This milestone will be complete when all 8 outcomes are computed from mom-baby ehr.

Top 10 most important features (Updated)
Below we list top 10 most important features whose absence can affect the prediction power of the cox-nnet model:

[ ] 1. Early-onset preeclampsia/ Late-onset preeclampsia: preeclampsia that was diagnosed before 34 weeks of gestation
[ ] 2. Severe PE/ Mild PE. Severe PE: O14.10, O14.12, O14.13 in ICD10, or 642.50, 642.51, 642.52, 642.53, 642.54 If a patient has severe preeclampsia at any time during pregnancy, she’s considered having severe PE; otherwise, she’s considered having mild preeclampsia.
[ ] 3. Number of fetuses: The number of fetuses at any point during a pregnancy. This does not necessarily reflect the number of babies that were born.
[ ] 4. Previous gravida counts: The number of times a woman has been pregnant regardless of the outcome. This number includes the current pregnancy.
[ ] 5. Previous parity counts: The number of pregnancies that reach a viable gestational age, regardless of pregnancy outcome. Multiple fetuses in the same pregnancy are counted as one parity count.
[ ] 6. Uncomplicated hypertension: Comorbidity from the most recent visit before pregnancy.
7. Complicated Diabetes: Comorbidity from the most recent visit before pregnancy.
8. Uncomplicated Diabetes: Comorbidity from the most recent visit before pregnancy.
9. Pulmonary circulation disorders: Comorbidity from the most recent visit before pregnancy.
10. Patient Age (year): Age at delivery

*Gestational age (week): The baby's gestational age in weeks at delivery. If gestational age is not directly available in the system, it can be computed using date of delivery minus date of last menstrual period.

*If a patient has multiple pregnancy records in the system, only the latest pregnancy information should be used in the model.

If a variable other than those listed above is missing, one can simply fill the column with 0 and the model will still work.