Abstract:
|
Data from the National Health and Nutrition Examination Survey (NHANES) have been linked to the Center for Medicare and Medicaid Services' Medicaid Enrollment and Claims Files for the survey years 1999-2004. This project examines the gains in estimate precision and quality when multiple imputation is used in conjunction with data-linkage. As not all survey participants provide sufficient information to be eligible for record linkage, multiple imputation methods estimate the Medicaid status of those who are not eligible to be linked to the administrative data. Medicaid status among children who are linkage eligible and linked to Medicaid enrollment during the month of interview determine the methods of missingness and imputation is executed using IVEware. Comparisons are drawn across the three different estimates: those made using Medicaid status from the survey interview, those made from linked data among those who are linkage eligible, and those made from linked data after imputation accounts for children who are not linkage eligible. In order to demonstrate how the method applies to the analysis of health outcomes, cotinine levels in children ages 3-17 are explored.
|