06_mortality_prediction.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "06-mortality-prediction",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/MIT-LCP/2019_hack_aotearoa_eicu/blob/master/06_mortality_prediction.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "metadata": {
        "id": "T3wdKZCPklNq",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# eICU Collaborative Research Database\n",
        "\n",
        "# Notebook 6: Mortality prediction\n",
        "\n",
        "This notebook explores how a logistic regression can be trained to predict in-hospital mortality of patients.\n"
      ]
    },
    {
      "metadata": {
        "id": "rG3HrM7GkwCH",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Load libraries and connect to the database"
      ]
    },
    {
      "metadata": {
        "id": "s-MoFA6NkkbZ",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Import libraries\n",
        "import numpy as np\n",
        "import os\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib.patches as patches\n",
        "import matplotlib.path as path\n",
        "\n",
        "# model building\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "from sklearn.pipeline import Pipeline\n",
        "from sklearn import preprocessing\n",
        "from sklearn import metrics\n",
        "from sklearn import impute\n",
        "\n",
        "# Make pandas dataframes prettier\n",
        "from IPython.display import display, HTML\n",
        "\n",
        "# Access data using Google BigQuery.\n",
        "from google.colab import auth\n",
        "from google.cloud import bigquery"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "jyBV_Q9DkyD3",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# authenticate\n",
        "auth.authenticate_user()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "cF1udJKhkzYq",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Set up environment variables\n",
        "project_id='new-zealand-2018-datathon'\n",
        "os.environ[\"GOOGLE_CLOUD_PROJECT\"]=project_id"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "GDEewAlvk0oT",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Helper function to read data from BigQuery into a DataFrame.\n",
        "def run_query(query):\n",
        "    return pd.io.gbq.read_gbq(query, project_id=project_id, verbose=False, \n",
        "                            configuration={'query':{'useLegacySql': False}})"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "LgcRCqxCk3HC",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Load the patient cohort\n",
        "\n",
        "In this example, we will load all data from the patient data, and link it to APACHE data to provide richer summary information."
      ]
    },
    {
      "metadata": {
        "id": "ReCl7-aek1-k",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Link the patient and apachepatientresult tables on patientunitstayid\n",
        "# using an inner join.\n",
        "query = \"\"\"\n",
        "SELECT p.unitadmitsource, p.gender, p.age, p.admissionweight, \n",
        "    p.unittype, p.unitstaytype, a.acutephysiologyscore,\n",
        "    a.apachescore, a.actualhospitalmortality\n",
        "FROM `physionet-data.eicu_crd_demo.patient` p\n",
        "INNER JOIN `physionet-data.eicu_crd_demo.apachepatientresult` a\n",
        "ON p.patientunitstayid = a.patientunitstayid\n",
        "WHERE apacheversion LIKE 'IVa'\n",
        "\"\"\"\n",
        "\n",
        "cohort = run_query(query)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "yxLctVBpk9sO",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "cohort.head()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "NPlwRV2buYb1",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Prepare the data for analysis"
      ]
    },
    {
      "metadata": {
        "id": "v3OJ4LDvueKu",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# review the data dataset\n",
        "print(cohort.info())"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "s4wQ6o_RvLph",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Encode the categorical data\n",
        "encoder = preprocessing.LabelEncoder()\n",
        "cohort['gender_code'] = encoder.fit_transform(cohort['gender'])\n",
        "cohort['admissionweight_code'] = encoder.fit_transform(cohort['admissionweight'])\n",
        "cohort['unittype_code'] = encoder.fit_transform(cohort['unittype'])\n",
        "cohort['apachescore_code'] = encoder.fit_transform(cohort['apachescore'])\n",
        "cohort['actualhospitalmortality_code'] = encoder.fit_transform(cohort['actualhospitalmortality'])\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "4ogi_ns-ylnP",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Handle the deidentified ages\n",
        "cohort['agenum'] = pd.to_numeric(cohort['age'], downcast='integer', errors='coerce')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "77M0QJQ5wcPQ",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Preview the encoded data\n",
        "cohort[['gender','gender_code']].head()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "GqvwTNPN3KZz",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Check the outcome variable\n",
        "cohort['actualhospitalmortality_code'].unique()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "ze7y5J4Ioz8u",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Create our train and test sets"
      ]
    },
    {
      "metadata": {
        "id": "i5zXkn_AlDJW",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "predictors = ['gender_code','agenum','apachescore_code','unittype_code',\n",
        "              'admissionweight_code']\n",
        "outcome = 'actualhospitalmortality_code'\n",
        "\n",
        "X = cohort[predictors]\n",
        "y = cohort[outcome]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "IHhIgDUwocmA",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=10)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "NvQWkuY6nkZ8",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Review the number of cases in each set\n",
        "print(\"Train data: {}\".format(len(X_train)))\n",
        "print(\"Test data: {}\".format(len(X_test)))\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "b2waK5qBqanC",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Build the model"
      ]
    },
    {
      "metadata": {
        "id": "EVoWS8HX3Sek",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Create an instance of the model\n",
        "model = LogisticRegression(solver='lbfgs')\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "sCWwZnc30ahA",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Impute missing values and scale using a pipeline\n",
        "estimator = Pipeline([(\"imputer\", impute.SimpleImputer(missing_values=np.nan, strategy=\"mean\")),\n",
        "                      (\"scaler\", preprocessing.StandardScaler()),\n",
        "                      (\"logistic_regression\", model)])\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "k_-lrq1A0rCh",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Fit the model to the training data\n",
        "estimator.fit(X_train, y_train)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "NMDmhhnlzr9Z",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Testing"
      ]
    },
    {
      "metadata": {
        "id": "Pf4G20-VGdvm",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "y_pred = estimator.predict(X_test)\n",
        "print('Accuracy of logistic regression classifier on the test set: {:.2f}'.format(estimator.score(X_test, y_test)))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "akOjuNz8qgMd",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "print(metrics.classification_report(y_test, y_pred))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "gH1w0RBH1JjT",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "logit_roc_auc = metrics.roc_auc_score(y_test, estimator.predict(X_test))\n",
        "fpr, tpr, thresholds = metrics.roc_curve(y_test, estimator.predict_proba(X_test)[:,1])\n",
        "\n",
        "plt.figure()\n",
        "plt.plot(fpr, tpr, label='Logistic Regression (area = %0.2f)' % logit_roc_auc)\n",
        "plt.plot([0, 1], [0, 1],'r--')\n",
        "plt.xlim([0.0, 1.0])\n",
        "plt.ylim([0.0, 1.05])\n",
        "\n",
        "plt.xlabel('False Positives')\n",
        "plt.ylabel('True Positives')\n",
        "plt.title('Receiver operating characteristic')\n",
        "plt.legend(loc=\"lower right\")\n",
        "plt.savefig('Log_ROC')\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "VvsNIjCDDIo_",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}