Machine learning methods to predict 30-day hospital readmission outcome among US adults with pneumonia: analysis of the national readmission database

Table 3 Features with variable importance in descending order of non-zero coefficient effect size selected by LASSO logit regression

LASSO model shrinks some coefficient estimates towards zero. Here, we see that only 9 of the 61 coefficient estimates are selected by LASSO Model. Variable importance is ranked by the absolute value of the coefficient for the LASSO model
(1) APRDRG_Severity: Severity of Illness Subclass based on 3M All Patient Refined DRG; (2) RENLFL_SEV: Elixhauser comorbidity measure of moderate and severe renal failure; (3) CANCER_METS: Elixhauser comorbidity measure of metastatic cancer; (4) HTN_CX: Elixhauser comorbidity measure of Hypertension, complicated; (5) I10_NDX: Number of ICD-10-CM diagnoses coded on the record; (6) I10_NPR: Number of procedures coded; (7) mortal_score: Elixhauser Comorbidity Index Score for in-hospital mortality; (8) LOS: Length of Stay; (9) DISPUNIFORMrr: Disposition of patient; (10) APRDRG_Risk_Mortality: Risk of Mortality Subclass based on 3M All Patient Refined DRG

ISSN: 1472-6947