Design of an Observational Study to Evaluate the Effectiveness of IAP During Labor Against Early Onset Disease in Newborns

• Activity Number: 84
• Type: Contributed
• Date/Time: Sunday, August 1, 2010 : 4:00 PM to 5:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Donald B. Rubin and Stephanie J. Schrag and Elizabeth R. Zell*+
• Companies: Harvard University and CDC and CDC
• Keywords: Propensity Score Methods ; Observational Data ; Causal Inference

Abstract

A sample survey was conducted to evaluate the implementation of new early-onset group B streptococcus (GBS) prevention guidelines. The resulting dataset allowed the assessment of whether intrapartum antibiotic prophylaxis (IAP) during labor is effective against early-onset GBS disease. A propensity score logistic regression was conducted, which included covariates associated with the survey design, published guideline indications for receipt of IAP, and prognostically relevant factors. Ten subclasses were created with approximately equal total survey weights in each. Diagnostics evaluated balance within each subclass and across all subclasses. The appropriate weighting factor for each subclass was the number of women recommended to receive IAP based on the new guidelines. Issues involved in designing a study from such observational weighted survey data will be described.