Cox Regression Demo Notebook.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# A Survival Analysis on The Worcester Heart Attack Study Patients\n",
    "### Jiaqing (Jack) Zhao\n",
    "\n",
    "***"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### A) Load the dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "> I am using the Worcester Heart Attack Study 500 dataset which is a public data set from the Worcester Heart Attack Study. The dataset recorded 500 individuals who were admitted to hospital because of having a heart attach. The covariates contained some of the health status from the individuals. We are trying to use this dataset to fit a Cox regression model on the number of days they can live starting from their admission day to the hospital.\n",
    "\n",
    "> Due to the fact that this dataset is for demo only I will not be discussing into any medical insights or nature of the covariates. Namely, we won't be doing inferences."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "from coxph import * # my implementation on Cox Regression\n",
    "import whas_clean # my script\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sksurv.nonparametric import kaplan_meier_estimator\n",
    "from sksurv.metrics import concordance_index_censored\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>age</th>\n",
       "      <th>gender</th>\n",
       "      <th>hr</th>\n",
       "      <th>sysbp</th>\n",
       "      <th>diasbp</th>\n",
       "      <th>bmi</th>\n",
       "      <th>cvd</th>\n",
       "      <th>afb</th>\n",
       "      <th>sho</th>\n",
       "      <th>...</th>\n",
       "      <th>miord</th>\n",
       "      <th>mitype</th>\n",
       "      <th>year</th>\n",
       "      <th>admitdate</th>\n",
       "      <th>disdate</th>\n",
       "      <th>fdate</th>\n",
       "      <th>los</th>\n",
       "      <th>dstat</th>\n",
       "      <th>lenfol</th>\n",
       "      <th>fstat</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>83</td>\n",
       "      <td>0</td>\n",
       "      <td>89</td>\n",
       "      <td>152</td>\n",
       "      <td>78</td>\n",
       "      <td>25.54051</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1/13/1997</td>\n",
       "      <td>1/18/1997</td>\n",
       "      <td>12/31/2002</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>2178</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>49</td>\n",
       "      <td>0</td>\n",
       "      <td>84</td>\n",
       "      <td>120</td>\n",
       "      <td>60</td>\n",
       "      <td>24.02398</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1/19/1997</td>\n",
       "      <td>1/24/1997</td>\n",
       "      <td>12/31/2002</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>2172</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>70</td>\n",
       "      <td>1</td>\n",
       "      <td>83</td>\n",
       "      <td>147</td>\n",
       "      <td>88</td>\n",
       "      <td>22.14290</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1/1/1997</td>\n",
       "      <td>1/6/1997</td>\n",
       "      <td>12/31/2002</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>2190</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   id  age  gender  hr  sysbp  diasbp       bmi  cvd  afb  sho  ...    miord  \\\n",
       "0   1   83       0  89    152      78  25.54051    1    1    0  ...        1   \n",
       "1   2   49       0  84    120      60  24.02398    1    0    0  ...        0   \n",
       "2   3   70       1  83    147      88  22.14290    0    0    0  ...        0   \n",
       "\n",
       "   mitype  year  admitdate    disdate       fdate los dstat  lenfol  fstat  \n",
       "0       0     1  1/13/1997  1/18/1997  12/31/2002   5     0    2178      0  \n",
       "1       1     1  1/19/1997  1/24/1997  12/31/2002   5     0    2172      0  \n",
       "2       1     1   1/1/1997   1/6/1997  12/31/2002   5     0    2190      0  \n",
       "\n",
       "[3 rows x 22 columns]"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "whas = pd.read_csv(\"whas500.csv\")\n",
    "whas.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### B) What is our y variable? What are the events and how should we calculate the timeline? "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> We have two sets of time to event variables. The first set records the individual's status at the day of discharge from the hospital and the second set contains the indivisual's status at the day of hospital follow up. It is crucial to decide which data to use for each individual. \n",
    ">- __Group 1__. Deceased at hospital <- Small group only 7.8%\n",
    ">- __Group 2__. Survived at discharge but found deceased during followup <- Not perfectly censored, our focus\n",
    ">> that is, if a person dies between the discharge date and followup date... how can we know?\n",
    ">- __Group 3__. Survived follow up <- Perfectly right censored\n",
    "\n",
    "> I can't find much information on the way the dataset is collected and how the follow ups were performed. So how to solve this problem?\n",
    "\n",
    "> - We can try to see if there is a difference in followups for the deceased and alive group and infer the validity of the follow up date as our final date for group 2\n",
    "> - Penalize very long time gaps between discharge and followup for group 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "whas['group'] = whas.apply(whas_clean.groups, axis=1)\n",
    "whas['after_dis'] = whas['lenfol'] - whas['los']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(whas[(whas['dstat'] == 0) & (whas['fstat'] == 1)]['lenfol'] - whas[(whas['dstat'] == 0) & (whas['fstat'] == 1)]['los'], hist=True, kde=True, \n",
    "             bins=int(180/5), color = 'darkblue', \n",
    "             hist_kws={'edgecolor':'black'},\n",
    "             kde_kws={'linewidth': 4}).set_title('Time gap between discharge and followup : Group 2. N = 176')\n",
    "plt.xlabel('Number of days')\n",
    "plt.ylabel('Density');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(whas[(whas['dstat'] == 0) & (whas['fstat'] == 0)]['lenfol'] - whas[(whas['dstat'] == 0) & (whas['fstat'] == 0)]['los'], hist=True, kde=True, \n",
    "             bins=int(180/5), color = 'darkblue', \n",
    "             hist_kws={'edgecolor':'black'},\n",
    "             kde_kws={'linewidth': 4}).set_title('Time gap between discharge and followup : Group 3. N = 285')\n",
    "plt.xlabel('Number of days')\n",
    "plt.ylabel('Density');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> We can clearly see the pattern in the second plot that there are three humps while in the first graph there aren't. It seems like hospital is following up people in three batches with a 500 to 750 days delay in between. Why are hospital giving different follow up times? Several hypothesis:\n",
    "> - Hospital follow up patients quickly for those who are at higher risks\n",
    "> - Has something to do with the cohort. Since there are 3 cohorts to this study (patients adimitted to hospital in 1997, 1999 ans 2001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x432 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "g = sns.FacetGrid(whas[whas[\"group\"] > 1], col=\"year\",row = \"group\")\n",
    "g = g.map(plt.hist, \"after_dis\")\n",
    "\n",
    "for j in range(3):\n",
    "\n",
    "    g.axes[1,j].set_xlabel('Number of days after discharge')\n",
    "    g.axes[1,j].set_ylabel('Frequency')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> __Summary__: it seems like hospitals follow up patients regularly for every 500 to 750 days yet only the latest follow-ups will be recorded. Such that for group 3, people from year 1 have survived 3 follow-ups and people from year 3 have only experienced one follow-up because their cohort is newer. \n",
    "\n",
    "> For people in group 2, if they have a follow-up record thats <= 500 days we know thay did not survive the first follow-up. If they have a follow-up record at ~ 1500 days we know they most likely survived the first 500 days checkpoint. Similarly if they have a follow-up record at ~ 2000 days we know they most likely survived the 1500 days checkpoint. This can also explain why the group 2 partient exhibit distributions that are skewed to the right. The distributions at 500, 1500 and 200 are not well distinguished for group 2 because of\n",
    ">> - Variation in time and\n",
    ">> - Lower count\n",
    "\n",
    "> - Our solution: penalize mildly\n",
    "> - Rubric: Assuming three distinct followup time distributions around ~500, ~1500 and ~2000. We get the mean and quantile of those three distributions from group 3 seperated by year. If any patient's reported follow-up time is larger than the 25th percentile of a distribution and smaller than the 75th of that distribution we resample a followup time (event time) from a normal distribution where the mean is the 25th percentile and a standard deviation of 100 (arbitrary). "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# penalize\n",
    "whas['time'] = whas.apply(whas_clean.penalize, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAo4AAADQCAYAAACJFeBOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAHnNJREFUeJzt3Xm0LHV57vHvw6CMIsiRIIrHASXEBMQjapzAARGjQJyvERSv5N4rBGLMCkZFEpNIvIq5xmgEJaAxRlFQFCfAAcEo42EWMYIJgwzRCDgD7/2jamuz2UN1n+rdvc/5ftbqtaura3irup/ev64xVYUkSZK0mPUmXYAkSZKWBxuOkiRJ6sSGoyRJkjqx4ShJkqRObDhKkiSpExuOkiRJ6sSGoyRJkjqx4biOS/LMJOcnuaT9+7R5hjsyySuWuLyRJNkxyb8l+XmS1026Hq2d1tLsvCzJxe0yfT3JzpOuSWuftTQ7+7TZWZ3kvCRPmnRN47LBpAtYVyXZoKrumHQdwC3Ac6vq+iSPAr4AbDfJgnpYNz8A/gjYt6eSNEXMzvx6WDdXA0+tqh8meTZwDPC4fqrTpJmd+fWwbs4ATqmqSvI7wMeAHfupbrq4xXEMkrwpyZVJzkrykZmtXkm+kuTvkpwHHJpkZZIvtb9SzkiyfTvc8UleMDC929u/uyc5M8mp7fT/MckavYdVdWFVXd8+vQzYOMm9h1jWM5PsMvD8rCQ7J9k0yXFJzklyYZJ92tdXJvlakgvax+8OLNvXkpwCXL6Gy3RTVZ0L/HJNpqOlZ3Ymnp2vV9UP26ffAB64JtPT0jE7E8/O7fXrW/FtCqy1t+Vzi2PPkjwWeD6wM7AhcAFw/sAg96qqVe2wnwZOqKoTkhwIvIvFt5LtBuwEfA/4PPD7wMdn1fBOYI85xv3XqjpqgWk/H7igqn6+SA2DPgC8AjgsySOAjarqoiR/A3ypqg5Mcl/gnCSnAzcBz6yqnyXZAfgIsKqd1q7Ao6rq6tkzSfJR4JFzzP/oqvrgEPVqSpmdqcvOq4DPDbE8mhCzMx3ZSbIf8Fbg/sBzhlieZcWGY/+eCHyqqn4G/KwN6aCPDnQ/gSaAAB8C3tZh+udU1XcBknwEeBKzAlxVfzxs0Ul+C/hbYM8hRz0ReFOSPwUOBI5v++8JPC+/PsZwI2B74Hrg3e2vxTuBRwxM65y5wgtQVS8esi4tP2anMfHsJNmDpuG41h6ntZYxO42JZqeqTgZOTvIU4C3AM4YZf7mw4bj0ftxhmDtoDyNodwnca+C12Zu/77E5fNhffkkeCJwM7F9V/96hvl/PvOonSU4D9gFeBDxmZrLA86vqylnzOhK4keaX8XrAzwZennfduMVRmJ0jWYLspDk+6/3As6vqvzovkKaZ2VnC/ztVdWaShybZuqpu6bBIy4oNx/6dDbwvyVtp1u/v0RxgPpevAy+h+dX3MuBrbf9raILwMeB5NLseZuyW5CE0uwxePNe0h/nl127OPxU4vKrO7jreLO8HPg18beD4qC8AhyQ5pD1Y+NFVdSGwBXBtVd2V5ABg/S4zcIvjOsHsNCaWnfZ4t5OAl1fVt4daEk2S2WlMMjsPB/69ne+uwL2BtfKHlyfH9Kw9KeMU4GKa44MuAX40z+CHAK9McjHwcuDQtv+xwFOTXESzW2HwF9G5wLuBK2jOgDx5DUs+GHg4cESaywisTnL/YSZQVecDtwL/NND7LTRfPBcnuax9DvAe4IB22Xak2y/hoST5jSTXAq8F3pjk2iT36Xs+6pfZ+ZWJZQc4Argf8J52ec4bwzzUM7PzK5PMzvOBS5OsBv4BePHAyTJrlaylyzVRSTarqtuTbAKcCRxUVRf0MN3dgddV1e+t6bRGmPeRwDVVdfwcrz0A+AqwY1XdtbSVaW1idqTRmB0tFbc4jscx7a+OC4BP9BHeaZVkf+CbwBsMr3pgdqTRmB0tCbc4qpP2V+d/V9XqSdciLSdmRxqN2ZlONhwlSZLUybLYVb3XXnsVzen/PnysK49emB0f6+CjF2bHxzr46GRZNBxvuWWtuwyStCTMjjQasyPNbVk0HCVJkjR5NhwlSZLUiQ1HSZIkdWLDUZIkSZ3YcJQkSVInNhwlSZLUyQaTLkDSZK08/NShx7nmqOeMoRJJ0rRzi6MkSZI6seEoSZKkTmw4SpIkqRMbjpIkSerEhqMkSZI6seEoSZKkTmw4SpIkqRMbjpIkSerEhqMkSZI6seEoSZKkTmw4SpIkqZOxNRyTPCjJl5NcnuSyJIe2/Y9Mcl2S1e1j73HVIEmSpP5sMMZp3wH8SVVdkGRz4Pwkp7WvvbOq3j7GeUuSJKlnY2s4VtUNwA1t921JrgC2G9f8JEmSNF5LcoxjkpXAo4Fvtr0OTnJxkuOSbDnPOAclOS/JeTfffPNSlCmtFcyONBqzIy1u7A3HJJsBnwAOq6pbgfcCDwN2odki+Y65xquqY6pqVVWtWrFixbjLlNYaZkcajdmRFjfWhmOSDWkajR+uqpMAqurGqrqzqu4CjgV2G2cNkiRJ6sc4z6oO8AHgiqo6eqD/tgOD7QdcOq4aJEmS1J9xnlX9RODlwCVJVrf9/hx4aZJdgAKuAf5wjDVIkiSpJ+M8q/osIHO89NlxzVOSJEnj451jJEmS1IkNR0mSJHViw1GSJEmd2HCUJElSJzYcJUmS1IkNR0mSJHViw1GSJEmd2HCUJElSJzYcJUmS1Mk4bzm4pFYefuoaT+Oao57TQyWSJElrJ7c4SpIkqRMbjpIkSerEhqMkSZI6seEoSZKkTmw4SpIkqRMbjpIkSepkbA3HJA9K8uUklye5LMmhbf+tkpyW5Kr275bjqkGSJEn9GecWxzuAP6mqnYDHA69JshNwOHBGVe0AnNE+lyRJ0pQbW8Oxqm6oqgva7tuAK4DtgH2AE9rBTgD2HVcNkiRJ6s+SHOOYZCXwaOCbwDZVdUP70veBbeYZ56Ak5yU57+abb16KMqW1gtmRRmN2pMWNveGYZDPgE8BhVXXr4GtVVUDNNV5VHVNVq6pq1YoVK8ZdprTWMDvSaMyOtLixNhyTbEjTaPxwVZ3U9r4xybbt69sCN42zBkmSJPWjU8MxyW8PO+EkAT4AXFFVRw+8dApwQNt9APCpYactratGyaIksyP1pesWx/ckOSfJ/0myRcdxngi8HHhaktXtY2/gKOCZSa4CntE+l9TNKFmUZHakXmzQZaCqenKSHYADgfOTnAP8U1WdtsA4ZwGZ5+WnD12ppJGyKMnsSH3pfIxjVV0FvBH4M+CpwLuSfCvJ74+rOEn3ZBal0Zgdac11Pcbxd5K8k+ZajE8DnltVv9l2v3OM9UkaYBal0ZgdqR+ddlUDfw+8H/jzqvrpTM+quj7JG8dSmaS5mEVpNGZH6kHXhuNzgJ9W1Z0ASdYDNqqqn1TVh8ZWnaTZzKI0GrMj9aBrw/F0mjOgb2+fbwJ8EfjdcRQlaV5mURpN79lZefip9+h3zVHPGXVy0rLQ9eSYjapqJmy03ZuMpyRJCzCL0mjMjtSDrg3HHyfZdeZJkscAP11geEnjYRal0ZgdqQddd1UfBpyY5HqaazP+BvDisVUlaT5mURqN2ZF60PUC4Ocm2RF4ZNvryqr65fjKkjQXsyiNxuxI/ei6xRHgscDKdpxdk1BVHxxLVZIWYhal0ZgdaQ11ajgm+RDwMGA1cGfbuwADJy0hsyiNxuxI/ei6xXEVsFNV1TiLkbQosyiNxuxIPeh6VvWlNAcSS5ossyiNxuxIPei6xXFr4PIk5wA/n+lZVc8bS1WS5mMWpdGYHakHXRuOR46zCEmdHTnpAqRl6shJFyCtDbpejuerSR4M7FBVpyfZBFh/vKVJms0sSqMxO1I/Oh3jmOTVwMeB97W9tgM+ucg4xyW5KcmlA/2OTHJdktXtY+9RC5fWRaNkUZLZkfrS9eSY1wBPBG4FqKqrgPsvMs7xwF5z9H9nVe3SPj7btVBJwGhZlGR2pF50bTj+vKp+MfMkyQY017+aV1WdCfxgDWqTdE9DZ1ESYHakXnRtOH41yZ8DGyd5JnAi8OkR53lwkovbXdlbjjgNaV3VZxaldYnZkXrQteF4OHAzcAnwh8BngTeOML/30ly5fxfgBuAd8w2Y5KAk5yU57+abbx5hVtJaadEsmh1pTmZH6kHXs6rvAo5tHyOrqhtnupMcC3xmgWGPAY4BWLVqlbsTJLpl0exI92R2pH50vVf11cxxLEhVPXSYmSXZtqpuaJ/uR3Mlf0kd9ZVFaV1jdqR+DHOv6hkbAS8EtlpohCQfAXYHtk5yLfBmYPcku9CE9xqa3QWSuhs6i5IAsyP1ouuu6v+a1evvkpwPHLHAOC+do/cHhqhN0iyjZFGS2ZH60nVX9a4DT9ej+eXWdWulpJ6YRWk0ZkfqR9fQDJ79fAfNbuYX9V6NpMWYRWk0ZkfqQddd1XuMuxBJizOL0mjMjtSPrruqX7vQ61V1dD/lSFqIWZRGY3akfgxzVvVjgVPa588FzgGuGkdRkuZlFqXRmB2pB10bjg8Edq2q2wCSHAmcWlV/MK7CJM3JLEqjMTtSD7recnAb4BcDz3/R9pO0tMyiNBqzI/Wg6xbHDwLnJDm5fb4vcMJ4SpK0ALMojcbsSD3oelb1Xyf5HPDkttcrq+rC8ZUlaS5mURqN2ZH60XVXNcAmwK1V9f+Aa5M8ZEw1SVqYWZRGY3akNdSp4ZjkzcCfAa9ve20I/PO4ipI0N7MojcbsSP3ousVxP+B5wI8Bqup6YPNxFSVpXmZRGo3ZkXrQteH4i6oqoACSbDq+kiQtwCxKozE7Ug+6Nhw/luR9wH2TvBo4HTh2fGVJmodZlEZjdqQedD2r+u1JngncCjwSOKKqThtrZZLuwSxKozE7Uj8WbTgmWR84vb1BvCGTJsQsSqMxO1J/Ft1VXVV3Ancl2WIJ6pE0D7MojcbsSP3peueY24FLkpxGe0YaQFX90XwjJDkO+D3gpqp6VNtvK+CjwErgGuBFVfXDkSqX1k1DZ1ESYHakXnRtOJ7UPoZxPPBumts8zTgcOKOqjkpyePv8z4acrrQuGyWLksyO1IsFG45Jtq+q/6iqoe/nWVVnJlk5q/c+wO5t9wnAV7DhKC1qTbIorcvMjtSvxbY4fhLYFSDJJ6rq+Ws4v22q6oa2+/vANvMNmOQg4CCA7bfffg1nKy17nbO4FNlZefipQ49zzVHPGUMl0qKmKjvScrfYyTEZ6H5onzMevBDrPK8fU1WrqmrVihUr+py1tBx1zqLZke7G7Eg9WqzhWPN0j+rGJNsCtH9v6mGa0rqg7yxK6wqzI/VosV3VOye5leYX28ZtN+3zqqr7DDm/U4ADgKPav58acnxpXdV3FqV1hdmRerRgw7Gq1h91wkk+QnMizNZJrgXeTNNg/FiSVwHfA1406vSldcmaZFFal5kdqV9dL8cztKp66TwvPX1c89Taa5STMWbz5AxJktbMoneOkSRJksCGoyRJkjqy4ShJkqRObDhKkiSpk7GdHKO1gyelSJKkGW5xlCRJUic2HCVJktSJDUdJkiR1YsNRkiRJnXhyjKSpNsoJWp6QJUnj4RZHSZIkdWLDUZIkSZ3YcJQkSVInNhwlSZLUiSfHDFib7pLSx7JIkqbDfN/p0/I/R+sOtzhKkiSpk4lscUxyDXAbcCdwR1WtmkQdkiRJ6m6Su6r3qKpbJjh/SZIkDcFd1ZIkSepkUlscC/hikgLeV1XHzB4gyUHAQQDbb7/9Epc3urXpBBvdU18nHY3zPZ7W7HjClqbdtGZHmiaT2uL4pKraFXg28JokT5k9QFUdU1WrqmrVihUrlr5CaZkyO9JozI60uIk0HKvquvbvTcDJwG6TqEOSJEndLXnDMcmmSTaf6Qb2BC5d6jokSZI0nEkc47gNcHKSmfn/S1V9fgJ1SJIkaQhL3nCsqu8COy/1fCVpHEY96ceT4NZOw9zhpY+7wcw1jTUdf9hpaN3i5XgkSZLUiQ1HSZIkdWLDUZIkSZ3YcJQkSVInNhwlSZLUyaRuOShJ67RRzsb2TFdJk+YWR0mSJHViw1GSJEmd2HCUJElSJzYcJUmS1Iknx0yhUW9hNq2mZXmmpQ6Nn7cBvDtPxJm8Yd4Dv6s0zdziKEmSpE5sOEqSJKkTG46SJEnqxIajJEmSOvHkGElqTftJCdNen6aDnxONk1scJUmS1MlEGo5J9kpyZZLvJDl8EjVIkiRpOEvecEyyPvAPwLOBnYCXJtlpqeuQJEnScCaxxXE34DtV9d2q+gXwr8A+E6hDkiRJQ0hVLe0MkxcAe1XV/2yfvxx4XFUdPGu4g4CD2qePBK5cYLJbA7eMody+LYc6rbE/a1LnLVW11ygjroXZWQ41wvKocznUCGanL8uhRlgedS6HGmEJsjO1Dcchp3leVa3qq8ZxWQ51WmN/lkOd1tif5VDncqgRlked1tif5VDncqgRlqbOSeyqvg540MDzB7b9JEmSNMUm0XA8F9ghyUOS3At4CXDKBOqQJEnSEJb8AuBVdUeSg4EvAOsDx1XVZWs42WPWvLIlsRzqtMb+LIc6rbE/y6HO5VAjLI86rbE/y6HO5VAjLEGdS36MoyRJkpYn7xwjSZKkTmw4SpIkqZNl33CcptsXJrkmySVJVic5r+23VZLTklzV/t2y7Z8k72rrvjjJrmOs67gkNyW5dKDf0HUlOaAd/qokByxBjUcmua5dn6uT7D3w2uvbGq9M8qyB/mP7PCR5UJIvJ7k8yWVJDm37T9W6HGJ5zM7idZmdfmo0O+Orxez0W6PZWUxVLdsHzck1/w48FLgXcBGw0wTruQbYela/twGHt92HA3/bdu8NfA4I8Hjgm2Os6ynArsClo9YFbAV8t/27Zdu95ZhrPBJ43RzD7tS+1/cGHtJ+BtYf9+cB2BbYte3eHPh2W8tUrcuOy2J2Rv9cTtX7bXbMjtkxO0u5Lpf7FsflcPvCfYAT2u4TgH0H+n+wGt8A7ptk23EUUFVnAj9Yw7qeBZxWVT+oqh8CpwEj3Z1hiBrnsw/wr1X186q6GvgOzWdhrJ+Hqrqhqi5ou28DrgC2Y8rWZUdmpwOz01uNZmdpmZ3Ra5yP2Wkt94bjdsB/Djy/tu03KQV8Mcn5aW5dBbBNVd3Qdn8f2KbtnnTtw9Y1qXoPbje3HzezKX4aakyyEng08E2Wz7ocNA01DDI7/TM74zENNQwyO/0zOwtY7g3HafOkqtoVeDbwmiRPGXyxmu3FU3f9o2mtC3gv8DBgF+AG4B2TLaeRZDPgE8BhVXXr4GtTvC6nndnpl9lZd5idfpmdRSz3huNU3b6wqq5r/94EnEyzCfvGmV0B7d+b2sEnXfuwdS15vVV1Y1XdWVV3AcfSrM+J1phkQ5rwfriqTmp7T/26nMM01PArZqdfZmespqGGXzE7/TI7i1vuDcepuX1hkk2TbD7TDewJXNrWM3P20gHAp9ruU4D92zOgHg/8aGCz81IYtq4vAHsm2bLddL9n229sZh17sx/N+pyp8SVJ7p3kIcAOwDmM+fOQJMAHgCuq6uiBl6Z+Xc7B7Ixu6t9vszNWZmd0U/9+m50OaonPAOv7QXMG0bdpzmp6wwTreCjN2VQXAZfN1ALcDzgDuAo4Hdiq7R/gH9q6LwFWjbG2j9Bscv8lzXENrxqlLuBAmgOCvwO8cglq/FBbw8VtGLYdGP4NbY1XAs9eis8D8CSa3QEXA6vbx97Tti6HWB6zM9rncqreb7NjdsyO2VnKdektByVJktTJct9VLUmSpCViw1GSJEmd2HCUJElSJzYcJUmS1IkNR0mSJHWyLBuOSd6aZI8k+yZ5/ZDjrkjyzSQXJnnyAsPtnuQza17tcJK8MMkVSb6cZJcke/c8/SOTvK7t/sskzxhy/JVJLl18yPFK8vEkD13C+d3e/n1Ako/3ML3B9+HtSZ62ptPsOF+zM/r0zc5o8zM7ZsfsjDa/qczOsmw4Ao8DvgE8FThzyHGfDlxSVY+uqq/1XtmaexXw6qrag+aWR0MFOMkGXYetqiOq6vQh61sjw9S3wDR+C1i/qr7bQ0lDqarrq+oFPU/274HDe57mfMzOPMzOeJkdswNmZxRTl51xXfxzHA/g/9JcBPM2motg3tY+P2KOYVcCX2pfPwPYniYQ/wHc3I6/8axx9gK+BVwAvAv4TNt/N+DfgAuBrwOPbPufCewyMP5ZwM40XywzF+q8ENh8jvo+CZxPc9HWg9p+RwC301xc9J2zan0xsClwHM3V6i8E9mnHewXNhUq/BHx1jnm9gebipGfRXPD0dW3/44EXtN1HAZe36+vtbb9taG5hNXOB2d9t1+sVNLdiugz44sx6BF5NcxX9i2huj7TJwHz+kebG7EcDK4DT2vHfD3wP2Lod9g/a5VsNvI8mqLOX52+AVww8v71dX5e17/WKtv/DgM+36/lrwI4D9byrfS+/O7AONmvHv4Dmwqn7DM5j4HN1adv9/oH3+WbgzW3/P23Xw8XAXyz2PrSvnQ/8htkxO5gds2N2zM4UZ2csQRvnA3gsTUt5Q+DsBYb7NHBA230g8MmBD/u75xh+I+A/aW4jFOBj/DrA9wE2aLufAXyi7T4A+Lu2+xHAeQPzfuLAh2KDOeY3c5X3jWluaXS/9vlXaK/0PrvW9oP7B233fdsPw6btcNfOTHPWfB7Tfhg3aZfjO8wKMM0V6K+EX10Q/r7t34/S3FAdYH1gi/YDfAftF1e7nmZqut/AfP8KOGRgPp+hDSPwbuD1bfdeNFfF3xr4zXbdbdi+9h5g/zmW6avAbw88L+BlbfcRM+uMJow7tN2PA740UM+JNFvcdwK+0/bfALhP2711u65m1sk9Ajww/wfTfKk9mOY2TsfQfIbWa5f7KQu9D+00jgWeb3bMDmbH7JgdszPF2VnjzbcTsCvNL4sdaVbafJ4A/H7b/SHgbYtMd0fg6qq6CiDJPwMHta9tAZyQZAeaD8uGbf8TgTcl+VOaL4nj2/5nA0cn+TBwUlVdO8f8/ijJfm33g2i+OP5rkRr3BJ43c4wCzZfO9m33aVX1gznGeTJwclX9pF2uue6h+SPgZ8AH2uNrZo6xeRqwP0BV3Qn8qL3H5dVVtbod5nyaDzXAo5L8Fc2Xy2bc/T6YJ7bTgOYWSvu10/18kh+2/Z9O80E/t7k9Jxvz6xu3D9qW5pfWjLtovmwA/hk4KclmNL9UT2ynBXDvgXE+Wc1N7C9Psk3bL8DfJHlKO83taH79fn+OGpoRko1oPgeHVNX3khxC8z5d2A6yGc17uzkLvw83AQ+Ybz49MTtmx+yMxuyYHbPTWjYNxyS70ATkgcAtNC3oJFkNPKGqfjrG2b8F+HJV7ZdkJc2vM6rqJ0lOA/YBXkTz4aOqjkpyKs1xImcneVZVfWtgWXan+QX5hHYaX6EJ42JC8+vgyrv1TB4H/HjUhauqO5LsRhOgFwAH04R3Pj8f6L6TJmjQvD/7VtVFSV4B7D4wXJf6ApxQVYsdeP5TFl5fRfOr67+rapd5hhlchpmEv4xmd8ZjquqXSa5ZZD7Q7Ao5qX59zE6At1bV+wYHSnLYItPZiGa5emd2mlExO2B2hmJ2mlExO2B2fmXZnBxTVavbN+PbNJt5vwQ8q6p2mSe8Xwde0na/jOZYg4V8C1iZ5GHt85cOvLYFcF3b/YpZ472f5riFc6vqhwBJHlZVl1TV39Icc7DjrHG2AH7YhndH4PHz1HQbzS+GGV8ADkn7UybJoxdZJmiOh9k3ycZJNgeeO3uA9lfSFlX1WeCPaY6XgWaT+/9uh1k/yRaLzGtz4IYkG9Ks8/mcTfOFR5I9gS0H5veCJPdvX9sqyYPnGP8K4OEDz9ej+eIB+B/AWVV1K3B1khe200qSnVnYFsBNbXj3oNkFMK8kr6E5juiogd5fAA5s1ylJtmuXZ7H34RE0u456Z3YAszPD7AzB7ABmZ4bZaS2bhiM0lzSg+eDfRXPA6eULDH4I8MokFwMvBw5daNpV9TOaXQSnJrmAu2+qfhvw1iQXMmsrbVWdD9wK/NNA78OSXNrO+5fA52bN7vPABkmuoDk4+BvzlPVlYKckq5O8mOYX6IbAxUkua58vqKouoNmcflFbx7lzDLY58Jm23rOA17b9DwX2SHIJza6BnRaZ3ZtoDkQ+m+YLcT5/AeyZ5vIKL6TZJH9b+36+EfhiW8tpNLsHZjuVe/6q3K2d3tOAv2z7vwx4VZKLaA5g3meR+j8MrGqXd/9FlgHgdcBvt+/P6iT/q6q+CPwL8G/tdD5OE/J534f2C+/hwHmLzG9kZsfstMzOkMyO2WmZnZlxqzlAUiNK8gCaXQg7tl8sWkSSewN3trsqngC8d4FN+3ONvzHNl9sTq+rOJLdX1Wbjqnfc0hxztGtVvWnStSwlszM8s3N3ZsfsdGV27m5NsrNsjnGcRkn2B/4aeK3hHcr2wMeSrAf8guZyCp1V1U+TvJnmIOL/GEN9S20D4B2TLmIpmZ2RmZ27Mzvqyuzc3cjZcYujJEmSOllWxzhKkiRpcmw4SpIkqRMbjpIkSerEhqMkSZI6seEoSZKkTv4/on6yYIMq1ZQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 648x216 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# check the distributions again\n",
    "g = sns.FacetGrid(whas[whas[\"group\"] == 2], col=\"year\",row ='group')\n",
    "g = g.map(plt.hist, \"time\")\n",
    "\n",
    "for j in range(3):\n",
    "\n",
    "    g.axes[0,j].set_xlabel('# of days after discharge (penalized)')\n",
    "    g.axes[0,j].set_ylabel('Frequency')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### C) Exploratory Analysis with the KM Curve"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Now we will do some exploratory analysis with the KM curve. We will breifly check the following aspects to see if anything interesting comes up\n",
    "> - overall trend\n",
    "> - by demographics (age, gender)\n",
    "> - by cohort\n",
    "> - by simple health features (initial heart rate, bmi)\n",
    "\n",
    "> We are using scikit-learn survival package to generate these graphs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Overall all KM Graph. \n",
    "> We can see a huge drop between t = 0 and t = 500. The trend gets slower after that."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "time, survival_prob = kaplan_meier_estimator(whas[\"fstat\"].values.astype('bool'), whas[\"time\"].values)\n",
    "plt.step(time, survival_prob, where=\"post\")\n",
    "plt.ylabel(\"est. probability of survival $\\hat{S}(t)$\")\n",
    "plt.xlabel(\"time $t$\");"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### KM curve by demographics\n",
    "> Higher age => higher risks. And females seem to have higher risks than males. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "gen_dic = {0:\"Male\",1:\"Female\"}\n",
    "whas_clean.display_km(gen_dic, \"gender\", whas)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "whas['age_group'] = whas.apply(whas_clean.age_group, axis=1)\n",
    "age_dic = {1:\"30~50\",2:\"51~70\",3:\"71~90\",4:\"91+\"}\n",
    "whas_clean.display_km(age_dic, \"age_group\", whas)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### KM curve by cohort\n",
    "> Pretty similar (drops due to shorter time span)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "year_dic = {1:\"1997\",2:\"1999\",3:\"2001\"}\n",
    "whas_clean.display_km(year_dic, \"year\", whas)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### KM curve by health status\n",
    "> Seems like higher heart rates and low body weights lead to more risks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "whas['hr_level'] = whas.apply(whas_clean.hr_level, axis=1)\n",
    "hr_dic = {1:\"too low\",2:\"low\",3:\"normal\",4:\"high\"}\n",
    "whas_clean.display_km(hr_dic, \"hr_level\", whas)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "whas['bmi_level'] = whas.apply(whas_clean.bmi_level, axis=1)\n",
    "bmi_dic = {1:\"under-weight\",2:\"normal\",3:\"overweight\",4:\"obese\"}\n",
    "whas_clean.display_km(bmi_dic, \"bmi_level\", whas)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### D) Predictive Modeling\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> I tried to implement the Cox regression myself. I modified the code from Scikit-survival because it is such a well-written package. I ended-up:\n",
    "> - adding comments, especially on the computation parts \n",
    "> - changing coding structure (merged several classes) and variable names (like changing *risk_set* to *sum_risk_j* to fit the original mathametical representations better)\n",
    "> - changing minor issues (non-essential, found some redundancy in sksurv code)\n",
    "> - stripping excessive functionalities and rewrote some functions for the sake of simplicity (e.g. we are not doing regularizations)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def cross_validation_cox(X,y,cv,y_name = 'after_dis'):\n",
    "    c_scores = []\n",
    "    coef_ = []\n",
    "    for i in range(cv):\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X,y, test_size = 0.2)\n",
    "        cpr = cox_ph_regressor().fit(X_train,y_train)\n",
    "        prediction = cpr.predict(X_test)\n",
    "        result = concordance_index_censored(y_test[\"fstat\"].values.astype(\"bool\"), y_test[y_name], prediction)\n",
    "        c_scores.append(float(result[0]))\n",
    "        coef_.append(cpr.coef_)\n",
    "    print(\"Average cross validation score: \" + str(sum(c_scores)/len(c_scores)))\n",
    "    print(\"Standard deviation of validation score:\" + str(np.std(c_scores)))\n",
    "    #return coef_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# cross validation\n",
    "cv = 5\n",
    "# prepare data \n",
    "X = whas.iloc[:,2:14]\n",
    "y = whas.iloc[:,[-7,-5]]\n",
    "y_original = whas.iloc[:,[-7,-8]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average cross validation score: 0.7757834684374055\n",
      "Standard deviation of validation score:0.026766735268062895\n"
     ]
    }
   ],
   "source": [
    "# cross validation score for our penalized y\n",
    "co_ = cross_validation_cox(X,y,cv)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average cross validation score: 0.7119700765848948\n",
      "Standard deviation of validation score:0.03480471683147719\n"
     ]
    }
   ],
   "source": [
    "# cross validation score for our original y\n",
    "cross_validation_cox(X,y_original,cv,\"lenfol\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> We can see that our penalized y yields better results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### E) Summary & Insights\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On the study...\n",
    "> - We used exploratory analysis to figure out the hospitals/researchers' routine to follow up patients and adjusted our time and events accordingly.\n",
    "> - From KM curve analysis we can see that visually there are huge differences between groups like gender, age, heart rate and etc. That gave us confidence in building a predictive model\n",
    "> - Our predictive models have a pretty good cross validation result. C-score of .776 is pretty high given that we train on only 400 data points."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On the methods...\n",
    "> - Another big part of this project is to understand how Cox Regression can be implemented in code\n",
    "> - We used Newton's method to minimize the negative log partial likelihood function for Cox regression\n",
    "\n",
    "<img src=\"img/ll.png\" alt=\"Drawing\" style=\"width: 300px;\"/>\n",
    "\n",
    "> - The predictions are relative risks, assessed by concordence index\n",
    "> - Possible to perform regularizations"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Other insights...\n",
    "> - With more data we can easily adapt other machine learning methods e.g. Random Survival Forest, SVM and etc...\n",
    "> - While working with our own data we can perform better feature engineering"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}