singleton births to mothers age 20 to 35. And on this one, we are looking at more years. We are looking at 1998 to 2002 instead of just 2000 to 2001. So you see it goes from a low of 5.3 in 2000 up to 6.1 in 2002 and that's for singleton births for mothers age 20 to 35. All other births show the same pattern it's 11.5 in 2000 and it goes up to 12 in 2002. So the same thing is happening to both groups.
Okay. In summary the increase of multiple births and births to women age 36 to 50 influence the increase in infant death rates from 2000 to 2001; however, most of the increase was not a result of increases in multiple births or births to older mothers. So that's part of the problem, but it's not the whole problem. So that's basically the conclusion.
We have a paper written on this analysis and it's available on the internet at the Florida Department of Health web site. You go to the Florida Department of Health web site and then you go to maternal and child health section and -- but if you want to skip all that you can type in this link and it will get you right to the PDF, just the paper. And in the paper, I show you the calculations how I came up with all the -- how we came up with all the numbers.
Okay. The second thing we looked at is the possibility that a shift is occurring between infant death and neonatal deaths with neonatal deaths in particular. So the theoretical rational behind this is that improved medical technology enables earlier detection of fetal distress and this results in an opportunity for medical intervention so they deliver a live infant. This may result in increased deliveries of very low birth weight fetuses in distress. So the idea is that in previous years, you may have had -- the birth outcome may have been a fetal death, but because things are changing, the same outcome may now be a live birth. So if that was happening you would expect to see a decrease in the fetal deaths and a corresponding increase in the live births in the infant deaths. So pregnancies that previously may have ended in fetal demise may therefore shift to an infant death outcome.
Okay. The methods, we used the birth certificates linked to death certificates and the fetal death certificates also. And the births and fetal deaths occurring in 1999, 2000, 2001 and 2002, those are the data files we used.
So we see here that fetal neonatal and post neonatal death rates Florida in 1999 to 2002, the fetal deaths have gone down somewhat, especially from 2001 to 2002. The neonatal deaths increased from 2000 from 4.5 up to 5, and the post neonatal deaths have been pretty much level. So this lends a little bit of evidence to our theoretical -- to our hypothesis.
This is the neonatal death rate and the fetal death rate ratio for birth weights one to 1499 grams in 1999 to 2002. And here we see that the rate for neonatal deaths has gone up since 2000 from 151.1 up to 173.7 and the fetal death ratio has gone down from the high in 2000 of 253.2 down to 230.6. So there is more evidence for that -- for the hypothesis of the shift.
Okay. Now here is another bit of evidence is the neonatal deaths occurring within the first five minutes of life, these are rates for 10,000 live births in1999 through 2002. And if the shift was occurring, you would expect to see it in the neonatal deaths in the first five minutes of life. So we can see that's going up pretty good too. It went from 2.80 in 1999 up to 4.17 in 2002. So this is another bit of evidence for the fetal death infant death shift.
Okay. In summary, the fetal death rates are declining, the neonatal death rates are increasing and the post neonatal death rates are basically level. Okay. We haven't done -- we've basically begun on this, the fetal death infant death shift analysis, and there's a lot more we would like to do. We just have the descriptive statistics that I've shown you so far, but we would like to do other things more than just descriptive statistics. So our plans are to do some logistic progression models for neonatal deaths, and we are going to look at factors including measures of fetal distress and medical interventions and the survival analysis of infant deaths. And that's it.