Ninth Annual Maternal and Child Health Epidemiology Conference / December 10-12, 2003

Recent Findings From PRAMS Surveillance Data 1996-2001

DENISE D'ANGELO:  Thank you, Amy.  Good afternoon.  I’m pleased to be presenting some recent findings from PRAMS.  The key Objectives for PRAMS are to analyze the data that we collect, to translate it into information that can be used for planning, evaluating, and changing public health, programs and policies and to disseminate that information to appropriate audiences.  Today I’ll be highlighting 2001 multi-state PRAMS data for several maternal and infant health indicators.  This is preliminary data and therefore, all of the PRAMS states are not represented.  Before I get started, I’d like to acknowledge my co-authors shown here, and the PRAMS working group shown here.  First, I will briefly describe the PRAMS methodology.  I’ll talk about the surveillance report series as a source of the data that I’ll be presenting and then I’ll present results from several PRAM states for 2001.  First, I’d like to give a brief overview of the PRAMS project.  PRAMS stands for Pregnancy Risk Assessment Monitoring System. 

It’s an ongoing population based state specific system.  The program is administered by CDC and states or cities participate through cooperative agreements.  The PRAMS dataset provides self-reported information on a variety of maternal behaviors and experiences that occur before, during, or shortly after pregnancy.  Women who deliver a live born infant are sampled at two to six months after delivery.  All PRAMS projects use a standardized data collection protocol.  The protocol addresses every component of the project from sampling to data collection to data analysis.  To increase response rates, PRAMS uses a mixed mode data collection methodology.  Up to three mailed surveys are sent to mothers in the sample and those who don’t respond to these mailings are followed up by telephone. 

Now, I’d like to talk a little bit about CDC’s role in utilizing PRAMS data.  We focus on analyzing and disseminating multi-state data.  Since 1995, CDC has produced a surveillance report for each year that presents data on a wide variety of topics for each state.  Over 20 indicators are presented in each report.  Prevalence data for the current year are presented by state and by selected socio-demographic characteristics.  When available, trend data have been reported by state as well.  Following the format of the surveillance reports, CDC has also published several MMWR surveillance summaries.  In November of this year, our most recent summary was published.  This report highlighted four indicators that were new in 2000.  All of these reports are available to download or order via the PRAMS website shown here, and we also brought copies of the most recent surveillance summary here, and for those of you who don’t want to carry the whole booklet home, we also have just a little one-page hand out that has the website address on there.  Now, I’d like to give a preview of the findings from the 2001 PRAMS surveillance data. 

I will present data from four indicators that we are tracking over time with regards to their Healthy People 2010 Objectives, and then show data for four indicators which were new beginning in 2000.  This map shows the 33 projects that participated in PRAMS in 2000, 32 states and New York City.  We will examine data from the 11 states that are highlighted in pink.  The states presented are those that had 2001 weighted data available with the response rate of at least 70% when we began putting this presentation together in November.  For each indicator, I’ll first present the prevalence estimates for 2001 for each of the 11 states, and I’ll present trend data from 1996 to 2001 only for the seven states that have at least three years of data, including 2001. 

Let’s start by looking at data on unintended pregnancy.  PRAMS collects data on unintended pregnancy among women who delivered a live infant.  It’s important to remember that the denominators here is women delivering a live infant and not all pregnant women.  For this analysis, we defined unintended pregnancy as both mis-timed and unwanted pregnancy.  In 2001, the prevalence of unintended pregnancy ranged from a low of 36.7% in Maine to a high of 48.9% in Alabama.  The red line on this and other prevalence graphs depicts the Healthy People 2010 goal for this indicator.  It doesn’t appear that any of the states are likely to meet the Healthy People 2010 Objective which calls for no more than 30% of all pregnancies to be unintended.  Now, let’s examine trends in unintended pregnancy.  Before we talk about the trend data, I’d like to take a minute to describe the presentation of the tables.  Across the X-axis we have the years ’96 through 2001, and the Y-axis depicts the prevalence.  There’s a line for each of the seven states that have trend data.  Most of the states have data for the full time period, however, newer states have data from ’98 through 2001.  I’ve identified those states where no significant changes were observed with yellow lines. 

While there are no significant changes on this graph, red lines will be used to note significant changes over time.  Looking at this table, we see that no state has seen a significant decrease in the prevalence of unintended pregnancy over time.  PRAMS collects data on the prevalence of physical abuse both during the 12 months before pregnancy and during pregnancy.  I’m going to show the prevalence of physical abuse during pregnancy.  Physical abuse is defined as pushing, hitting, slapping, kicking, choking, or any other way of physically hurting someone, and here we are reporting abuse during pregnancy.  In 2001, the prevalence of physical abuse by a husband or partner during pregnancy ranged from 2.6% in Michigan to 5.8% in Alabama.  The Healthy People 2010 goal is to reduce physical assault by an intimate partner to no more than 3.3 per 1,000 persons age 12 and over.  The red line is a little high there but that’s about as close as I could get it.  No states demonstrated a statistically significant change in the prevalence of physical abuse by a husband or partner during pregnancy during this time period. 

Now, we’ll look at the prevalence of entry into prenatal care after the first trimester of pregnancy.  There’s a wide range in the prevalence of entry into prenatal care after the first trimester, with the lowest prevalence being 16.5% in Maine and the highest 28.2% in New Mexico.  The Healthy People 2010 Objective is to increase to at least 90% the proportion of pregnant women who receive prenatal care in the first trimester, or inversely to reduce the proportion that enter after the first trimester to no more than 10%.  PRAMS states shown here still need to make some progress in order to meet that goal.  There is some good news to report in the prevalence of entering prenatal care after the first trimester, however.  While most states do not show a significant change over time, we see that the state of North Carolina has made some progress in this area with a steady and significant decrease in the prevalence of late entry into prenatal care.  Now, I’m going to talk about an infant health indicator, infant back sleep position.  Back sleep position has been known to reduce the risk of Sudden Infant Death Syndrome.  This is measured by a new mother’s answer to the question, “How do you most often lay your baby down to sleep?” 

In 2001, the prevalence of placing infants to sleep on their backs ranged from 47.9% in Alabama, to 76.9% in Maine.  The Healthy People 2010 Objective calls for at least 70% of healthy term infants to be placed to sleep on their backs.  In 2001, four states met or exceeded this goal:  Colorado, Maine, Michigan, and Nebraska.  Three states were also very close at about 68%.  Those are Hawaii, Illinois, and New Mexico.  We see more good news when we examine trends in infant sleep back sleep position.  Statistically significant increases in the prevalence in of placing infants to sleep on their back were observed in all seven states, even those falling short of meeting the Healthy People 2010 Objective.  The PRAMS Phase IV questionnaire which was implemented in states with January 2000 births allowed us to collect some new maternal and child health data.  Today, I’ll be presenting prevalence data for these indicators for 2000 and 2001.  The new indicators are listed below and we’ll talk about them in more details (inaudible).  We collect data on the prevalence of complications during pregnancy and the severity of these complications. 

Here we are reporting the prevalence of at least one of these nine possible complications where the respondent indicated that she stayed in the hospital for at least one day because of the problem.  I’d like to note that the survey question also includes car accidents as a possible complication, but that is not included for this report.  On the following graphs, you can see two bars for most states.  The blue bar represents 2000 data and the pink bar represents 2001 data, and I’ll be speaking about the 2001 results.  In 2001, pregnancy related complications requiring a hospital stay of at least one day range from 10.5% in Maine to a high of 15.3% in Alabama.  Now, I’d like to look at a newly added infant alpha indicator available in 2000 and 2001.  This is the prevalence of infants receiving a checkup in the first week after leaving the hospital.  The decrease in the length of hospital delivering the past decade, there’s been a ongoing debate about the safety of early hospital discharge of mothers and newborns.  For infants, appropriate follow up is recommended by the American Academy of Pediatrics and the American College of Obstetrics and Gynecology within one week of leaving the hospital for infants discharged within 48 hours of delivery.  In 2001, the prevalence of first week infant checkup ranged from a low of 64.9% in Alabama to a high of 88% in Colorado. 

Now, we’ll look at postpartum contraceptive use.  Contraceptive behavior is of interest because of its relationship to unintended pregnancy, abortion, and sexually transmitted diseases.  In addition, postpartum contraceptive use is critical in preventing pregnancy shortly after childbirth.  It can pose health risks to mothers and infants.  Postpartum contraceptive use is defined as the respondent reporting “yes” to doing something to keep from getting pregnant at the time of the survey which is two to six months postpartum.  In 2001, the prevalence of postpartum contraceptive use ranged from 78.6% in Hawaii to 90.1% in Alabama.  The last new indicator that I’ll present today is on folic acid intake.  In order to measure pre-pregnancy folic acid intake, we added a question on multi-vitamin use in the month before pregnancy to the Phase IV survey.  Here we are reporting the prevalence of women saying that she took a multi-vitamin four or more times per week during the month before pregnancy.  In 2001, the prevalence of multi-vitamin use ranged from a low of 26.4% in Alabama and a high of 42.3% in Maine.  The Healthy People 2010 goal is to increase the proportion of pregnancies began with an optimum folic acid level to 80%.  The optimum level was defined as at least 400 micrograms everyday in the month prior to becoming pregnant.  Although multi-vitamin use does not account for folic acid intake through regular foods, none of the states shown here are close to meeting this goal. 

In conclusion, I want to point out the value of PRAMS in providing population based state level data on a variety of issues that may not be available from other data sources.  The data can be used to measure progress toward Healthy People 2010 Objectives.  A number of states have state negotiated Title V performance measures that can be monitored with PRAMS data as well.  The data are useful for program planning and guiding policy decisions.  I want to stress the importance of collecting surveillance data on behaviors and experiences that affect maternal and infant health.  With these data we can better understand what is happening in our states and provide direction for improving the health of mothers and infants.  Again, all PRAM states with 2001 data are not included in this report but we plan to have the final full report for 2001 available sometime next year.