Skip to main content
. 2022 Aug 25;12:14489. doi: 10.1038/s41598-022-18793-6

Figure 1.

Figure 1

Schematic representation of EHR processing. Starting from the EHR subset with sufficient information for classification (‘Included EHR’) EHR are classified into 3 CKD status groups based on clinical diagnosis, risk and other pointers for CKD in the record. Subsequently records with sufficient visits and blood/urine data prior to the evaluation visit are selected. Finally the EHR are split randomly in training and test sets. For the ‘CKD probable’ status group the training set is not shown as it was not used in the study. EHR, electronic health record.