The PRAMS sample of women who have had a recent live birth is drawn from the jurisdiction’s birth certificate file. Each participating site samples between 1,000 and 3,000 women per year. Women from some groups are sampled at a higher rate to ensure adequate data are available in smaller but higher risk populations. Selected women are first contacted by mail. If there is no response to repeated mailings, they are contacted and interviewed by telephone. Data collection procedures and instruments are standardized to allow comparisons between sites.
The Pregnancy Risk Assessment Monitoring System (PRAMS): Overview of Design and Methodology [PDF – 612 KB]
tailored design method. New York: John Wiley and Sons, 2000.
One strength of the PRAMS surveillance system is the standardized data collection methodology. This standardized approach allows for comparisons among sites and for optimal use of the data for single-jurisdiction or multi-jurisdiction analysis. The standardized data collection methodology is prescribed in the CDC Model Surveillance Protocol.1 Each site follows the protocol, but also has the opportunity to customize some portions of it to tailor the procedures to the needs of the jurisdiction.
PRAMS combines two modes of data collection; a survey conducted by mailed questionnaire with multiple follow-up attempts, and a survey by telephone. The principles and practices of mail/telephone survey methodology used by CDC are based primarily on the research of Don Dillman.2 A key aspect of his approach is to make numerous and varied contacts with sampled mothers. Here is the sequence of contacts for PRAMS surveillance:
- Preletter. This letter introduces PRAMS to the mother and informs her that a questionnaire will soon arrive.
- Initial Mail Questionnaire Packet. This packet is sent to all sampled mothers 3 to 7 days after the preletter. Its contents are described below.
- Tickler. The tickler serves as a thank you and a reminder note. It is sent 7 to 10 days after the initial mail packet.
- Second Mail Questionnaire Packet. This packet is sent to all sampled mothers who have not yet responded 7 to 14 days after the tickler has been sent.
- Third Mail Questionnaire Packet. This third packet is sent to all remaining nonrespondents 7 to 14 days after the second questionnaire.
- Telephone Follow-up. Telephone follow-up is initiated for all mail nonrespondents 7 to 14 days after mailing the last questionnaire.
The series of mailings begins commences two to four months after delivery. The survey contains questions about the early postpartum period; thus, the mailings are timed to ensure that all women will be able to respond for this period. The data collection cycle from the mailing of the pre-letter to the close of telephone follow-up lasts approximately 60 to 95 days. Each month, a stratified sample is drawn from the current birth certificate file. For each of these monthly samples, or “batches,” this sequence of mailing contacts is attempted. All data collection activities are managed and tracked using the PRAMS Integrated Data Collection System (PIDS), a web-based customized tracking system. PIDS was developed by CDC and housed on CDC secured servers. PIDS is designed to schedule and track data collection activities, record mail and telephone responses, including write-in comment data, generate reports to summarize operational data, manage call attempts for telephone interviews, and prepare customized letters.
The mail packets include the following items:
- Multipurpose cover letter: This letter describes PRAMS and its purpose, explains how the mother was chosen and why, solicits the mother’s cooperation, describes the procedures for completing and returning the questionnaire, explains any incentive or reward, and provides a telephone number for additional information. This letter is modified slightly for the second and third mailings, primarily by adding an additional appeal for response. Beginning in 2004, the multipurpose cover letter was divided into two documents; an introductory letter and an informed consent information sheet.
- Survey booklet: The survey booklet carries similar specifications for each site. The booklet is 14 pages or less in length, has a colorful cover designed by each site, and is slightly smaller than an 8½” by 11″ sheet of paper. It contains an extra page at the end for respondents to provide extra comments. A self-addressed return envelope with postage is provided for the easy return of the questionnaire.
- Q&A brochure: A question and answer brochure is included to provide additional information about PRAMS. It contains answers to the most frequently asked questions about the survey. The brochure can be an important tool to encourage participation.
- Calendar: A 3-year calendar is provided to serve as a memory aid for answering the questions.
- Incentive and reward: All participating sites have adopted the use of some type of participation incentive (sent to all sampled mothers) or reward (sent to all mothers who responded to the survey or respondents). Examples of participation incentives and rewards include: coupons for certified birth certificates, participation in a raffle for a cash award, postage stamps, bibs, cash (a dollar bill), magnetic picture frames, diapers, and gift cards.
Telephone follow-up begins after the mailing of the third questionnaire packet. Sites use a variety of sources to obtain valid or current telephone numbers of participants. Calls to a particular number are staggered over different times of the day and different days of the week. The calling period for a batch runs two to four weeks. Up to 15 call attempts are made for each viable telephone number. Often, telephone interviewers arrange call-back interviews to accommodate the mother’s schedule.
The original PRAMS questionnaire was developed in 1987. Topics addressed in the PRAMS questionnaire include barriers to and content of prenatal care, obstetric history, maternal use of alcohol and cigarettes, physical abuse, contraception, economic status, maternal stress, and early infant development and health status. Learn more about the questionnaires and download the latest PRAMS questionnaires.
Each jurisdiction has their unique sampling approach, so the data are weighted by jurisdiction. PRAMS aggregate data are not weighted to provide national estimates. The weighting process described below is completed for each jurisdiction.
Each participating site draws a stratified random sample of 100 to 250 new mothers every month from a frame of eligible birth certificates. Sampling plans are tailored to each site’s research and programmatic needs. Sites identify stratification variable(s) of interest for oversampling of sub-population(s). Commonly used stratification variables include birth weight, maternal race and ethnicity, Medicaid status, and geography. Annual sample sizes range from 1,000 to 3,000 per site, and are divided among two to six strata. Typically, the annual sample is large enough for estimating site-wide risk factor proportions within 3.5% at 95% confidence. Estimated proportions within strata are slightly less precise (typically, they are estimated within 5% at 95% confidence).
Survey responses are linked to extracted birth certificate data items to create the complete Analytic Research File (ARF) for analysis. Thus, the PRAMS data set contains a wealth of demographic and medical information collected through the jurisdiction’s vital records system. The availability of birth certificate information is the basis for drawing stratified samples and, ultimately, for generalizing results to the sites’ entire population of births. The availability of births data for all sampled individuals, whether they responded or not, is key to deriving nonresponse weights.
For each respondent, the initial sampling weight is the reciprocal of the sampling fraction applied to the stratum. Sampling fractions in PRAMS typically range from 1 in 1 (for low birth weight strata in small jurisdictions) to about 1 in 211 (for normal birth weight, nonminority strata in populous jurisdictions). The corresponding sampling weights then would range from 1 to 211.
Nonresponse adjustment factors attempt to compensate for the tendency of women having certain characteristics (such as being unmarried or of lower education) to respond at lower rates than women without those characteristics. If multivariate analysis shows that these characteristics affect the propensity to respond in a particular stratum, the adjustment factor is the ratio of the sample size in that category to the number of respondents in the category. If analysis shows that no characteristic distinguishes respondents from nonrespondents, the adjustment factor is the ratio of the sample size in that stratum to the number of respondents in the stratum. In the first case, each category so identified has an adjustment factor; in the second, there is a single factor for the whole stratum.
The rationale for applying nonresponse weights is the assumption that nonrespondents would have provided similar answers, on average, to respondents’ answers for that stratum and adjustment category. So that cells with few respondents are not distorted by a few women’s answers, small categories are collapsed until each cell contains at least 25 respondents. The magnitude of the adjustment for nonresponse depends on the response rate for a category. If 80% (or 4/5) of the women in a category responded, the nonresponse weight is 1.2 (or 5/4). Categories with lower response rates have higher nonresponse weights.
Frame omission studies are carried out to look for problems that occur during frame construction. The frame noncoverage weights are derived by comparing frame files for a year of births to the calendar year birth tape that sites provided to CDC. Omitted records are usually due to late processing and are evenly scattered across the site, but sometimes they are clustered by particular hospitals or counties or even times of the year. The effect of the noncoverage weights is to bring totals estimated from sample data in line with known totals from the birth tape. In mail/telephone surveillance, the magnitude of noncoverage is small (typically from 1% to 5%), so the adjustment factor for noncoverage is not much greater than 1.
Multiplying together the sampling, nonresponse, and noncoverage components of the weight yields the analysis weight. The weight can be interpreted as the number of women like herself in the population that each respondent represents.
Analyzing PRAMS data requires software that takes into account the complex sampling designs that sites employ. Such software utilizes first-order Taylor series approximations to calculate appropriate standard errors for the estimates it produces.
- Centers for Disease Control and Prevention. PRAMS model surveillance protocol, 2005 CATI version. Unpublished.
- Dillman DA. Mail and Internet surveys: the tailored design method. New York: John Wiley and Sons, 2000.