Exposure measurement error when assessing current glucocorticoid use using UK primary care electronic prescription data.

Misclassification of current use of oral GCs was low but sufficient to lead to significant bias. Researchers should take care to assess the likely impact of exposure misclassification on their analyses.