MEGACare

The data and source code for MEGACare: Knowledge-guided Multi-view Hypergraph Predictive Framework for Healthcare. Related code and data will be published in a new repository after review.

Setup

1. Create the rdkit conda environment

conda create -c conda-forge -n MEGACare  rdkit  && conda activate MEGACare

2. Install dependecies

Install the required packages

pip install rdkit-pypi, scikit-learn, dill, dnc

Finally, install other packages if necessary

pip install [xxx] # any required package if necessary

3. Data

Go to https://physionet.org/content/mimiciii/1.4/ to download the MIMIC-III dataset (You may need to get the certificate)

cd ./data
wget -r -N -c -np --user [account] --ask-password https://physionet.org/files/mimiciii/1.4/

Processing the data to get a complete records_final.pkl

Go into the folder and unzip three main files

cd ./physionet.org/files/mimiciii/1.4
gzip -d PROCEDURES_ICD.csv.gz # Procedure information
gzip -d PRESCRIPTIONS.csv.gz  # Medication information
gzip -d DIAGNOSES_ICD.csv.gz  # Diagnosis information

4. Folder Specification

• data/
  • Input:
    • PRESCRIPTIONS.csv
    • DIAGNOSES_ICD.csv
    • PROCEDURES_ICD.csv
    • RXCUI2atc4.csv
    • drug-atc.csv
    • ndc2RXCUI.txt
    • drugbank_drugs_info.csv
    • drug-DDI.csv
  • Output:
    • atc3toSMILES.pkl
    • ADDI.pkl
    • SDDI.pkl
    • records_final.pkl: we only provide the first 100 entries as examples here. We cannot distribute the whole MIMIC-III data.
    • voc_final.pkl
• src/
  • Baselines:
    • LR.py
    • CNN.py
    • RNN.py
    • GRAM.py
    • KAME.py
    • Dipole.py
    • RETAIN.py
    • GAMENet.py
    • SafeDrug.py
    • MICRON.py
    • COGNet.py
    • LEAP.py
    • Retain.py
    • processdata_new.py
    • Pre-trainMPNN.py
    • proposedmethod.py: Our method.
  • Setting file
    • model.py
    • SafeDrug_model.py
    • COGNet_model.py
    • Statistic_ddi_rate_in_mimic.py
    • util.py
    • layer.py

The current statistics are shown below:

#patients  6,350
#clinical events  15,031
#diagnosis  1,958
#med(ATC-3rd)  132
#procedure 1,430
#avg of diagnoses  10.5089
#avg of medicines  11.1864
#avg of procedures  3.8436
#avg of vists  2.3672
#max of diagnoses  128
#max of medicines  64
#max of procedures  50
#max of visit  29

5. Run our project

python proposedmethod.py

usage: proposedmethod.py [-h] [--Test] [--model_name MODEL_NAME]
                   [--resume_path RESUME_PATH] [--lr LR]
                   [--target_addi TARGET_ADDI] [--target_sddi TARGET_SDDI] [--kp KP] [--dim DIM]

4. Tips

Welcome to contact me jialunwu96@163.com for any question.