Skip to main content

Table 2 Features selected in the conventional logistic regression

From: Predicting 30-days mortality for MIMIC-III patients with sepsis-3: a machine learning approach using XGboost

  OR_with_CI p value
(Intercept) 52,913.003 (87.92–33,934,517.782) < 0.001
Sofa 1.142 (1.106–1.179) < 0.001
Aniongap_min 1.078 (1.043–1.115) < 0.001
Creatinine_min 0.676 (0.592–0.767) < 0.001
Chloride_min 0.98 (0.962–0.999) 0.03393
Hematocrit_min 1.113 (1.053–1.178) < 0.001
Hemoglobin_min 0.748 (0.623–0.895) 0.00169
Hemoglobin_max 0.926 (0.863–0.992) 0.02993
Lactate_min 1.308 (1.194–1.435) < 0.001
Potassium_min 1.179 (1.001–1.389) 0.04922
Sodium_max 1.046 (1.019–1.074) < 0.001
Bun_min 1.033 (1.018–1.048) < 0.001
Bun_max 0.986 (0.973–0.997) 0.01542
Wbc_min 1.062 (1.036–1.09) < 0.001
Wbc_max 0.969 (0.952–0.987) < 0.001
Heartrate_min 0.987 (0.977–0.997) 0.0111
Heartrate_mean 1.022 (1.011–1.033) < 0.001
Sysbp_min 0.991 (0.984–0.998) 0.00839
Meanbp_min 0.992 (0.985–1) 0.0468
Resprate_mean 1.062 (1.038–1.086) < 0.001
Tempc_min 0.897 (0.81–0.993) 0.03242
Tempc_max 0.781 (0.698–0.873) < 0.001
Spo2_mean 0.947 (0.909–0.986) 0.00839
Age 1.029 (1.022–1.035) < 0.001
Diabetes 0.779 (0.639–0.948) 0.01328
Vent 1.824 (1.48–2.251) < 0.001
  1. OR odds ratio, CI confidence interval, SOFA sequential organ failure assessment, bun blood urea nitrogen, wbc white blood cell, sysbp systolic blood pressure, meanbp mean blood pressure, resprate respiratary rate, Spo2 oxyhemoglobin saturation, vent ventilation, Max maximum, Min minimum