AMCHP 2005 ANNUAL CONFERENCE
DELIVERING RESULTS, IMPROVING PREGNANCY & BIRTH
February 19-23, 2005
BETSY WOOD: I also thank you for letting me share some ways that Florida has used FIMR, the FIMR process as a touch point for some additional community activities that have been going on.
As an overview, the Florida Department of Health course is charged with the tracking and monitoring the statewide maternal and child health outcomes. But we have a very rich and diverse makeup and it's very necessary for local communities to develop and maintain their own expertise for identifying factors that influence these outcomes.
Just as an overview, the key partners in Florida's MCH program are 67 county health departments along with statutorily mandated healthy start infant and prenatal coalitions that provide oversight for our MCH symptoms of care. We have 31 coalitions around the State that cover 65 of our 67 counties and in the two counties that don't have coalitions, their health departments oversee the symptoms of care there
In this session, I'd like to talk about the types of analyses that we routinely perform with statewide data, the use of these analyses that identify areas throughout the State that we targeted for additional technical assistance the initial work with one local community subsequent development of our infant mortality technical assistance teams, and finally I want to talk about how one of these counties, using information that they got from the technical assistance teams, develop their own local evaluation of infant mortality.
In 2000, we reached our best infant mortality rate that we ever had and that was 7.0 deaths per thousand. However, 2001, our infant mortality rate went up to 7.3 and in 2002 and 2003 it went up to 7.5 and stayed at 7.5 last year or in 2003. This graph shows that Florida's rates are fairly close to national rates.
And as you can see in 2001, the national rate began to curve upward just like Florida's did and of course that's very concerning to all of us.
As we are looking at these changes, it's very important for us to consider trends among the categories that make up infant mortality. Both our neonatal and post neonatal death rates as well as the trends in our fetal death ratio for the same time period. And in Florida, our fetal deaths are defined deaths to a fetus at or about 20 weeks gestation.
Our data and evaluation team look at all these, use all these methods to look at and try to understand our outcomes and why they are going the way they are. And I want to address each one individually now.
Florida's used the PPR approach to analyze fetal and infant death both at our state level and individual county levels. Over the five-year period that we've been using PPOR, the findings have remained consistent and just as many other places, it suggests that the most critical category for improvements lies in maternal health.
So with these findings, we've been looking at ways to provide pre and inner conceptional services to our population, especially for women who experienced poor birth outcomes. And in the last year, we developed standards and guidelines for our Healthy Start programs as well as technical assistant guidelines for our county health departments in delivering inner conceptional education and counseling for most at risk women.
We also look at birth weight specific mortality. Any increase in the percentage of infants, our smallest infants born have the greatest influence of infant mortality. In addition, changes in the infant mortality rate within each category can be influential as well. So we need to look at both of these factors to see about -- to tease out clues for influencing factors in our infant mortality rates.
We also look at cause of death trend analysis. The leading causes of death for infants have remained pretty constant with some noted decline in our sudden infant death syndrome.
And this is a five year look at our cause of death with of course perinatal conditions and congenital and anonymous lease continue to be the -- cause the majority of deaths in Florida. We need to continue to emphasize the necessity of addressing health for women prior to become pregnant. While many of the underlying factors of prematurity are unknown, we still need to continue to have our public health focus on smoking cessation, taking folic acid and our communities continue to do this.
Our SIDS rates are very slowly declining, and while there are issues regarding classification of SIDS rates, our communities are working to increase public awareness through campaigns such as our Back to Sleep and save sleeping campaigns. And we are confident that part of this decrease is due to these community efforts.
And as Dr. Wise said GIS mapping can be a very powerful tool. This shows our infant mortality rates with the darkest areas being the areas of the highest -- with the highest infant mortality. This map was developed using our chart system, which is a web based community health assessment resource tool which is out on our Florida internet and it's available for all communities and it's a rich data source for communities. However you've got to learn how to do the mapping and the geocoding.
In addition, each year through the university, a contract with the University of Florida, we develop an official matched file which links birth certificate, fetal and infant death certificate information Healthy Start prenatal and infant screenings. Our Healthy Start prenatal services, Medicaid eligibility file, WIC services and census track information. This file was created by county and has been available for a couple of years for local use but all counties and all coalitions are not aware and not real experienced in using that file.
In addition, each year the department prepares an analysis of infant death and low birth weight by county and coalition area. It provides an adjustment for the demographic areas of race, marital status and education. This allows for a comparison among regions that takes into account the high risk population residing in each of these regions. It's called our actual versus expected analysis, and Dan Thompson who actually created that is here in the audience. The purpose of the analysis is to identify geographic areas in the State where the low birth rate and the infant mortality rate are statistically higher than we would have suspected. And that way, we could target possible interventions in those particular counties.
As I said, in 2000, we had our best infant mortality, rate and it began to go up in 2000. And in 2002, our secretary said, um, talk to me about our infant mortality rates going up. So we used the trend analysis and actual versus expected analysis, and we identified ten counties that might need additional services or support, and our secretary had telephone conversations with the executive directors of our Healthy Start coalitions and our deputy state health office had similar conversations with our health departments in those ten counties, and that is the boss of the Health Department directors in those counties. And basically, they ask, tell me what's happening in your county. Why with are the infant mortality rates either trending up or higher than I would have expected them to be? And as you can imagine , there were several counties who said, I know why, this is happening and this is happening and it may be due to this and it may be due to that. And they had ideas and potential ideas about why this happened. And then there were other counties who said, not really sure.
So it really pointed out that we had a great discrepancy in our data capacity locally right about after that happened one of our Healthy Start commission in one of the counties noticed a spike in infant mortality and they requested on site assistance to review the infant and fetal deaths in that community. So we used a slightly modify FIMR process pulling together state and local experts. The partners including the local Healthy Start commission, the local Health Department and the Department of Health staff. We look at the vital stats data for all of 2002 and identified all known fetal and infant deaths. We developed an abstraction tool that was slightly modified from the normal FIMR abstraction tool. And then our Healthy Start coalitions were able to set up meetings with the local hospitals and schedule the abstractions. We got copies of the birth and death certificates and abstracted records from the hospitals as well as from the medical examiner, and they were both local hospitals as well as the regional referral center that provided high risk obstetrical and neonatal intensive care for that county. They abstracted 67 cases in a week.
They looked at the trends in fetal and infant mortality and birth weight specific mortality, trends in prematurity, very low birth weight and low birth weight, trends in maternal age, congenital malformation and causes of death. They categorize causes of death using the ICD 10 codes and they geocoded the deaths to look for potential geographic clustering.
The findings from this project included that we needed to have an emphasis in the importance of reviewing these indicators on a more frequent basis than what we had -- than an annual basis. There were indeed shifts occurring in birth weight categories however the shift didn't occur in the same category for the fetal deaths and the infant deaths. There was a hypothesis that much of this was because of amiss classification or fetal death versus life birth and that did not bear out.
The geographic mapping assisted the community in targeting areas for intervention. The issues related to co-sleeping were identified and the community has since created a public awareness campaign looking at this issue. There were noted increases in maternal age associated with infant mortality, and there was a higher prevalence for birth defects for the years studied.
The recommendations for this as you can see it is a FIMR process we used really the recommendation based on the finding included a need for more active birth surveillance a continued birth weight mortality analysis, the utilization of local expertise and community partners to interpret data and a more frequent use of FIMR activities to review deaths. At that point they were doing the FIMR process every three years and more three went local reviews of indicators.
This experience with the local process really shored up our efforts to expand the State and local collaboration to additional areas throughout the State. After we had finished this local review, the 2003 data were available so we looked at the 2003 data as well as the previous data and we identified 15 counties that may very well need some additional targeted technical assistance. We then surveyed the data capacity of each of these 15 counties and it very nicely fell out into 34 teams based on the level of local expertise and evaluation resources. Each team was assigned a leader from the State health office to coordinate communication and training.
The teams now consist of a representative from the local Health Department, the local Healthy Start coalition, the Department of Health contract manager who has a close alliances with the coalition involved, as well the MCH qualify improvement liaison who has a close relationship with the county a member of our evaluation staff and an epidemiologies was used in a consulting capacity at the beginning of this process.
We conducted an orientation call with each of the three teams talking about our process with the Brevard county spike in infant mortality talking about some of the things we learned and techniques we learned to use and then offering technical assistance for such techniques as birth weight specific mortality and using the official matched file. And then the counties and the coalition, this was a purely voluntary process and the counties were able to identify what they had needs for. We are continuing to meet periodically via phone calls, sometimes as a group with all five counties and coalitions together, and more often in individual like strange sessions. Of course, as all things in Florida, the hurricane sort of slowed down our process, but the teams continued to work together and right now, all the teams have successfully opened and manipulated our official matched files for their areas, they are able to analyze their outcomes looking at birth weight specific mortality. Many of them are looking at and learning to do the cause of death analysis, and some still don't know how to use GIS mapping yet. That's another thing that we are going to be providing for them.
Now, one county, Pope County in Florida used the specific techniques that they learned during this infant mortality team. To look at the to 9.7 deaths in 2002. And what they found is that there was a significant number of fetal and infant death that was occurring in counties outside of Pope County so they were not being captured in the feel in a process which is an ongoing process and continues to be an ongoing process in that particular county. So they looked at the basic examination of fetal and infant death trends overtime. This is a chart showing 1997 to 2002. Then through the use of the official matched file, they were able to start looking more closely at risk factors that influenced those rates and specifically this is for 2002 and 2003. They looked at the characteristics of infant deaths and they considered shifts in birth rate categories as well as changes in prematurity. The community considered changes in cause of death for infants in their county. As you can see, the county did have some significant changes in the cause of deaths between the two years that were examined. They considered the influence of race and ethnicity in examining cause of death, and then they developed community findings which included that -- and this is what the coalition then reported to their community that in 2002 there was an increase in the infant death rate and a decrease in the fetal death rate. There an was increase in the percentage of mothers over 35 in 2002. There was an increase in multiple births in 2002. The leading cause of infant death in 2000 and 2003 was prematurity. The premature births were higher for African Americans than whites in both years and both years approximately 50 percent of the deaths occurred outside of level three hospitals.
These are descriptive analyses, but they were tools learned in the infant mortality review teams. The community began to examine infant mortality in new ways. They are developing skills that not only analyze the local community data, but also the capacity for presenting these findings in a way that's meaningful for their community. One of the things we are doing now is performing an analysis at the zip code level so they can target intervention to the highest risk community or the highest risk neighborhoods within the community.
In summary, we all know fetal and infant mortality are influenced by many factors. State level analysis may not be exact enough to discern local issues impacting these outcomes. By working with local communities, providing assistance for the local collaborative efforts, states can support the critical function of the local level analyses and problem solving.
Thank you.