Background

In 1987, the Centers for Disease Control and Prevention (CDC) implemented the Pregnancy Risk Assessment Monitoring System (PRAMS) to help state health departments establish and maintain an epidemiologic surveillance system of selected maternal behaviors and experiences to supplement data from vital records. PRAMS was developed in response to distressing statistics on infant mortality and low birthweight. The U.S. infant mortality rate was no longer declining as rapidly as it had in past years, and the prevalence of low-birthweight infants showed little change. In addition, maternal behaviors such as smoking and drug use were recognized as contributors to these slow rates of decline.

Purpose

Because PRAMS data are population-based, findings from data analyses can be generalized to an entire state's population of women having a live birth. In addition, the use of standardized data collection methods allows for comparisons among states. In each participating state, PRAMS data supplement information from vital records and can be used to plan, monitor, and evaluate policies and programs designed to reduce adverse pregnancy outcomes and to improve the health of babies and mothers. Findings from analyses of PRAMS data can be used to enhance states' understanding of maternal behaviors and experiences and their relationship with adverse pregnancy outcomes.

History

PRAMS is administered by the Division of Reproductive Health (DRH), National Center for Chronic Disease Prevention and Health Promotion (NCCDPHP) at CDC. PRAMS operates through cooperative agreements between CDC and the states, which have been awarded funds competitively. At the state level, PRAMS management and operating structures may cross multiple organizational units, including maternal and child health (MCH) and vital statistics.

Since the program's inception, the number of participating health departments has grown from 6 in 1987 (5 states and the District of Columbia) to 30 in 2004 (29 states and New York City). In 2004, participants were Alabama, Alaska, Arkansas, Colorado, Florida, Georgia, Hawaii, Illinois, Louisiana, Maine, Maryland, Michigan, Minnesota, Mississippi, Nebraska, New Jersey, New Mexico, New York, New York City, North Carolina, Ohio, Oklahoma, Oregon, Rhode Island, South Carolina, Texas, Utah, Vermont, Washington, and West Virginia. In 2004, live births in these states and New York City represented approximately 62% of all live births in the United States.

In 2002, the year of this report, PRAMS participants included 31 states and New York City. Thirty of the 32 participants conducted traditional PRAMS surveillance, while two states—Montana and North Dakota—conducted point-in-time surveys. PRAMS participants for 2002 were Alabama, Alaska, Arkansas, Colorado, Florida, Georgia, Hawaii, Illinois, Louisiana, Maine, Maryland, Michigan, Minnesota, Mississippi, Montana, Nebraska, New Jersey, New Mexico, New York, New York City, North Carolina, North Dakota, Ohio, Oklahoma, Oregon, Rhode Island, South Carolina, Texas, Utah, Vermont, Washington, and West Virginia (see map below). In 2002, PRAMS surveillance covered 62% of all live births in the United States.

PRAMS Participants in 2002

The United States map shows all 32 states that collected PRAMS data in 2002. Data are included for a total of 27 states with fully implemented PRAMS data collection procedures and weighted response rates of at least 70 percent: Alabama, Alaska, Arkansas, Colorado, Florida, Hawaii, Illinois, Louisiana, Maine, Maryland, Michigan, Minnesota, Montana, Nebraska, New Jersey, New Mexico, New York, North Carolina, North Dakota, Ohio, Oklahoma, Rhode Island, South Carolina, Utah, Vermont, Washington, and West Virginia. Five additional states collected data but are not included in this report because their weighted response rates are less than 70 percent: Georgia, Mississippi, New York City, Oregon, and Texas.

Methodology

Each participating state and New York City uses a standardized data collection method developed by CDC.¹ PRAMS staff in each state collect data through statewide mailings and follow up with nonrespondents by telephone. Every month, a stratified sample of 100 to 300 new mothers per state is selected from eligible birth certificates. At 2 to 6 months after delivery, PRAMS staff in each state mail each mother in the sample a package containing a letter introducing the survey and the 14-page survey itself. Mothers who do not respond to the first mailed survey are mailed a second questionnaire package, and in most states, a third package is mailed to mothers who do not respond to the second request. PRAMS interviewers telephone mothers who do not respond to any of the mailed surveys and administer the questionnaire by telephone.

The PRAMS questionnaire addresses many topics, including unintended (mistimed and unwanted) pregnancy, barriers to and content of prenatal care, infant sleeping position, obstetric history, pregnancy-related complications, maternal use of alcohol and tobacco, multivitamin use, economic status, maternal stress, postbirth follow-up and well-baby care, and pre- and postpartum contraceptive use. Not all topics are included in this report. The questionnaire consists of a core component and a state-specific component. The core portion is used by all participating PRAMS states. Each state develops a state-specific portion that addresses particular state data needs.

Since the program's inception, the PRAMS questionnaire has undergone several revisions, referred to as phases. Revisions to the questionnaire have been made to capture data on recent guidelines or emerging MCH issues (such as knowledge of folic acid's relationship to birth defects) and to improve respondents' comprehension of questions. In January 2000, states implemented the fourth phase of the questionnaire. The 32 indicators presented in this report are from the core component of the Phase 2 (1993–1995), Phase 3 (1996–1999), and Phase 4 questionnaires (2000–2002).

Validity of specific questions is addressed through pretesting. New questions are tested through cognitive interviewing, in which respondents are asked to describe their understanding of a question's meaning and how they arrived at their response. Based on the results of the cognitive testing, questions are revised. A second round of testing involves administering the questionnaire to respondents who are asked to complete it and provide feedback. Questions are then finalized for use on the survey. Prior to the next revision cycle, questions are evaluated for item nonresponse, write-in responses, and whether respondents correctly followed the skip patterns in the survey. Using these criteria, questions that perform poorly are revised accordingly and pretested before being included in the questionnaire.

In 2002, two states (Montana and North Dakota) conducted point-in-time PRAMS surveys for births during only a portion of the calendar year. North Dakota sampled births that occurred January through April 2002 and Montana sampled births that occurred February through May 2002. Except for the truncated time period, sampling and follow-up were done according to the standard PRAMS protocol. The data were weighted to represent the total number of live births that occurred in 2002 in each of these states.

Additional information on PRAMS can be found in the appendixes. Appendix A describes the PRAMS data collection methodology and questionnaire revisions. Appendix B lists the 2002 stratification variables, total sample sizes, and weighted response rates for each state. Appendix C identifies the corresponding PRAMS question number from the PRAMS Phase 4 core questionnaire for each indicator in this report, defines each indicator, and specifies which indicators have associated Healthy People 2010 objectives or Title V Maternal and Child Health Block Grant performance measures. Appendix D is the Phase 4 core questionnaire. Appendix E lists which states participated in PRAMS, by year. Finally, Appendix F defines the abbreviations used in this report.

Technical Notes

This report includes data from Alabama, Alaska, Arkansas, Colorado, Florida, Hawaii, Illinois, Louisiana, Maine, Maryland, Michigan, Minnesota, Montana, Nebraska, New Jersey, New Mexico, New York, North Carolina, North Dakota, Ohio, Oklahoma, Rhode Island, South Carolina, Utah, Vermont, Washington, and West Virginia. These 27 states had fully implemented PRAMS data collection procedures in 2002 and achieved weighted response rates of at least 70% (range: from 70% to 88%). The weighted response rate indicates the proportion of women sampled who completed a survey, adjusted for sample design.

For most of the indicators in this report, the wording of the questions changed little between the Phase 3 and Phase 4 versions of the PRAMS survey. However, small changes in the wording of some questions affected the following indicators: contraceptive use, alcohol use, tobacco use, and breastfeeding. Therefore, tables for these indicators include 2002 prevalence data and trend data for the period covered by the Phase 4 questionnaire (2000–2002).

The multistate prevalence table for each of the 32 indicators presents state estimates and confidence intervals (CIs) using 2002 data. A bar graph of 2002 prevalence for each state accompanies the multistate prevalence table. The 2002 data for Minnesota (May through December) and New Jersey (July through December) represent births for only a portion of 2002.

In addition, a multistate trend table complements the trend table for each indicator. Depending on the indicator, the trend table presents data for 1993–2002 (8 indicators), 1996–2002 (5 indicators), or 2000–2002 (19 indicators). In addition to the state-specific exceptions to the 2002 data noted above, there are states with exceptions for 1997, 1998, and 2001 data. The 1997 data for North Carolina represent only a partial year (July through December) and the 1998 data for New Mexico represent births from July 1997 through December 1998. For 2001, data for Maryland (February through December) and Michigan (July through December) represent births for only a portion of the calendar year, while 2001 data for Vermont include births from October 2000 through December 2001.

In the state-specific tables presenting the social, demographic, and outcome characteristics of the PRAMS-eligible population and PRAMS respondents, information on maternal age, race, ethnicity, education, marital status, and parity and birthweight were obtained from state birth certificate data provided to CDC. For all states, births to out-of-state residents or births occurring out of state (except in Alaska) are excluded from the description of the PRAMS-eligible population. The PRAMS-eligible population is described using five race categories (white, black or African American, American Indian, Asian or Pacific Islander, and all other races) for all states, except Alaska. Alaska reports separately data for Alaska Natives.

For the state-specific sections of the report, data for racial groups comprising at least 5% of the state's birth population are reported separately. For 19 of the 27 states featured in the report, the three race categories are white, black or African American, and all other races. Eight of the 27 states (Alaska, Hawaii, Montana, New Jersey, New Mexico, North Dakota, Oklahoma, and Washington) have additional race categories for 2002. Further, data on Medicaid status of PRAMS respondents were obtained from the PRAMS questionnaire; a Medicaid recipient was defined as a woman who reported that she received Medicaid just before she became pregnant or that Medicaid paid for her prenatal care or the delivery.

The detailed summary tables present data for all states using three race categories: white, black or African American, and all other races. The "all other races" category includes American Indians, Asian or Pacific Islanders, Alaska Natives, Native Hawaiians, and other races.

Except for the tables describing the PRAMS-eligible population in each state, all tables in the report are produced using weighted PRAMS data. Percentages and standard errors are calculated for the characteristic of interest using PROC CROSSTAB in SUDAAN.² The 95% CIs were computed using the formula CI = percentage ± (1.96 x standard error). The number of respondents is the number of mothers who answered that PRAMS question. All missing (blank and "don't know") observations are excluded. The percentage of missing values is noted when it equals or exceeds 10%. Because estimates based on small samples are imprecise and may be biased, estimates for which the number of respondents is fewer than 30 are not reported. In the detailed summary tables, estimates based on sample sizes between 30 and 60 are reported, but they include a note stating that the estimates may be unreliable. In the trend tables (1993–2002, 1996–2002, or 2000–2002), trend statistics are presented for states with three or more data points for an indicator. The P value indicates a test for linear trend and was calculated using PROC LOGISTIC in SUDAAN.²

PRAMS data are representative of women whose pregnancies resulted in a live birth and are not generalizable to all pregnant women. For one reporting area, data are not representative of the entire state: New York data are for upstate New York only and exclude New York City, which has an autonomous vital records agency.

References

1. Centers for Disease Control and Prevention (CDC). PRAMS Model Surveillance Protocol, 1999 (Unpublished). 1999.
   
   
2. Research Triangle Institute (RTI). SUDAAN Language Manual, Release 9.0. Research Triangle Park, NC: RTI; 2004.