Adapting the pregnancy risk assessment monitoring survey to enhance locally available data: methods

Despite the increasing emphasis on pre- and interconception planning, perinatal data available to local municipalities and organizations is often limited to that on the birth certificate. A partnership between a local health department and an academic medical center sought to overcome this gap. Using the core questions from the Pregnancy Risk Assessment Monitoring System (PRAMS) and a stratified random sample methodology (by income) in a county with ~8,000 annual births we mailed 2,462 surveys to mothers who gave birth between May 2009 and April 2010. Mailings occurred at 4-5 months postpartum. Low income mothers (those with a Medicaid-funded delivery and/or prenatal WIC enrollment) were oversampled based on a projected response rate of 35% (rate for non-low income was 55%). Over 1,000 usable surveys were returned and linked with birth certificate data. Target response rates were achieved. 9.4% of addresses for low income mothers were undeliverable (vs. 4.2% of non-low income). Both low and non-low income respondents were more likely to be over age 18 and White. After statistical adjustments the survey dataset was demographically similar to the original birth data. Personnel and non-personnel costs per usable survey exceeded $20. Collecting local data using a modified PRAMS methodology is feasible but requires expertise in survey, data management and birth certificate data and local knowledge about survey response patterns. These types of data can serve to inform policy and program planning and provide data to support relevant funding requests.