{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "2f4a0417-ff9b-41df-adc7-0589cb24d156",
   "metadata": {},
   "source": [
    "# How much of the inter-hospital variation in thrombolysis use do in-hospital processes explain\n",
    "\n",
    "Aims:\n",
    "\n",
    "* Investigate the correlation (explained variance) between hospital model process parameters and the variation in use of thrombolysis between hospitals."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "511496fe-6e63-4f0b-9c59-6a265cdb948b",
   "metadata": {},
   "source": [
    "## Load data and pivot by scenario"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "10d83928-c22f-4bd4-ac0a-5211e94af239",
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "# Load data\n",
    "scenarios = pd.read_csv('output/key_scenario_results.csv')\n",
    "\n",
    "# Performance data\n",
    "performance = pd.read_csv(\n",
    "    'hosp_performance_output/hospital_performance.csv', index_col='stroke_team')\n",
    "\n",
    "performance['hosp_speed'] = (\n",
    "    np.exp(performance['arrival_scan_arrival_mins_mu']) +\n",
    "    np.exp(performance['scan_needle_mins_mu']))\n",
    "\n",
    "# Decision data\n",
    "decision =  pd.read_csv(\n",
    "    '../random_forest/predictions/corhort_rates.csv', index_col='hospital')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "cababd31-3092-4818-9df1-b93e8a6265cb",
   "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>stroke_team</th>\n",
       "      <th>scenario</th>\n",
       "      <th>admissions</th>\n",
       "      <th>thrombolysis_rate</th>\n",
       "      <th>additional_good_outcomes_per_1000_patients</th>\n",
       "      <th>patients_receiving_thrombolysis</th>\n",
       "      <th>add_good_outcomes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>AGNOF1041H</td>\n",
       "      <td>base</td>\n",
       "      <td>671.666667</td>\n",
       "      <td>15.11</td>\n",
       "      <td>12.72</td>\n",
       "      <td>101.488833</td>\n",
       "      <td>8.543600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>AKCGO9726K</td>\n",
       "      <td>base</td>\n",
       "      <td>1143.333333</td>\n",
       "      <td>15.06</td>\n",
       "      <td>13.43</td>\n",
       "      <td>172.186000</td>\n",
       "      <td>15.354967</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>AOBTM3098N</td>\n",
       "      <td>base</td>\n",
       "      <td>500.666667</td>\n",
       "      <td>7.81</td>\n",
       "      <td>5.74</td>\n",
       "      <td>39.102067</td>\n",
       "      <td>2.873827</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>APXEE8191H</td>\n",
       "      <td>base</td>\n",
       "      <td>439.333333</td>\n",
       "      <td>10.08</td>\n",
       "      <td>7.35</td>\n",
       "      <td>44.284800</td>\n",
       "      <td>3.229100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ATDID5461S</td>\n",
       "      <td>base</td>\n",
       "      <td>275.666667</td>\n",
       "      <td>9.20</td>\n",
       "      <td>6.42</td>\n",
       "      <td>25.361333</td>\n",
       "      <td>1.769780</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  stroke_team scenario   admissions  thrombolysis_rate  \\\n",
       "0  AGNOF1041H     base   671.666667              15.11   \n",
       "1  AKCGO9726K     base  1143.333333              15.06   \n",
       "2  AOBTM3098N     base   500.666667               7.81   \n",
       "3  APXEE8191H     base   439.333333              10.08   \n",
       "4  ATDID5461S     base   275.666667               9.20   \n",
       "\n",
       "   additional_good_outcomes_per_1000_patients  \\\n",
       "0                                       12.72   \n",
       "1                                       13.43   \n",
       "2                                        5.74   \n",
       "3                                        7.35   \n",
       "4                                        6.42   \n",
       "\n",
       "   patients_receiving_thrombolysis  add_good_outcomes  \n",
       "0                       101.488833           8.543600  \n",
       "1                       172.186000          15.354967  \n",
       "2                        39.102067           2.873827  \n",
       "3                        44.284800           3.229100  \n",
       "4                        25.361333           1.769780  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scenarios.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "b5833678-ce27-4ee6-9ede-e7fe97fb03f5",
   "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>thrombolysis_rate</th>\n",
       "      <th>admissions</th>\n",
       "      <th>80_plus</th>\n",
       "      <th>onset_known</th>\n",
       "      <th>known_arrival_within_4hrs</th>\n",
       "      <th>onset_arrival_mins_mu</th>\n",
       "      <th>onset_arrival_mins_sigma</th>\n",
       "      <th>scan_within_4_hrs</th>\n",
       "      <th>arrival_scan_arrival_mins_mu</th>\n",
       "      <th>arrival_scan_arrival_mins_sigma</th>\n",
       "      <th>onset_scan_4_hrs</th>\n",
       "      <th>eligable</th>\n",
       "      <th>scan_needle_mins_mu</th>\n",
       "      <th>scan_needle_mins_sigma</th>\n",
       "      <th>hosp_speed</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>stroke_team</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>AGNOF1041H</th>\n",
       "      <td>0.154839</td>\n",
       "      <td>671.666667</td>\n",
       "      <td>0.425459</td>\n",
       "      <td>0.635236</td>\n",
       "      <td>0.681250</td>\n",
       "      <td>4.576874</td>\n",
       "      <td>0.557598</td>\n",
       "      <td>0.965596</td>\n",
       "      <td>1.665700</td>\n",
       "      <td>1.497966</td>\n",
       "      <td>0.935867</td>\n",
       "      <td>0.388325</td>\n",
       "      <td>3.669602</td>\n",
       "      <td>0.664462</td>\n",
       "      <td>44.525658</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AKCGO9726K</th>\n",
       "      <td>0.158892</td>\n",
       "      <td>1143.333333</td>\n",
       "      <td>0.395658</td>\n",
       "      <td>0.970845</td>\n",
       "      <td>0.428829</td>\n",
       "      <td>4.625486</td>\n",
       "      <td>0.597451</td>\n",
       "      <td>0.955882</td>\n",
       "      <td>2.834183</td>\n",
       "      <td>0.999719</td>\n",
       "      <td>0.908425</td>\n",
       "      <td>0.419355</td>\n",
       "      <td>2.904479</td>\n",
       "      <td>0.874818</td>\n",
       "      <td>35.272215</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AOBTM3098N</th>\n",
       "      <td>0.085885</td>\n",
       "      <td>500.666667</td>\n",
       "      <td>0.485470</td>\n",
       "      <td>0.619174</td>\n",
       "      <td>0.629032</td>\n",
       "      <td>4.603918</td>\n",
       "      <td>0.584882</td>\n",
       "      <td>0.935043</td>\n",
       "      <td>3.471419</td>\n",
       "      <td>1.254744</td>\n",
       "      <td>0.846435</td>\n",
       "      <td>0.267819</td>\n",
       "      <td>3.694918</td>\n",
       "      <td>0.518929</td>\n",
       "      <td>72.424682</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>APXEE8191H</th>\n",
       "      <td>0.098634</td>\n",
       "      <td>439.333333</td>\n",
       "      <td>0.515679</td>\n",
       "      <td>0.716237</td>\n",
       "      <td>0.608051</td>\n",
       "      <td>4.590357</td>\n",
       "      <td>0.496452</td>\n",
       "      <td>0.966899</td>\n",
       "      <td>3.312930</td>\n",
       "      <td>0.714465</td>\n",
       "      <td>0.904505</td>\n",
       "      <td>0.258964</td>\n",
       "      <td>3.585094</td>\n",
       "      <td>0.751204</td>\n",
       "      <td>63.522230</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ATDID5461S</th>\n",
       "      <td>0.090689</td>\n",
       "      <td>275.666667</td>\n",
       "      <td>0.533546</td>\n",
       "      <td>0.573156</td>\n",
       "      <td>0.660338</td>\n",
       "      <td>4.427826</td>\n",
       "      <td>0.591373</td>\n",
       "      <td>0.878594</td>\n",
       "      <td>4.125690</td>\n",
       "      <td>0.549301</td>\n",
       "      <td>0.865455</td>\n",
       "      <td>0.315126</td>\n",
       "      <td>3.497262</td>\n",
       "      <td>0.608126</td>\n",
       "      <td>94.935429</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             thrombolysis_rate   admissions   80_plus  onset_known  \\\n",
       "stroke_team                                                          \n",
       "AGNOF1041H            0.154839   671.666667  0.425459     0.635236   \n",
       "AKCGO9726K            0.158892  1143.333333  0.395658     0.970845   \n",
       "AOBTM3098N            0.085885   500.666667  0.485470     0.619174   \n",
       "APXEE8191H            0.098634   439.333333  0.515679     0.716237   \n",
       "ATDID5461S            0.090689   275.666667  0.533546     0.573156   \n",
       "\n",
       "             known_arrival_within_4hrs  onset_arrival_mins_mu  \\\n",
       "stroke_team                                                     \n",
       "AGNOF1041H                    0.681250               4.576874   \n",
       "AKCGO9726K                    0.428829               4.625486   \n",
       "AOBTM3098N                    0.629032               4.603918   \n",
       "APXEE8191H                    0.608051               4.590357   \n",
       "ATDID5461S                    0.660338               4.427826   \n",
       "\n",
       "             onset_arrival_mins_sigma  scan_within_4_hrs  \\\n",
       "stroke_team                                                \n",
       "AGNOF1041H                   0.557598           0.965596   \n",
       "AKCGO9726K                   0.597451           0.955882   \n",
       "AOBTM3098N                   0.584882           0.935043   \n",
       "APXEE8191H                   0.496452           0.966899   \n",
       "ATDID5461S                   0.591373           0.878594   \n",
       "\n",
       "             arrival_scan_arrival_mins_mu  arrival_scan_arrival_mins_sigma  \\\n",
       "stroke_team                                                                  \n",
       "AGNOF1041H                       1.665700                         1.497966   \n",
       "AKCGO9726K                       2.834183                         0.999719   \n",
       "AOBTM3098N                       3.471419                         1.254744   \n",
       "APXEE8191H                       3.312930                         0.714465   \n",
       "ATDID5461S                       4.125690                         0.549301   \n",
       "\n",
       "             onset_scan_4_hrs  eligable  scan_needle_mins_mu  \\\n",
       "stroke_team                                                    \n",
       "AGNOF1041H           0.935867  0.388325             3.669602   \n",
       "AKCGO9726K           0.908425  0.419355             2.904479   \n",
       "AOBTM3098N           0.846435  0.267819             3.694918   \n",
       "APXEE8191H           0.904505  0.258964             3.585094   \n",
       "ATDID5461S           0.865455  0.315126             3.497262   \n",
       "\n",
       "             scan_needle_mins_sigma  hosp_speed  \n",
       "stroke_team                                      \n",
       "AGNOF1041H                 0.664462   44.525658  \n",
       "AKCGO9726K                 0.874818   35.272215  \n",
       "AOBTM3098N                 0.518929   72.424682  \n",
       "APXEE8191H                 0.751204   63.522230  \n",
       "ATDID5461S                 0.608126   94.935429  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "performance.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "b2b8685a-0d35-4d52-9b5d-ef1af6c4af89",
   "metadata": {},
   "outputs": [],
   "source": [
    "rx = scenarios.pivot(\n",
    "    index='stroke_team', columns='scenario', values='thrombolysis_rate')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "7bec2ac9-91b2-4852-8af6-3fe00acd841e",
   "metadata": {},
   "outputs": [],
   "source": [
    "rx = rx.merge(performance[['hosp_speed', 'onset_known']], left_index=True, right_index=True)\n",
    "rx = rx.merge(decision['cohort_rate'], left_index=True, right_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "3a936f0a-a665-40d7-87b1-1bd6a71f4a1b",
   "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>base</th>\n",
       "      <th>benchmark</th>\n",
       "      <th>onset</th>\n",
       "      <th>onset_benchmark</th>\n",
       "      <th>same_patient_characteristics</th>\n",
       "      <th>speed</th>\n",
       "      <th>speed_benchmark</th>\n",
       "      <th>speed_onset</th>\n",
       "      <th>speed_onset_benchmark</th>\n",
       "      <th>hosp_speed</th>\n",
       "      <th>onset_known</th>\n",
       "      <th>cohort_rate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>AGNOF1041H</th>\n",
       "      <td>15.11</td>\n",
       "      <td>20.38</td>\n",
       "      <td>18.17</td>\n",
       "      <td>24.09</td>\n",
       "      <td>11.04</td>\n",
       "      <td>15.21</td>\n",
       "      <td>20.14</td>\n",
       "      <td>17.89</td>\n",
       "      <td>23.90</td>\n",
       "      <td>44.525658</td>\n",
       "      <td>0.635236</td>\n",
       "      <td>27.76</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AKCGO9726K</th>\n",
       "      <td>15.06</td>\n",
       "      <td>14.18</td>\n",
       "      <td>14.92</td>\n",
       "      <td>14.27</td>\n",
       "      <td>22.22</td>\n",
       "      <td>15.32</td>\n",
       "      <td>14.69</td>\n",
       "      <td>15.36</td>\n",
       "      <td>14.83</td>\n",
       "      <td>35.272215</td>\n",
       "      <td>0.970845</td>\n",
       "      <td>37.45</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AOBTM3098N</th>\n",
       "      <td>7.81</td>\n",
       "      <td>11.88</td>\n",
       "      <td>9.39</td>\n",
       "      <td>14.52</td>\n",
       "      <td>8.74</td>\n",
       "      <td>9.39</td>\n",
       "      <td>13.78</td>\n",
       "      <td>11.40</td>\n",
       "      <td>17.17</td>\n",
       "      <td>72.424682</td>\n",
       "      <td>0.619174</td>\n",
       "      <td>26.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>APXEE8191H</th>\n",
       "      <td>10.08</td>\n",
       "      <td>13.06</td>\n",
       "      <td>10.76</td>\n",
       "      <td>13.35</td>\n",
       "      <td>13.24</td>\n",
       "      <td>10.15</td>\n",
       "      <td>12.80</td>\n",
       "      <td>10.90</td>\n",
       "      <td>13.54</td>\n",
       "      <td>63.522230</td>\n",
       "      <td>0.716237</td>\n",
       "      <td>29.97</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ATDID5461S</th>\n",
       "      <td>9.20</td>\n",
       "      <td>9.92</td>\n",
       "      <td>11.81</td>\n",
       "      <td>13.35</td>\n",
       "      <td>7.60</td>\n",
       "      <td>11.10</td>\n",
       "      <td>11.79</td>\n",
       "      <td>14.64</td>\n",
       "      <td>15.95</td>\n",
       "      <td>94.935429</td>\n",
       "      <td>0.573156</td>\n",
       "      <td>25.92</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             base  benchmark  onset  onset_benchmark  \\\n",
       "AGNOF1041H  15.11      20.38  18.17            24.09   \n",
       "AKCGO9726K  15.06      14.18  14.92            14.27   \n",
       "AOBTM3098N   7.81      11.88   9.39            14.52   \n",
       "APXEE8191H  10.08      13.06  10.76            13.35   \n",
       "ATDID5461S   9.20       9.92  11.81            13.35   \n",
       "\n",
       "            same_patient_characteristics  speed  speed_benchmark  speed_onset  \\\n",
       "AGNOF1041H                         11.04  15.21            20.14        17.89   \n",
       "AKCGO9726K                         22.22  15.32            14.69        15.36   \n",
       "AOBTM3098N                          8.74   9.39            13.78        11.40   \n",
       "APXEE8191H                         13.24  10.15            12.80        10.90   \n",
       "ATDID5461S                          7.60  11.10            11.79        14.64   \n",
       "\n",
       "            speed_onset_benchmark  hosp_speed  onset_known  cohort_rate  \n",
       "AGNOF1041H                  23.90   44.525658     0.635236        27.76  \n",
       "AKCGO9726K                  14.83   35.272215     0.970845        37.45  \n",
       "AOBTM3098N                  17.17   72.424682     0.619174        26.00  \n",
       "APXEE8191H                  13.54   63.522230     0.716237        29.97  \n",
       "ATDID5461S                  15.95   94.935429     0.573156        25.92  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rx.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3f657cda-6be1-4355-b343-399be0c94534",
   "metadata": {},
   "source": [
    "## Calculate difference between each hospital's thrombolysis rate and the mean thrombolysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "4793ecc7-85cd-445a-8865-15705b079c85",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean thrombolysis: 11.22\n"
     ]
    }
   ],
   "source": [
    "mean_rx = rx['base'].mean()\n",
    "print (f'Mean thrombolysis: {mean_rx:0.2f}')\n",
    "\n",
    "rx['diff_from_mean'] = rx['base'] - mean_rx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "506d0ff9-5190-4813-acb8-bc6c7c7253b9",
   "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>base</th>\n",
       "      <th>benchmark</th>\n",
       "      <th>onset</th>\n",
       "      <th>onset_benchmark</th>\n",
       "      <th>same_patient_characteristics</th>\n",
       "      <th>speed</th>\n",
       "      <th>speed_benchmark</th>\n",
       "      <th>speed_onset</th>\n",
       "      <th>speed_onset_benchmark</th>\n",
       "      <th>hosp_speed</th>\n",
       "      <th>onset_known</th>\n",
       "      <th>cohort_rate</th>\n",
       "      <th>diff_from_mean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>AGNOF1041H</th>\n",
       "      <td>15.11</td>\n",
       "      <td>20.38</td>\n",
       "      <td>18.17</td>\n",
       "      <td>24.09</td>\n",
       "      <td>11.04</td>\n",
       "      <td>15.21</td>\n",
       "      <td>20.14</td>\n",
       "      <td>17.89</td>\n",
       "      <td>23.90</td>\n",
       "      <td>44.525658</td>\n",
       "      <td>0.635236</td>\n",
       "      <td>27.76</td>\n",
       "      <td>3.889545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AKCGO9726K</th>\n",
       "      <td>15.06</td>\n",
       "      <td>14.18</td>\n",
       "      <td>14.92</td>\n",
       "      <td>14.27</td>\n",
       "      <td>22.22</td>\n",
       "      <td>15.32</td>\n",
       "      <td>14.69</td>\n",
       "      <td>15.36</td>\n",
       "      <td>14.83</td>\n",
       "      <td>35.272215</td>\n",
       "      <td>0.970845</td>\n",
       "      <td>37.45</td>\n",
       "      <td>3.839545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AOBTM3098N</th>\n",
       "      <td>7.81</td>\n",
       "      <td>11.88</td>\n",
       "      <td>9.39</td>\n",
       "      <td>14.52</td>\n",
       "      <td>8.74</td>\n",
       "      <td>9.39</td>\n",
       "      <td>13.78</td>\n",
       "      <td>11.40</td>\n",
       "      <td>17.17</td>\n",
       "      <td>72.424682</td>\n",
       "      <td>0.619174</td>\n",
       "      <td>26.00</td>\n",
       "      <td>-3.410455</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>APXEE8191H</th>\n",
       "      <td>10.08</td>\n",
       "      <td>13.06</td>\n",
       "      <td>10.76</td>\n",
       "      <td>13.35</td>\n",
       "      <td>13.24</td>\n",
       "      <td>10.15</td>\n",
       "      <td>12.80</td>\n",
       "      <td>10.90</td>\n",
       "      <td>13.54</td>\n",
       "      <td>63.522230</td>\n",
       "      <td>0.716237</td>\n",
       "      <td>29.97</td>\n",
       "      <td>-1.140455</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ATDID5461S</th>\n",
       "      <td>9.20</td>\n",
       "      <td>9.92</td>\n",
       "      <td>11.81</td>\n",
       "      <td>13.35</td>\n",
       "      <td>7.60</td>\n",
       "      <td>11.10</td>\n",
       "      <td>11.79</td>\n",
       "      <td>14.64</td>\n",
       "      <td>15.95</td>\n",
       "      <td>94.935429</td>\n",
       "      <td>0.573156</td>\n",
       "      <td>25.92</td>\n",
       "      <td>-2.020455</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             base  benchmark  onset  onset_benchmark  \\\n",
       "AGNOF1041H  15.11      20.38  18.17            24.09   \n",
       "AKCGO9726K  15.06      14.18  14.92            14.27   \n",
       "AOBTM3098N   7.81      11.88   9.39            14.52   \n",
       "APXEE8191H  10.08      13.06  10.76            13.35   \n",
       "ATDID5461S   9.20       9.92  11.81            13.35   \n",
       "\n",
       "            same_patient_characteristics  speed  speed_benchmark  speed_onset  \\\n",
       "AGNOF1041H                         11.04  15.21            20.14        17.89   \n",
       "AKCGO9726K                         22.22  15.32            14.69        15.36   \n",
       "AOBTM3098N                          8.74   9.39            13.78        11.40   \n",
       "APXEE8191H                         13.24  10.15            12.80        10.90   \n",
       "ATDID5461S                          7.60  11.10            11.79        14.64   \n",
       "\n",
       "            speed_onset_benchmark  hosp_speed  onset_known  cohort_rate  \\\n",
       "AGNOF1041H                  23.90   44.525658     0.635236        27.76   \n",
       "AKCGO9726K                  14.83   35.272215     0.970845        37.45   \n",
       "AOBTM3098N                  17.17   72.424682     0.619174        26.00   \n",
       "APXEE8191H                  13.54   63.522230     0.716237        29.97   \n",
       "ATDID5461S                  15.95   94.935429     0.573156        25.92   \n",
       "\n",
       "            diff_from_mean  \n",
       "AGNOF1041H        3.889545  \n",
       "AKCGO9726K        3.839545  \n",
       "AOBTM3098N       -3.410455  \n",
       "APXEE8191H       -1.140455  \n",
       "ATDID5461S       -2.020455  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rx.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7ecaff49-3863-4e9a-a83a-0d3f49df6e7d",
   "metadata": {},
   "source": [
    "## How much variation is explained by differences in decision making?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "51c22917-e171-4e04-8486-3a6000acd465",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.399\n"
     ]
    }
   ],
   "source": [
    "diff_explained_by_decison_making = (np.corrcoef(\n",
    "    rx['cohort_rate'], rx['diff_from_mean'])[1,0]) ** 2\n",
    "\n",
    "print(f'{diff_explained_by_decison_making:0.3}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1eb4d10e-cced-4442-bffd-e865bea40615",
   "metadata": {},
   "source": [
    "## How much variation is explained by differences in speed?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "94a68a29-5267-4224-bc92-d65fca7c4154",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.254\n"
     ]
    }
   ],
   "source": [
    "diff_explained_by_speed = (np.corrcoef(\n",
    "    rx['diff_from_mean'], rx['hosp_speed'])[1,0]) ** 2\n",
    "\n",
    "print(f'{diff_explained_by_speed:0.3}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3378bad2-f690-4c00-88b0-e8de8d38d763",
   "metadata": {},
   "source": [
    "## How much variation is explained by determination of stroke onset time?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "94cc637f-424b-4d9f-a739-cab3232b4b5c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0474\n"
     ]
    }
   ],
   "source": [
    "diff_explained_by_determination_of_onset = (np.corrcoef(\n",
    "    rx['diff_from_mean'], rx['onset_known'])[1,0]) ** 2\n",
    "\n",
    "print(f'{diff_explained_by_determination_of_onset:0.3}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cdc97983-ec89-4a2c-bef7-48dedd30c907",
   "metadata": {},
   "source": [
    "## Plot relationships"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "1b226988-c9b7-4836-ba58-b5b69f9d2bba",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from scipy import stats\n",
    "fig = plt.figure(figsize=(15,6))\n",
    "\n",
    "ax1 = fig.add_subplot(131)\n",
    "x = rx['cohort_rate']\n",
    "y = rx['diff_from_mean']\n",
    "gradient, intercept, r_value, p_value, std_err = stats.linregress(x,y)\n",
    "y_fit=intercept + (x*gradient)\n",
    "plt.plot(x, y, 'o', label='original data')\n",
    "plt.plot(x, y_fit, 'r', label='fitted line')\n",
    "text='Slope: %.2f\\nR-Squared: %.3f\\nP value: %.3f' %(gradient,r_value**2,p_value)\n",
    "plt.text(35, -8, text)\n",
    "ax1.set_title('Relationship between thrombolysis use and\\ndecision making')\n",
    "ax1.set_xlabel('Predicted 10k cohort thrombolysis rate')\n",
    "ax1.set_ylabel('Difference in thrombolysis from mean')\n",
    "\n",
    "ax2 = fig.add_subplot(132)\n",
    "x = rx['hosp_speed']\n",
    "y = rx['diff_from_mean']\n",
    "gradient, intercept, r_value, p_value, std_err = stats.linregress(x,y)\n",
    "y_fit=intercept + (x*gradient)\n",
    "plt.plot(x, y, 'o', label='original data')\n",
    "plt.plot(x, y_fit, 'r', label='fitted line')\n",
    "text='Slope: %.2f\\nR-Squared: %.3f\\nP value: %.3f' %(gradient,r_value**2,p_value)\n",
    "plt.text(77, 10, text)\n",
    "ax2.set_title('Relationship between thrombolysis use and\\npathway speed')\n",
    "ax2.set_xlabel('Arrival to scan + scan to needle (mins)')\n",
    "ax2.set_ylabel('Difference in thrombolysis from mean')\n",
    "\n",
    "ax3 = fig.add_subplot(133)\n",
    "x = rx['onset_known']\n",
    "y = rx['diff_from_mean']\n",
    "gradient, intercept, r_value, p_value, std_err = stats.linregress(x,y)\n",
    "y_fit=intercept + (x*gradient)\n",
    "plt.plot(x, y, 'o', label='original data')\n",
    "plt.plot(x, y_fit, 'r', label='fitted line')\n",
    "text='Slope: %.2f\\nR-Squared: %.3f\\nP value: %.3f' %(gradient,r_value**2,p_value)\n",
    "plt.text(0.38,10, text)\n",
    "ax3.set_title('Relationship between thrombolysis use and\\ndetermination of stroke onset')\n",
    "ax3.set_xlabel('Proportion patients with determined stroke onset')\n",
    "ax3.set_ylabel('Difference in thrombolysis from mean')\n",
    "\n",
    "\n",
    "plt.tight_layout(pad=2)\n",
    "plt.savefig('./output/model_correlations.jpg', dpi=300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5131da5e-0164-4b3a-8e6f-4dcf371da328",
   "metadata": {},
   "source": [
    "## Conclusions\n",
    "\n",
    "* In-hospital process parameters partly explain the inter-hospital variation in thrombolysis use.\n",
    "\n",
    "* The strongest relationship is between decision-making, as described by the predicted thrombolysis use of a standard 10k cohort of patients, with R-square of 0.40.\n",
    "\n",
    "* Pathway speed is the next strongest predictor of thrombolysis use, with an R-square of 0.25.\n",
    "\n",
    "* Determination of stroke onset time is the weakest predictor of thrombolysis use (R-square of 0.05), but is still statistically significant (p=0.012)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}