A Method for Missing Data Problems in Epidemiologic Studies with Historical Exposure

• Activity Number: 253
• Type: Contributed
• Date/Time: Monday, July 30, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Xinbo Zhang*+ and Bryan Langholz and Myles Cockburn
• Companies: Oregon Health and Science University and University of Southern California School of Medicine and University of Southern California
• Keywords: Missing Data ; Case-Control Study ; Exposure History ; Cox Regression ; Logistic Regression ; Asymptotic Efficiency

Abstract

In this study we consider methods for a specific missing pattern where missing values occur across the exposure history of an individual. We propose a missing indicator induced intensity (MIND) method under the rare disease assumption. The essential part is the parametrization of the induced intensity, which actually reflects the missing mechanism, common missing mechanism assumption such as MCAR and MAR are no longer required. The MIND method is compared against simple imputation methods in a Monte Carlo simulation study and demonstrates superior in term of bias and efficiency. The method is shown to reach an asymptotic efficiency equal to the expected non-missing proportion under cohort design. Under nested case-control sampling design, the asymptotic efficiency varies but stays close to cohort design. Under rare disease assumption, the method can be bridged back to case-control design based on logistic model. The method is then applied to the University of Southern California prostate cancer-pesticide pilot study to assess its performance. The MIND method is overall efficient and flexible in solving the missing data problem where the missingness occurs in the exposure history.