Parameter Adjustment of Case-Control Analysis to Adjust for Different Pathogenicity of Multiple Exposures

• Activity Number: 341 - Statistical Applications in Infectious Disease Epidemiology
• Type: Contributed
• Date/Time: Tuesday, August 1, 2017 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Diana Lam* and William Blackwelder
• Companies: University of Maryland, Baltimore and University of Maryland, Baltimore
• Keywords: case-control study ; EM algorithm ; conditional logistic regression ; co-infections

Abstract

In case-control studies, cases may have multiple exposures with varying magnitudes of influence on case status. In the Global Enteric Multicenter Study (GEMS), our cases were defined as those with moderate-to-severe diarrhea (MSD) but a proportion of our cases had multiple pathogen exposures such as rotavirus and giardia. While rotavirus is a main cause of MSD, giardia is not; however, it was extremely prevalent in our study population. Diarrhea in a case with both rotavirus and giardia exposure was most likely caused by the rotavirus. We propose conditional logistic regression parameter adjustment using the EM algorithm and apply it to the GEMS data. To evaluate our adjustment, we perform a simulation to compare how our method to regular conditional logistic regression and the true values.