A Bayesian Approach to Estimating Red Blood Cell Folate Concentration for Optimal Reduction of Neural Tube Defect Risk When Precise Folate Concentration Measures Are Not Available

• Activity Number: 338
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Owen Devine*+ and Krista Crider
• Companies: CDC and CDC
• Keywords: Imputation ; Missing Data ; Imprecise Data

Abstract

The goal of this analysis is to estimate the association between red blood cell (RBC) folate concentration and the risk of neural tube birth defects (NTDs) leading to development of target population-level folate concentrations for reducing NTD risk. The analysis was based on data from a large intervention study focused on relating folic acid intake and reduction in NTD risk. However, the participants' RBC folate concentrations were not measured. We used a Bayesian approach to impute missing concentrations based on informative priors developed in an independent study focused on the relationship between folic acid intake and RBC concentration. Because folic acid uptake is affected by genetic variations in the MTHFR gene, and genetic information was not available, we also developed data-based priors for this missing information for incorporation into the imputation process. The Bayesian imputation approach facilitated realistic estimation of the uncertainty associated with estimated target RBC folate concentrations reflecting both lack of knowledge on the true association between folate concentration and NTD risk and the use of imprecise information in the estimation process.