Fig. 1From: The effect of high prevalence of missing data on estimation of the coefficients of a logistic regression model when using multiple imputationMean estimated regression coefficient (when cholesterol is subject to missingness)Back to article page