AMCHP 2006 ANNUAL CONFERENCE
EARLY CHILDHOOD: BUILDING THE FOUNDATION FOR LIFELONG HEALTH
March 4-8, 2006
VIOLANDA I. GRIGORESCU: Good morning. Before I became a State epidemiologist, I worked at county level. And I had a chance to work with City Match. I loved their workshop and training. When we were given kind of assumptions, or type of fake-up situations, if you like. So, let’s say, let’s assume, if somebody to your today, your director or your health Commissioner, and asks you, what happened in your state and how is working your program, Safe Sleep program, or what happens with SIDS, and you have to give an answer in five minutes. What do you do? And you may not have access to medical examiner data, but then you have to find some data sources.
As you know, we have different data sources with different information, and then you turn to epidemiologist and say “Now help me, and tell me, what should I reply?” I told you this just because I’d like you to understand that in Michigan, when we began looking to this issue, this is how actually came about. Our MCH director was wondering what happened in Michigan and how can we understand better the decreasing seats, and if our programs had any impact. So we began looking at this I would say for more than one year now. What I’m going to share with you today is just a part of our work. But I’d like to show you what you do actually, what data sources you can use to have an understanding of what happened when you don’t have access to medical examiners’ information and you don’t know senior investigations data.
So, these are objectives for our presentation today, and I’m going to skip this. I’d like to go to the problem. Michigan was unsuccessful in significantly reducing infant mortality. And I’m saying 96, 2003, because you’ll understand from my graph, we were able to decrease infant mortality from 1999 to 1996, but then it was steady around eight with small fluctuations, and in 2003 we actually noticed an increase from 8.1 to 8.5. Always we’re above nation, infant mortality. You know there are two components, neonatal death and post-neonatal death. So will always look at those two to see what we are doing in each of them.
Neonatal recorded fluctuations over time, but post-neonatal was steady, especially from 1999. It didn’t have much changes, and it didn’t change even in 2003, so we were like having the same post-neonatal mortality around 2.5 before, like with small fluctuations, while neonatal noticed different changes. And we also addressed both. And we have different projects going on, trying to understand both, neonatal and post-neonatal. But I’m going to address more post-neonatal today. SIDS rate declined 71% since 1994. And these graphs actually depict very clearly that from 2002 to 2003, when we noticed an increasing overall infant mortality, we noticed an important decrease in SIDS, almost half decrease, overall, and by race.
So the question was, where did SIDS death go? What happened? If we have a steady post-neonatal mortality, and we have such a significant SIDS decrease, something probably happened. Those SIDS death who don’t show up anymore at SIDS must be in other groups. We must have a shift in diagnostic, it was assumed. So how can we explore this further and understand what actually happened? Of course we did use the link file. This was the first thought, what data we have access to. The link file death cohort link to lifers. And we did look to ICD10 codes, as we know, and we used National Center for Health Studies and Groupings, trying to understand what happened in post-neonatal. And I’m talking just about post-neonatal debts because as we all know, SIDS is the main causing post-neonatal time.
(Inaudible) did look and these are just couple groups, not all of them, we did look to see if we understand where we have decreased, where we have increased. I mean, sorry. We noticed we have a significant decrease that we mentioned in SIDS. We have some increases in other groups, and I’m having my conclusions here. Forty-five less SIDS deaths. Some increase in some groups, like digestive heart and secretory systems, and determine death, a little bit of increase. We have less death in other groups. But we didn’t find any increase in suffocation death, regardless of different feedbacks we received from our CDR group, Child Death Review people, who kept reminding us that we have an increase in suffocation death. And in fact, those studies (inaudible) didn’t show up when we looked to ICD10 codes.
So conclusion was, the simple math, looking to these ICD9 codes and trying to get that (inaudible), it didn’t help us, if we look through this. What codes we did use, and what groups we should be focused more on. Based on some prior work and based on our work with CDR actually at some point that I don’t have with me today, but we ended up matching and trying to look at both data sets and understand how groups in child death rate may match with vital statistic groupings.
These are (inaudible), and you all know ICD10 codes for SIDS as well as accidental suffocation in bed as well as undetermined death. So we thought, did we use the right ICD10 codes for these groups? Because based on our work with CDR, we identify some of death, called it a suffocation but not having the known W-75 code for accidental suffocation in bed. So the question is that probably there are some algorithms we have to develop to understand how we would be able to look at vital statistics using the information we have, but including all codes we may need to include to better understand if there is a shift and what the shift is.
And in this simple table, I just wanted to share with you like what accidental suffocation, let’s say, can be a suffocation if it’s confinal tapped, with is other W code. It can be other unspecified threat to breathing. Yes, it can be coded as such too. And you all know that in vital statistics, information comes from death certificate. So it’s a matter of how even medical examiners record the cause of death in vital statistics for us to be able to use it and to be able to offer a trend or to have an understanding what happened for the entire state all over all the entire state. So it’s very important to look further and to explore further these type of coding issues, if you like. And we are in a process and trying to understand from CPR how we can do this grouping in a different way and trying to use ICD10 codes in a more useful manner. So yes, it was a linked file. We all know. We have, Michigan is very strong in having different data sources and good people in analyzing different data.
But the question is if these ICD10 codes as I said, tell us the true story. It was very difficult to explain. And we are still working in this collaborative effort with CDR, trying to understand what would be the best ICD10 codes we should be using to understand what happened in Michigan. So we need more information, we need to explore further, we need to even understand scene investigation findings. But then, that is going to take time. And as I told you at the beginning, you have a question for (inaudible), you have to find some answers to give back to your director.
So, what we would say about SIDS, what we would say about accidental suffocation, what we would say about safe sleep. So then we thought that this can be found in another data source. It’s not just vital statistics. And as I said, we have different data sources, and we have to be smart enough to be able to use for different answers we have to give. So then, we understood then we can use PRAMS to understand the safe sleep in what women believe about or know about this. I don’t know how many of you know about PRAMS, but I assume the many of you already know and have used PRAMS. But I thought I would give you a little brief update on Pregnancy Risk Assessment Monitoring System, which is PRAMS. It’s a surveillance project of the Centers for Disease Control and Prevention in State Health Departments, and collects state-specific population days’ data for maternal attitudes and experience before, during, and shortly after pregnancy. And we use it in Michigan quite a bit.
PRAMS’ sample of women who have had a recent live birth and assemblies run from the state’s per certificate file. We in Michigan, each participating state, for example, between 1,300 and 3,400 women, we sample between 2,000, 3,000 women. And we have assembled some groups of women, and in Michigan, has been a change in 2000. We did sample for low birth weight and race, and now we sample for race and geographical distribution. Geographical location. It is a mail first, followed by a phone interview. You can find all information about PRAMS at CDC Website, CDC PRAMS. I like to talk a little that about question now because you have to understand what question we did use to understand the sleeping behaviors in the Michigan.
The original PRAMS questionnaire was developed in 1987. It was revised many times, and in April, 2004, the fifth phase was implemented. This is what we use now. The questionnaire consist of two parts, and we are going to use, I’m going to share with you what we used in Michigan, two type of questions. Core questions and state-added questions. State-added questions can be standard questions, developed by CDC, and we borrow those questions we feel like are more important for our states, or each state can develop their own questions.
So, we use one core question. We say how do you most often lay your baby down to sleep now? On his or her side, on his or her back, on his or her stomach? We found out that more than 70% of women in Michigan reported they’re laying back their baby. Which was good sign. But then of course we did different certification looking by age, and I’m not going to share with you everything we’ve done using PRAMS for this. I just want to show how you can use this information to find out more about behaviors in your state. So the group with the highest prevalence of having baby laying back was in 30, 35 years. We assume these are women who are, you know, more mature to understand better probably the message you send it out. The lowest prevalence was in very young women below 18 years of age.
When I looked to race ethnicity, it was a little bit different I would say than some findings Melinda shared with you. In black and Hispanic women in Michigan, they have the lowest prevalence of laying baby back. But then over time, when we looked at these trends, we can see an increase overall, and an increase by race in Michigan, a constant trend up of laying back babies to sleep, which means, or efforts have the results. And we have more African-American women who actually learn. But, if you’ll look to the difference between white and black, there always is about 20% difference. So they keep growing. But we were not able to close the gap.
Now we look to another question, standup question. How often does your new baby sleep in the same bed with you or anyone else? Always, almost always, sometimes, rarely, never? We decided to group these responses because the numbers were too small if you use individual responses. So the majority, the way we group, the majority showed up to be rarely, never. More than 50% say rarely never bed-share. Now when we look by age, of course the same age group were thought as being the most prevalent, or the group who better understand I would say, the messages, and they reported really never a higher percent, and the lowest was again below 18 years of age.
We look by race, ethnicity again, and look, is like a reflection, if you like, of the prior question, the higher prevalence of those who say always bed-shared was in black, not Hispanic women. By education, if we look to this specific question, the highest prevalence of those who said always bed-shared is for those who had below high school education. So uneducated black, white, black and non-Hispanic women, and younger are the groups that needs to be more targeted.
So by using all this information, I shared, then especially using PRAMS information I shared with you, we conclude that we learn that Back to Sleep Campaign in Michigan has changed, has improved the behavior of many mothers, and they put infant back to sleep. However, we learned that though still we have groups, younger women, uneducated, and black non-Hispanic women, who need a different message. We need to be more effective in sending the message out to these groups of women. And we decide that we should explore further the curriculum we had in Michigan for the Safe Sleep Campaign, trying to address these specific groups of women.
That was really timely. Because in Michigan as I said, being concerned of the increasing infant mortality, the steady post-neonatal death and decreasing SIDS, the question was, how are we doing and what we should be doing. So the information that we were able to provide were really in time for this Infant Safe Sleep Campaign that I guess Sandra is going to share with you a little bit more. And yeah, as we learned, as I said, what groups are at high risk and what groups should be more targeted, and we learned when, in what groups the high prevalence of riskiest behavior still exist.
So, the whole message of what I share with you today is that we have data resources available. Different data sources. And we have to use our resources in the best way to understand what issue we have in our state. Because each state has something specific. In Oklahoma they found out whites have more bed-sharing issues. In our data in Michigan, we find out that we have more black, non-Hispanic women. And yes, as I said, we continue to work with CDR to better understand the information they can provide from scene investigation.
But again, as Melinda pointed out, we don’t know how biased that is, depending on the time when the scene investigation is performed, depending on the time investigation is. So we have to allow us to look at these diverse sources of information to understand which can offer what, and to best use the information and data we have for our programs and prevention strategies. I’d like to acknowledge our epidemiologists, who work with me on PRAMS, as well as our PRAMS coordinator. And I’d like to acknowledge everybody who was helping, as in Michigan, to move forward with all this so-called investigation. And I think you all passed the test. I said at the beginning, what you say to your health department director when he ask you what happened in your state. Thank you.
And I turn to Sandra, I turn to Sandra and she’s going to talk with you more about what happened in Michigan based on the information we shared with here. Thank you.