AMCHP 2005 ANNUAL CONFERENCE
DELIVERING RESULTS, IMPROVING PREGNANCY & BIRTH
February 19-23, 2005
JENNIFER SKALA: All right. How many of you are familiar with the PPOR approach? All right. Quite a number. Could you even show me how familiar on a scale one to five, with one not at all and five very much so? Familiar? Just show with your hands. One through five. Wide range. Not very many ones though. So I won't take much time on explaining all the rates and how we got there. But at least I want to, during my presentation, give you enough. So you're able to understand the basics of the approach before Ohio and Florida tell you their stories.
So today, during this whole session, we're going to hit on these learning objectives. And again my part is to learn, teach you how to do the basics of the approach and make sure you don't have any questions, and then the Ohio and Florida stories will let you articulate how PPOR will be used or can be used to influence programs and policy as well as acknowledge methods to use to integrate data to yield better information. And then define strategies is really what this is all about, how states and locals can work together to build political will and stronger partnerships to address maternal health and prematurity.
So CityMatCH took this on. This is our mission: To improve the health and well‑being of urban women, children and families by strengthening public health organizations and leaders in their communities. Again, this happened back in 1996, initially when Dr. Bill Sappenfield came to CityMatCH from CDC and took an approach that was actually developed by Dr. Brian McCarthy with the World Health Organization that was being used in developing and developed countries and took about four years to revamp it, test it, pilot it and validate it for use in the United States, in urban communities. So that's why we say it's to be used at the urban level. Again, there's all kinds of approaches and strategies to use it at the state and local, but again it has to, with the way it was validated was for use in the urban communities.
In 2000, which Amy mentioned was when the practice collaborative was started that we took 12 cities across the country and made sure that we could take this full six‑step approach and put it into practice, make sure that we knew the best practices, make sure that it was really a valid approach that could make a difference. So, again, it's a full six‑step approach, and the reason I say that is because a lot of people just do the analysis and say, yes, I've done PPOR. And we want to make sure today you understand that the way it was designed was to take it straight from analysis to action. It's really for an MCHS epidemiologist, it's probably the simplest analytic approach ever. But again it's to be used with the community to move it to action.
So here are the six steps. These have been revised throughout the years. And most recently even, so that it aligns with the MCH planning cycle. So the steps that I'm going to talk about specifically today are the steps one, assure analytic and community readiness way before you even jump into the approach and how to actually conduct the analytic phases of the approach.
And then the Ohio and Florida story will not explicitly go through steps three through six but show you how it actually moved into action and how it's sustained to work on infant mortality in their communities.
So step one is to assure analytic and community readiness. Again, this has a lot of things that we learned through the practice collaborative, and we developed a checklist of all the things you need to consider before you even leap into doing the approach. It is the framework, really, that puts everyone on the same page to work on this together. And if you don't have that checklist or that readiness started at the beginning, then you miss a lot of things that need to really occur to move this to action.
Here's just a few things you need to consider analytically, to be ready. Your fetal death files; you have to investigate those. This approach doesn't just look at infant deaths but also includes the fetal deaths into the matrix of how it's displayed. So you have to investigate those make sure you don't have any restrictions on your gestational ages, because we advise on a cutoff of 24 weeks or less should not be included because of all the variability that we found when we ran the national numbers.
You have to have a link birth death file; and as Columbus will attest to, and Ohio , that doesn't exist at the state level currently or it did at one time but it does now because of this but it was a long process to take on. Those that are unlinked, those infant death files that are unlinked you want to investigate because you don't want to miss any information those are critical deaths that occurred; if they're missing out there, you have to investigate those. And then the forth item here is those critical number of events. And fortunately or unfortunately you have to have at least 60 deaths that have occurred in your time period that you're looking at. And, again, this is at a community level. We advise not to go over the three‑year time period, because systems change, services change, all kinds of things change in three years and it really has to tell you valuable information. So again because of the statistically significant numbers, they get too small if you go under 60 total deaths.
And again you have to make sure your maternal age and your maternal education is all available to you so what your data is telling you is the truth about what's going on in your community. So any of those poor quality issues in your vital statistics need to be investigated.
So that's a lot. Took a long time for us to just make sure our data was in order.
The community needs to be on board. This is crucial. While all the data is being looked at, you can get your community ready to do this. So here's just a few of the things that we have suggested. But there needs to be leadership and trained staff and they all need to be on the same page about what truly is the fetal infant mortality problem in your community. Then they develop the work plan together and understand it. They know what their roles and responsibilities are. They put up the resources as a full partnership and commit to this collaboration, and they champion it. Infant mortality has been around for a long time and it is so, this is a way of revitalizing that energy, by looking at it in a new way. So this is one of the tools that we put together. And you'll see the five tent poles of: Reasoning, roles, resources, risk and rewards and results. And this again has, it's a tool that asks the questions that a community needs to ask themselves before they take on this approach. And if you don't have a Palladian tent, there's not room for everybody to get into that tent. That's what we're saying. We make sure that if any of those tent poles are a little bit lower, you know exactly which direction you need to work on so you're all working together on it. So if your resources are not adequately fit to take on any action, you know right away let's work on that first.
So that's step one, just all the readiness work that needs to happen. I'm sorry if I'm talking fast, but we usually do this in one and a half day workshop to train people how to do PPOR and I'm trying to do it in about 15 minutes.
So bear with me. We'll be around afterwards for any questions you have about the basics.
Here's the analytic phases. There's actually two analytic phase of the PPR approach. The reason these phases are different and the way we put them, we want to make sure you know it includes fetal and infant deaths again with the cutoff of 24 weeks gestation, focuses on those very low birth weight babies, the really tiny babies, and gives emphasis to those and be able to discover what you need to do about those. It generates a map, which I'll show you in a minute, first targeting strategic action. It puts everyone on the same page and leaps you right into action.
Again, it examines birth weight and gestational age at the same time. Different than normal infant mortality rates. This is how a lot of times we see infant mortality rates at CityMatCH is it comes across as just the rate over a time period broken down by race and ethnicity. It tells you what's going on with those subgroups and those races and ethnicities, but it doesn't tell you what you need to do about it. So, again, this is the two dimensions of birth weight along the side and age of death at the top. And it forms these four periods of risk. So you have the maternal health prematurity, the blue box across the top. You're looking at those, again, those under 1500 gram babies, those tiny babies, across the continuum of the fetal death period which is 24 weeks to birth, the first column; the neonatal period, birth to a month; and the postnatal period, which is the month to a year. So all those deaths that occur along that continuum that are tiny births we're saying all should be attributed to maternal health prematurity. The labels on these four periods of risk are suggesting prevention. So maternal care are those larger babies over 1500 grams, these babies that are in vitro that fall in the fetal death time period, again 24 weeks to birth. So the pink box you know right away you need to handle issues of maternal care if that's where your rates are falling, if that's where the numbers of deaths are occurring. Newborn care is the yellow box, the neonatal deaths that are birth to a month that are again 15, the larger babies, 15 grams or more, 1500 grams or more, and then infant health deaths which are a lot of times the SIDS deaths that you're familiar with that happen from a month to a year of age.
So that's the basics. Then we want to make sure you get that before we go any further. Is there any questions about that?
UNIDENTIFIED SPEAKER: It's going to sound like a picky question but in terms of data, it's not, when you say one month, are you defining it as 28 days or ‑‑
JENNIFER SKALA: Actually, 28 days, yes.
UNIDENTIFIED SPEAKER: I wondered why, (inaudible) we do the 20 to 24 weekers too, why is it 24 weeks?
JENNIFER SKALA: When we did the validation of the model, we looked at all the deaths and infant births in the country and linked files using MCHS files. There was way too much variability between the 20 and 24 weekers to be able to say that those are attributed to that cell. They're just for validity reasons. But a lot of communities know what's happening in those 20 to 24 weeks. What this model doesn't include are any spontaneous or induced abortions. So that's another reason. There was just too much unknown going on between 20 and 24 weeks to include it into this and say we can stand behind it. But, like I said, a lot of communities know what's going on, and they include 20 to 24.
Okay. Here's how it moves a community to action. It takes, if most of your deaths are occurring maternal health prematurity, you know right away you should be looking into preconception health issues or the health behaviors of the mother or sometimes into the perinatal care system. And I'll explain the second phase of the approach which is a Kittagoian analysis, or further investigation that actually leads you to go one way or the other in your prevention strategies. Maternal care suggests you should go look into the prenatal care system or high risk referrals or obstetric care. The newborn care ‑‑ the yellow is hard to see, I'm sorry about that ‑‑ and the perinatal management, neonatal care or pediatric surgery issues that need to be looked into. And infant health again, sleep, physician, SIDS, congenital anomalies, injury, breast feeding, all those things that we have been doing a really good job about. Again, this puts everyone on the same page to be able to say let's go in this direction right away.
What we want to make sure though is that you just don't stop there. It's the best way to get everyone on the same page but then you ask all the questions that you need to ask to go through both phases. The first phase is identifying the populations with overly high numbers and rates of mortality that we call this the MAP, which I just showed you. Then I'll show you how to look at the GAP analysis. Phase two explains why the excess deaths are really occurring. That's where a FIMR comes in. That's why it's important to do these in conjunction with each other.
So I heard Amy mention we have lots of handouts, but we wanted to make sure you walked away with the exercise that we actually do in our training so that you can get your hands around the number, because oftentimes you just let your data person run the numbers, but it's important for everyone to really understand this and get ‑‑ it's simple. And so we gave you the answer sheet and we gave you everything you need to know about the numbers. And this is true data in an urban county that we used as an example. So if you have that sheet available, you'll see you have this first page that is table one. And you look at something like the blue box, the maternal health prematurity category. And you say we don't look at under 500 gram babies, and we don't look at under 24 week babies that die. So you get rid of automatically the first column and the first row of that table, of the numbers. So then you're able to just add 35 plus 57, plus five, and you get 97 deaths. Those are both fetal and infant deaths.
The maternal care, you just looked at the pink box. You look at just those that are 24 weeks or more in the fetal death category so it's that second column that are 15 grams or higher, and you add those together. So you have 34 plus 14 equals 48. I don't want to lose anyone here, but just know that you can, the numbers are actually in your answer key and they're also in this presentation, so you can flip back and forth from it. But, again, if you want to stay afterwards and go through the exercise, it's important to understand why it's different. And from the number of fours and fives I saw on the hands, I know a lot of you are already familiar with it. The important thing is the 97, 48, 44 in the newborn care and 47 in infant health all equal the 236 fetal infant deaths.
You times that by a thousand and you divide it by the denominator which is the live births and fetal, which is 23,282, you get the 10.2 overall rate. Then you also see that all the rates within each period of risk also equals 10.2. Sometimes we have some rounding issues. But that's close enough to see how you can contribute what is contributing most to your overall rate, and in this case you can see maternal health prematurity followed by the maternal care in this urban county and as well by infant health.
Okay. Are we all right? We're now onto another way of using just the MAP, is you can break it down by race and ethnicity group. You have the white non‑Hispanic rate of 8.6, and you do the same numbers for the black non‑Hispanic rate which is 17.6. Automatically you can say my goodness what is going on, it's not surprising to any of the people that have been working with us for a number of years, but again to the community and to the people that you bring on board, you can automatically say why is the rate more than double for black non‑Hispanic versus white non‑Hispanic. The same in each category it was alarming to see even the infant health, the green cell, the bottom to the right, 1.6 for white non‑Hispanic versus the 4.0for the black non‑Hispanic. So again leads you straight to what you need to do first. Not that you need to conquer it all. Wish we could, but at least tells you what you can do first.
So now it's about the GAP analysis. You first ask which women and infants do have the best outcomes. You assume that all infants in this country should have similar best outcomes. And you then choose a comparison group which we actually call a reference group and who have already achieved those best outcomes, and we compare those to your target population in your own community or state.
So then you calculate the excess deaths by looking at the target group minusing the comparison or reference group and that's your opportunity GAP.
For external reference group which the practice collaborative used we developed one using all, first, deaths that occurred in the United States . Right here it says for urban cities and counties that bottom bullet, but actually this, we updated it so states can use this reference group as well. It's all births that were residents of the United States .
The maternal characteristics that we used were 20 or more years of age, 13 or more years of education, and of non‑Hispanic women. Again, those were the best outcomes we saw. Equaled a rate of 5.9. You remember just now, it was 10.2 for our urban county. Now we can automatically compare it to this best outcome group of 5.9. I'll show you how we do that. You just lay out your rates for urban county along the top and subtract your external reference group, and then you subtract them to show that all the excess rates that are occurring overall is 4.3. Maternal health prematurity is 2.0, which is the largest contributor. But there's still something going on with maternal care in this urban county, as well as newborn care and infant health. But again focusing on maternal health maturity, which is the hardest one to conquer, as we all know, is the first one we need to focus on. But this urban county is actually Omaha , Nebraska , which is where CityMatCH ‑‑ we always reveal that, I think. But we actually devised a plan to conquer each one or address each one. Maybe not conquer but address.
You can also do this by your race and ethnicity subgroups. You can show the white, lay out all the rates across the top. Your black, non‑Hispanic rates and then your Hispanic and other races and compare that to your external reference group. Then on this next slide you'll show the excess in each of those categories.
So right away you can see again these other races what's going on in maternal health prematurity compared to your white non‑Hispanic group. You can look at your infant health group; blacks, it's way more than the other rates, and so on and so forth.
That's a lot of numbers. I just wanted to show in the exercise it actually shows you how to convert these rates then to numbers so you can say, because you can only compare rates to each other. You can't compare numbers to each other so it's important to get to the numbers because that's what people understand in your communities is to say you could actually have done something different or maybe saved this number of deaths. So in the white area if you had the rate of 2.8, and the numbers are actually given in your exercise, and you times that by the live births and fetal deaths, divide that by a thousand, then you get 45 actual numbers of deaths that maybe could have been saved. 101 overall in this urban county that we need to be working on. And there's the numbers broken down by race and ethnicity.
The one thing I want to point out is it's not just important to do an external reference group. A lot of times in our practice collaborative they said it's like comparing apples to oranges where that's not what's going on in our community why would we use the best outcomes happening all over the country. It's important to do an internal reference group. You can see here, using the external reference group you'll remember there was a problem in all of the categories, because we really do want to compare us to the best outcomes that can be happening. But in our own community there's really something going on, because, look, there's no excess in maternal care, or in newborn care. And so if we want to just look at the internal reference group, we would have not known to do anything about either of those. But we've been finding, investigating even more in our further analysis that we really need to be doing something about the maternal care issues of actually white non‑Hispanic women. And that would have been completely missed had we just done an external reference group.
And there's all the numbers for the rest of the exercise. And since I have only four minutes left, I just want to make sure you know: You don't end there. You don't end there. You actually take this to a phase 2 analysis, which includes more in depth targeted interventions and actions that include a number of analyses. For instance, maternal health prematurity. This explains why the excess is occurring. You do what you call a Kittagowa analysis. That actually looks at birth weight specific mortality versus birth weight distribution, which I know some of the sessions have already addressed. Is it women's health issues or is it NICU survival issues. And in urban county, it's women's health, 70 percent. But 30 percent, something is going on once those babies are actually born that they're dying in the NICU's and it leads you to straight to action.
Causes of death are those ICD‑10 and used to be 9 codes when we started this out that leads you to sleep position, as well as congenital anomalies and injuries, and you actually have that information and you can just attribute the numbers accordingly.
So without that, you don't know exactly what to do. So it's so important to integrate all these other key efforts accordingly. And, most importantly, this fetal infant mortality review we found to be most beneficial. It revitalizes the FIMRs to be able to paint the faces behind the numbers, to be able to get the qualitative information is very much important.
This is what we found through the practice collaborative. You can read it later. It's about capacity building and partnerships and evidence based. It's a new way of looking at fetal infant mortality to promote change for perinatal health. So that's the first two steps of the approach in 15 minutes, and I hope you stay after if you have any questions.
But step three through six will be addressed in the two stories that we have to bring to you through the practice collaborative, which we promoted the state local partnership in Ohio that really did build better data capacity, and then in Florida they integrated, they have so much going on and they use this to integrate all their efforts to be able to prevent, to better prevent fetal infant deaths. Is there time for questions? Any questions? At this time?
Okay. Thank you.