Use of negative control outcomes to detect mismeasured confounders in administrative data.

Challenge

Mismeasured confounders in administrative claims data are a recognized source of bias, but the utility of NCOs as a detection tool for confounders that are present but systematically miscoded had not been formally evaluated.

Solution

Target RWE researchers presented an oral analysis demonstrating how NCOs can be used to detect the presence of mismeasured confounders in administrative data, extending the NCO framework beyond its traditional unmeasured confounding application.