The dataset includes dozens of columns that indicate the technological level of the facility at which the patient was seen. We discussed this in the EHR technology and meaningful use section in Chapter 2, Healthcare Foundations. We omit these columns since they are valued on a per-hospital basis rather than a per-encounter basis:

emr_cols = [
    'EBILLANYE','EMRED','HHSMUE','EHRINSE','EDEMOGE',
    'EDEMOGER','EPROLSTE','EPROLSTER','EVITALE','EVITALER',
    'ESMOKEE','ESMOKEER','EPNOTESE','EPNOTESER','EMEDALGE',
    'EMEDALGER','ECPOEE','ECPOEER','ESCRIPE','ESCRIPER',
    'EWARNE','EWARNER','EREMINDE','EREMINDER','ECTOEE',
    'ECTOEER','EORDERE','EORDERER','ERESULTE','ERESULTER',
    'EGRAPHE','EGRAPHER','EIMGRESE','EIMGRESER','EPTEDUE',
    'EPTEDUER','ECQME','ECQMER','EGENLISTE','EGENLISTER',
    'EIMMREGE','EIMMREGER','ESUME','ESUMER','EMSGE',
    'EMSGER','EHLTHINFOE','EHLTHINFOER','EPTRECE','EPTRECER',
    'EMEDIDE','EMEDIDER','ESHAREE','ESHAREEHRE','ESHAREWEBE',
    'ESHAREOTHE','ESHAREUNKE','ESHAREREFE','LABRESE1','LABRESE2',
    'LABRESE3','LABRESE4','LABRESUNKE','LABRESREFE','IMAGREPE1',
    'IMAGREPE2','IMAGREPE3','IMAGREPE4','IMAGREPUNKE','IMAGREPREFE',
    'PTPROBE1','PTPROBE2','PTPROBE3','PTPROBE4','PTPROBUNKE',
    'PTPROBREFE','MEDLISTE1','MEDLISTE2','MEDLISTE3','MEDLISTE4',
    'MEDLISTUNKE','MEDLISTREFE','ALGLISTE1','ALGLISTE2','ALGLISTE3',
    'ALGLISTE4','ALGLISTUNKE','ALGLISTREFE','EDPRIM','EDINFO',
    'MUINC','MUYEAR'
]

X_train.drop(emr_cols, axis=1, inplace=True)
X_test.drop(emr_cols, axis=1, inplace=True)