Each month approximately 200 mothers who are Missouri residents and recently delivered a live-born infant during the preceding 2-4 months are randomly selected from a file of birth certificate records using stratified systematic sampling. Mothers, who are mentally incapacitated, have died, or had a multiple birth greater than three gestations are excluded from the sampling frame. Annual sample sizes range from 2,000 to 3,000. The Missouri PRAMS sample is stratified by infant birth weight and the mother’s area of residence. Infants with low birth weight (<2,500 grams) and infants whose mothers lived in rural areas were oversampled to permit separate estimates of these subgroups of interest based on adequate sample size.
PRAMS uses two modes of data collection. The initial data collection method is a mail survey. Up to three self-administered surveys are mailed to sampled women. Women not responding to any of the three mailings are contacted by telephone. The sequence of contacts for PRAMS surveillance is:
1. Pre-letter. This is a brief letter that is sent a few days to one week in advance of the questionnaire. It informs the woman that a survey is forthcoming while providing only minimum detail about the project itself.
2. First Questionnaire Mailing. The questionnaire is sent with a cover letter that describes the project. Other appropriate materials (e.g., informed consent document, incentive, question-and-answer brochure, resource list, calendar, return envelope) are included as well.
3. Tickler (Reminder Letter). This letter is sent to non-respondents 7 to 10 days after the questionnaire to remind the woman to complete the questionnaire and to offer thanks in case the questionnaire was already returned.
4. Second Questionnaire Mailing. Another questionnaire is mailed one to two weeks after the tickler. A cover letter informs the woman that the previous questionnaire has not been returned and includes a stronger appeal for participation. Research does not support the inclusion of another incentive in this mailing packet. However, the informed consent document, question-and-answer brochure, resource list, calendar, and return envelope are included with this mailing.
5. Third Questionnaire Mailing. One to two weeks after the second questionnaire is mailed, a third and final questionnaire is mailed to the woman. A cover letter informs the person that the previous questionnaire has not been returned and includes a stronger appeal for participation. Research does not support the inclusion of another incentive in this mailing packet. However, the informed consent document, question-and-answer brochure, resource list, calendar, and return envelope are included with this mailing.
6. Telephone Follow-up. Seven to ten days after the final mailing, telephone calls are initiated with any mail non-respondents. Up to 15 telephone calls are made in an attempt to reach the mother and persuade her to complete a telephone interview. Participants entering the telephone survey are not eligible for the incentive.
The timing and nature of the mail and telephone contacts is designed to elicit the best response rates possible and minimize recall bias. To maintain data integrity, the entire data collection period will not exceed 95 days from the date of the pre-letter. The PRAMTrac software will assist in setting and managing the contact schedule. The Missouri PRAMS survey follows the standardized data collection methodology provided in the CDC PRAMS Model Surveillance Protocol distributed in August 2009. The complete protocol can be accessed at the following web address: http://www.cdc.gov/prams/.
This standardized approach allows for comparisons among states. Multiplying together the sampling, non-response, and non-coverage 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 needs to take into account analysis weight and complex sampling designs. Data presented in Missouri PRAMS report and other types of publications are weighted to provide estimates representative of Missouri resident women who have recently delivered a live birth.