AMCHP 2006 ANNUAL CONFERENCE
EARLY CHILDHOOD: BUILDING THE FOUNDATION FOR LIFELONG HEALTH
March 4-8, 2006
EMILY LU: Let me try to speak into the microphone, since my English is not. Beth thank you for giving us such a great background, and I really want to thank them too this is giving me an opportunity to hear more about Massachusetts, and I am a CST fellow assigned to Massachusetts. And so Beth already mentioned the study question. I just want to go through the source of data that we have in Massachusetts.
Source of data. Currently for this project we use three different data sources. MASCHP which is Massachusetts Community Health Information Profile. The second data set is coming from Vital Registry. Third coming from community sources.
MASCHP is a publicly accessible information data system that anyone can download just go to our department website and it's very easy. There are two types of reports that's available on MASCHP. One is instant topics, that includes perinatal data that has birth outcomes for specific communities in Massachusetts. And the second type, the custom reports, here is where you can actually go in and do your own, create your own specialized reports.
Vital Registry contains data from live births, link live to infant death files as well as fetal deaths.
A wealth of information is also available in the community, and this includes qualitative data such as oral histories, key informant interviews, focus groups, as well as historical records and community reports.
So we use various methods listed here to try to help us answer our study question. Those include analytical methods and process methods. Analytical methods we are using to try to help us understand the community and to frame the issue. The process method is helpful to help us contextualize the data.
Analytical methods we use quantitative and qualitative methods. Quantitative methods include bivariant analysis, population attributable risk. Perinatal period risk approach and qualitative method includes focus groups. Bivariant analysis, it's an analytical tool that enables us to explore the relationship, the association between two variables. And I want to show you guys some data from here on.
First this is a histogram showing racial disparities in birth outcomes in Springfield, Massachusetts and the birth outcomes will focus on ‑‑ you can see it's on the X axis.
Very preterm births is being defined as less than 32 weeks gestational age. Preterm births which Beth already defined is less than 37 weeks gestational age. Very low birth weight is less than 1500 grams. Low birth weight is less than 2500 grams. And fetal infant deaths goes up to one years of age.
And the Y axis is showing the percent of these average birth outcomes. And you can see that there is a large disparity that exists between black and white for very preterm births, the percent of very preterm births for black is about 2.2 times higher than what's observed in white. For preterm birth, it's about 1.5 times compared black to white. For very low birth weight, it's about twice as high as when you compare black to white.
For low birth weight it's about 1.3 times higher. And for fetal infant mortality, it's about three times higher for blacks as compared to white.
So here we identify the disparity. On the next result I'm showing we're just focusing on preterm birth by selecting maternal characteristics. And here you can see on the X axis are the selected maternal characteristics that we focused on that includes education, maternal education, maternal age. Insurance. Prenatal care, as well as smoking. And you can see from here that except prenatal care, higher percent of preterm births is associated with mothers having less than high school education, younger than 20 years old, being on public insurance and smoking during pregnancy.
For prenatal care it doesn't seem to make as big a difference on the preterm birth in Springfield, Massachusetts.
Next slide here we are showing again selective maternal characteristics by race and ethnicity in Springfield Massachusetts. And here you can see that we also, we broke it into two strata, black and white and also looking at the selected maternal characteristics.
Here we actually are exploring other factors that might be associated with average birth outcomes other than race and ethnicity. So here you can see that black moms were more likely to be having less than high school education, being younger than 20 years of age. More likely to be on public insurance and more likely to receive less than adequate prenatal care. And except smoking seemed to be, there seemed to be a lower rate of smoking in black communities, in black populations than white. So this kind of gives us, there may be other confounding factors besides being black that's affecting giving them average birth outcomes.
Here, since we decided to use another statistical tool to help us understand better of the poor birth outcomes in Springfield Massachusetts. Population attributable risks is defined as a statistics that shows the percent of poor birth outcomes that could be reduced if the access risks of poor birth outcome among one group was reduced to the same level risk of a poor birth outcome in other group.
Here is the results we have about PPOR for very low birth weight. PAR. Sorry. PAR. I just wanted to make a comment on PAR. So PAR is actually very interesting tool that allow us to both measure the degree of association between the factor and the outcome that we're interested in as well as taking into account the prevalence of the factor. So, for example, one example that Karen used a lot is there's a strong association, strong relative risk associated with mothers using crack cocaine and having delivered preterm birth but as we all know the prevalence of crack cocaine users is probably pretty low so therefore the PAR when you take into account both, it will be pretty insignificant.
So here you can see that out of, oh, the risks indicators, being black actually had the, actually the highest risk for very low birth weight is contributed for being black, and that's 36.2%. And followed by being Hispanic, on public health insurance and so on and so forth. So, in other words, if we could reduce the access risk associated with being black to the same level of their reference group, we could reduce very low birth weight by 36.2%.
Next I just want a show of hands if any of you have ever used perinatal peer risk assessment approach in your state. Okay. Great. So it's basically a map of fetal infant mortality using birth weight and timing of death. And you can see from here that using birth weight and timing of death we can divide the infant mortality into four areas of possible intervention areas, and those are maternal health or prematurity, maternal care, newborn care and infant care.
Here we further have broken it down into let's say if ‑‑ let's say, for example, your community identified that they would like to work on improving maternal health, then there are a couple of intervention focuses that you can think of, such as pre-conceptional health, perinatal care and same thing so on so forth for other areas to focus your intervention programs on.
Here I'm showing the results of our PPOR analysis, and overall the fetal infant mortality in Springfield, Massachusetts is over the last five years is it's 8.4 deaths per thousand live births plus infant death. And here you can see that the highest number of deaths is rate of deaths is found in the maternal health prematurity box, followed by infant health 2.6 per thousand. And the other two boxes had very small number of deaths.
And that could be due to ‑‑ so the other two is maternal care and newborn care, and this could be attributed to having really good obstetrical care and really good neonatal intensive care management in Massachusetts hospitals.
Here we decided that we needed to use as part of the PPOR analysis that we should use a reference group that will help us do a benchmark and measure our improvement against reducing excess deaths in fetal infant mortality here we're trying to find a reference group that's achievable and acceptable for the community. Remember, this is really ‑‑ we really want to give the community ownership of taking this and mobilizing and focus interventions that they think will work for them.
So for this case, our reference group we decided to use Massachusetts resident mothers who are non-Hispanic white, who have more than ‑‑ who have 13 years of education or more who are older than 20 years of age. And here you can see for the reference group overall we had 4.2 deaths per thousand for reference, for the reference group and compared to the blacks in Springfield, Massachusetts had 12.7 deaths per thousand live births plus fetal death. And subtracting that, the excess death was 8.5 per thousand.
And it's also ‑‑ I also want to point out the highest access deaths still in the maternal health prematurity Box 3.7 followed by infant health, 2.9 per thousand births.
Next we use the same internal reference group and compare that using, compare that with the white populations in Springfield, Massachusetts. And the excess deaths is lower. It's 2.5 per thousand birth, and interestingly, infant health is the highest box that had highest deaths for this group.
Here in summary, we are basically showing what I just presented now in the histogram format, comparing the disparities between black and white and you can see that the disparity does exist in all four areas and most prominently in the maternal health section.
Although it's very interesting that the infant health, in the infant health section, both black and white had comparable number of excess deaths in the infants.
Here I just want to say this is a more of a theoretical mind set, helping the community quantify the burden of fetal infant deaths. So we have the excess rate of deaths per thousand from this community and we also know the total number of live birth and fetal deaths in Massachusetts, so we could actually estimate the number of deaths we could prevent if we implement, concentrate our interventions that address the disparities in Springfield, Massachusetts.
At this point, this ends my portion of the talk and I'm going to give the podium to my colleague, Karen Downs.