Fig. 3From: Addressing health disparities using multiply imputed injury surveillance dataBrier Skill Score (BSS)a comparison for evaluating imputation performance on simulation datab by using the different models (JM and FCS). a BSS indicates the degree of skill improvement. A BSS range from 0 to 1: 0 means no improvement in accuracy and 1 means a perfect accuracy of prediction. bSimulation data was developed by imposing the missing data patterns on the subset of fully observed data in National Electronic Injury Surveillance System-All Injury Program (NEISS-AIP) 2018. JM: joint modelling. FCS: fully conditional specification. LOC: location where the injury occurred. RACE: race and ethnicity of patient. CAUSE: external cause of injury. BDYPT: primary body part affected. TYPE: work-relatedness. AGE: patient age in year. DISP: disposition at emergency department discharge. SEX: gender of patientBack to article page