analysis/generate_results_jun.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sn\n",
    "import numpy as np\n",
    "from datetime import datetime, timedelta\n",
    "os.chdir('../')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 289,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('output/input_2021-06-01.csv.gz', parse_dates=[\"ae_attendance_date\", \"ae_attendance_hosp_discharge_date\", \"emergency_covid_hospital_admission_date\", \"ae_before_date\", \"ae_before_with_hospital_discharge_date\"])\n",
    "from pandas.testing import assert_frame_equal\n",
    "if not os.path.exists('output/jun'):\n",
    "    os.mkdir('output/jun')\n",
    "\n",
    "def produce_matrix(df_list, names, filename, size=(7, 7)):\n",
    "   \n",
    "    confusion_matrix = []\n",
    "    for i in df_list:\n",
    "        \n",
    "        \n",
    "        row = []\n",
    "        for j in df_list:\n",
    "            equal = len(set(i['patient_id']) & set(j['patient_id']))\n",
    "            row.append(int(round_to_5(equal)))\n",
    "\n",
    "        confusion_matrix.append(row)\n",
    "    \n",
    "\n",
    "    plt.figure(figsize=size)\n",
    "    mask = np.zeros_like(confusion_matrix)\n",
    "    mask[np.triu_indices_from(mask, k=1)] = True\n",
    "\n",
    "    with sn.axes_style(\"white\"):\n",
    "\n",
    "        sn.heatmap(confusion_matrix, mask=mask, annot=True, fmt=\"d\", xticklabels=names, yticklabels=names)\n",
    "        plt.yticks(rotation=0) \n",
    "        plt.savefig(f\"output/jun/{filename}.jpeg\", bbox_inches='tight')\n",
    "        plt.show()\n",
    "        \n",
    "def round_to_5(number):\n",
    "    return 5 * round((number / 5))\n",
    "\n",
    "def performance_metrics(tp, fp, tn, fn):\n",
    "    \n",
    "    tp = round_to_5(len(tp))\n",
    "    fp = round_to_5(len(fp))\n",
    "    tn = round_to_5(len(tn))\n",
    "    fn = round_to_5(len(fn))\n",
    "    \n",
    "    if (tp+fn) > 0:\n",
    "        sensitivity = round((tp/(tp+fn)) *100, 2)\n",
    "    else:\n",
    "        sensitivity = 0\n",
    "        \n",
    "    if (tn+fp) > 0:\n",
    "        specificity = round((tn/(tn+fp)) *100, 2)\n",
    "    else:\n",
    "        specificity = 0\n",
    "    \n",
    "    if (fp+tp) > 0:\n",
    "        fdr = round((fp/(fp+tp)) *100, 2)\n",
    "    else:\n",
    "        fdr=0\n",
    "        \n",
    "    return [tp, fp, tn, fn,sensitivity, specificity, fdr]\n",
    "\n",
    "\n",
    "ae_discharge_dict = {\n",
    "    1066341000000100: \"Ambulatory Emergency Care\",\n",
    "    19712007: \"Patient Transfer\",\n",
    "    183919006: \"Hospice\",\n",
    "    1066361000000104: \"High dependency unit\",\n",
    "    305398007: \"Mortuary\",\n",
    "    1066381000000108: \"Special baby care unit\",\n",
    "    1066331000000109: \"Emergency department short stay ward\",\n",
    "    306705005: \"Police custody\",\n",
    "    306706006: \"Ward\",\n",
    "    306689006: \"Home\",\n",
    "    306694006: \"Nursing Home\",\n",
    "    306691003: \"Residential Home\",\n",
    "    1066351000000102: \"Hospital at home\",\n",
    "    1066401000000108: \"Neonatal ICU\",\n",
    "    1066371000000106: \"Coronary Care Unit\",\n",
    "    50861005: \"Legal Custody\",\n",
    "    1066391000000105: \"ICU\",\n",
    "    \"missing\": \"missing\",\n",
    "}\n",
    "\n",
    "hosp_discharge_list = [str(306706006), str(1066331000000109), str(1066391000000105), str(1066361000000104)]\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Number of emergency covid hospital admissions and those that can be identified in A&E attendances\n",
    "### Any A&E attendance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 290,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of emergency covid hospital admissions: 800\n",
      "Number with A&E attendance before admission: 500 (62.5%)\n",
      "---\n",
      "Number of A&E attendances: 400\n"
     ]
    }
   ],
   "source": [
    "num_emergency_cov_admissions = round_to_5(df['emergency_covid_hospital_admission_date'].count())\n",
    "ae_before = (df.loc[\n",
    "    df[\"ae_before_date\"].notnull()\n",
    ", \"patient_id\"])\n",
    "print(f'Number of emergency covid hospital admissions: {num_emergency_cov_admissions}')\n",
    "print(f'Number with A&E attendance before admission: {round_to_5(len(ae_before))} ({round((round_to_5(len(ae_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "ae = round_to_5(len(df[\n",
    "    (df[\"ae_attendance_date\"].notnull())\n",
    "]))\n",
    "\n",
    "ae_with_hosp_discharge = round_to_5(len(df[\n",
    "    (df[\"ae_attendance_hosp_discharge_date\"].notnull())\n",
    "]))\n",
    "\n",
    "summary_table = pd.DataFrame({\n",
    "    \"emergency covid hospital admission\": num_emergency_cov_admissions,\n",
    "    \"ae attendance before admission\": round_to_5(len(ae_before)),\n",
    "    \"ae_attendance\": ae, \n",
    "    \"ae attendance with hospital discharge\": ae_with_hosp_discharge,\n",
    "    \n",
    "}, index=[\"count\"])\n",
    "\n",
    "summary_table.to_csv(\"output/jun/summary_table.csv\")\n",
    "\n",
    "print('---')\n",
    "print(f'Number of A&E attendances: {ae}')\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "tp_ae = df[\n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())  \n",
    "]\n",
    "\n",
    "fn_ae_hosp_discharge = df[\n",
    "    ~(\n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_hosp_discharge = df[\n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    ) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_hosp_discharge = df[\n",
    "    ~(\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    ) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "ae_stats = performance_metrics(tp_ae, fp_ae, tn_ae, fn_ae)\n",
    "\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_ae))} ({ae_stats[6]}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 1 - Discharge destination\n",
    "Restrict to those that attended A&E and had a hospital discharge destination (hospital ward, emergency short stay ward, icu)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 291,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with hospital discharge before admission: 500 (62.5%)\n",
      "False positives (FDR): 0 (0%)\n"
     ]
    }
   ],
   "source": [
    "ae_hosp_before = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_date\"].notnull()\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with hospital discharge before admission: {round_to_5(len(ae_hosp_before))} ({round((round_to_5(len(ae_hosp_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_hosp_discharge = df[\n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())  \n",
    "]\n",
    "\n",
    "fn_ae_hosp_discharge = df[\n",
    "    ~(\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_hosp_discharge = df[\n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_hosp_discharge = df[\n",
    "    ~(df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "discharge_destination = performance_metrics(tp_ae_hosp_discharge, fp_ae_hosp_discharge, tn_ae_hosp_discharge, fn_ae_hosp_discharge)\n",
    "\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_ae_hosp_discharge))} ({discharge_destination[6]}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 2 - Diagnosis code\n",
    "Restrict to those with a COVID-19 diagnosis code or respiratory-related diagnosis code from the [emergency care diagnosis refset](https://termbrowser.nhs.uk/?perspective=full&conceptId1=991411000000109&edition=uk-edition&release=v20220216&server=https://termbrowser.nhs.uk/sct-browser-api/snomed&langRefset=900000000000508004)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 292,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance and covid diagnosis code before admission: 500 (62.5%)\n",
      "Number with A&E attendance and respiratory code before admission: 500 (62.5%)\n",
      "False positives (FDR) - A&E attendance and covid diagnosis code: 105 (84.0%)\n",
      "False positives (FDR) - A&E attendance and respiratory diagnosis code: 100 (83.33%)\n"
     ]
    }
   ],
   "source": [
    "ae_cov_before = (df.loc[\n",
    "    df[\"ae_before_with_covid\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "ae_resp_before = (df.loc[\n",
    "    df[\"ae_before_with_resp\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance and covid diagnosis code before admission: {round_to_5(len(ae_cov_before))} ({round((round_to_5(len(ae_cov_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "print(f'Number with A&E attendance and respiratory code before admission: {round_to_5(len(ae_resp_before))} ({round((round_to_5(len(ae_resp_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_cov = df[\n",
    "    (df[\"ae_cov\"]==1) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_cov = df[\n",
    "    ~(\n",
    "        (df[\"ae_cov\"]==1) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_cov = df[\n",
    "    (df[\"ae_cov\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_cov = df[\n",
    "    ~(df[\"ae_cov\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "tp_ae_resp = df[\n",
    "    (df[\"ae_resp\"]==1) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_resp = df[\n",
    "    ~(\n",
    "        (df[\"ae_resp\"]==1) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_resp = df[\n",
    "    (df[\"ae_resp\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_resp = df[\n",
    "    ~(df[\"ae_resp\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "diagnosis_code_cov = performance_metrics(tp_ae_cov, fp_ae_cov, tn_ae_cov, fn_ae_cov)\n",
    "diagnosis_code_resp = performance_metrics(tp_ae_resp, fp_ae_resp, tn_ae_resp, fn_ae_resp)\n",
    "\n",
    "print(f'False positives (FDR) - A&E attendance and covid diagnosis code: {round_to_5(len(fp_ae_cov))} ({diagnosis_code_cov[6]}%)')\n",
    "print(f'False positives (FDR) - A&E attendance and respiratory diagnosis code: {round_to_5(len(fp_ae_resp))} ({diagnosis_code_resp[6]}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 3 - Identification in primary care\n",
    "\n",
    "Restrict to those with a probable or suspected COVID-19 code recorded in primary care in the 2 weeks prior to A&E attendance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 293,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance and recent suspected covid code in primary care before admission: 500 (62.5%)\n",
      "Number with A&E attendance and recent probable covid code in primary care before admission: 500 (62.5%)\n",
      "False positives (FDR) - A&E attendance and probable covid code in primary care: 82.61%)\n",
      "False positives (FDR) - A&E attendance and suspected covid code in primary care: 86.36%)\n"
     ]
    }
   ],
   "source": [
    "ae_prim_care_probable_before = (df.loc[\n",
    "    df[\"ae_before_with_prim_care_probable\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "ae_prim_care_suspected_before = (df.loc[\n",
    "    df[\"ae_before_with_prim_care_suspected\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance and recent suspected covid code in primary care before admission: {round_to_5(len(ae_prim_care_suspected_before))} ({round((round_to_5(len(ae_prim_care_suspected_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "print(f'Number with A&E attendance and recent probable covid code in primary care before admission: {round_to_5(len(ae_prim_care_probable_before))} ({round((round_to_5(len(ae_prim_care_probable_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_suspected = df[\n",
    "    (df[\"suspected_covid_primary_care_before_ae\"]==1) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_suspected = df[\n",
    "    ~(\n",
    "        (df[\"suspected_covid_primary_care_before_ae\"]==1) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_suspected = df[\n",
    "    (df[\"suspected_covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_suspected = df[\n",
    "    ~(df[\"suspected_covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "tp_ae_probable = df[\n",
    "    (df[\"covid_primary_care_before_ae\"]==1) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_probable = df[\n",
    "    ~(\n",
    "        (df[\"covid_primary_care_before_ae\"]==1) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_probable = df[\n",
    "    (df[\"covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_probable = df[\n",
    "    ~(df[\"covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "prim_care_suspected = performance_metrics(tp_ae_suspected, fp_ae_suspected, tn_ae_suspected, fn_ae_suspected)\n",
    "prim_care_probable = performance_metrics(tp_ae_probable, fp_ae_probable, tn_ae_probable, fn_ae_probable)\n",
    "print(f'False positives (FDR) - A&E attendance and probable covid code in primary care: {prim_care_probable[6]}%)')\n",
    "print(f'False positives (FDR) - A&E attendance and suspected covid code in primary care: {prim_care_suspected[6]}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 4 - Positive test result\n",
    "\n",
    "Restrict to those with a positive COVID-19 test result recorded up to 4 weeks prior to A&E attendance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 294,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance and recent positive test (2 weeks) before admission: 500 (62.5%)\n",
      "Number with A&E attendance and recent positive test (4 weeks) before admissio: 500 (62.5%)\n",
      "False positives (FDR) - A&E attendance and recent positive test (2 weeks): 100 (83.33%)\n",
      "False positives (FDR) - A&E attendance and recent positive test (4 weeks): 95 (86.36%)\n"
     ]
    }
   ],
   "source": [
    "ae_pos_test_before_2_weeks = (df.loc[\n",
    "    df[\"ae_before_with_positive_covid_test_2_weeks\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "ae_pos_test_before = (df.loc[\n",
    "    df[\"ae_before_with_positive_covid_test\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "\n",
    "print(f'Number with A&E attendance and recent positive test (2 weeks) before admission: {round_to_5(len(ae_pos_test_before_2_weeks))} ({round((round_to_5(len(ae_pos_test_before_2_weeks))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "print(f'Number with A&E attendance and recent positive test (4 weeks) before admissio: {round_to_5(len(ae_pos_test_before))} ({round((round_to_5(len(ae_pos_test_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_test = df[\n",
    "    (df[\"pos_test\"]==1) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_test = df[\n",
    "    ~(\n",
    "        (df[\"pos_test\"]==1) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_test = df[\n",
    "    (df[\"pos_test\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_test = df[\n",
    "    ~(df[\"pos_test\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tp_ae_test_2_weeks = df[\n",
    "    (df[\"pos_test_2_weeks\"]==1) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_test_2_weeks = df[\n",
    "    ~(\n",
    "        (df[\"pos_test_2_weeks\"]==1) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_test_2_weeks = df[\n",
    "    (df[\"pos_test_2_weeks\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_test_2_weeks = df[\n",
    "    ~(df[\"pos_test_2_weeks\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "positive_test_2_weeks = performance_metrics(tp_ae_test_2_weeks, fp_ae_test_2_weeks, tn_ae_test_2_weeks, fn_ae_test_2_weeks)\n",
    "positive_test_4_weeks = performance_metrics(tp_ae_test, fp_ae_test, tn_ae_test, fn_ae_test)\n",
    "\n",
    "print(f'False positives (FDR) - A&E attendance and recent positive test (2 weeks): {round_to_5(len(fp_ae_test_2_weeks))} ({positive_test_2_weeks[6]}%)')\n",
    "print(f'False positives (FDR) - A&E attendance and recent positive test (4 weeks): {round_to_5(len(fp_ae_test))} ({positive_test_4_weeks[6]}%)')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "####  Combining approaches"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Diagnosis codes (resp or cov) with recent positive test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 295,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with either covid or respiratory diagnosis code and recent positive test (2 weeks) before admission: 375 (46.88%)\n",
      "False positives (FDR): 75 (83.33%)\n"
     ]
    }
   ],
   "source": [
    "ae_hos_resp_cov_plus_test_2_weeks = (df.loc[(\n",
    "        (df[\"ae_before_with_covid\"]==1) | \n",
    "        (df[\"ae_before_with_resp\"]==1) \n",
    "    )&\n",
    "    (df[\"ae_before_with_positive_covid_test_2_weeks\"]==1)\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with either covid or respiratory diagnosis code and recent positive test (2 weeks) before admission: {round_to_5(len(ae_hos_resp_cov_plus_test_2_weeks))} ({round((round_to_5(len(ae_hos_resp_cov_plus_test_2_weeks))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)    \n",
    "        )&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    ~(\n",
    "        (\n",
    "            (\n",
    "                (df[\"ae_cov\"]==1) |\n",
    "                (df[\"ae_resp\"]==1)    \n",
    "            )&\n",
    "            (df[\"pos_test_2_weeks\"]==1)\n",
    "        ) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)    \n",
    "        )&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    ~((\n",
    "        (\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)    \n",
    "        )&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "code_plus_test = performance_metrics(tp_ae_hos_resp_cov_plus_test_2_weeks, fp_ae_hos_resp_cov_plus_test_2_weeks, tn_ae_hos_resp_cov_plus_test_2_weeks, fn_ae_hos_resp_cov_plus_test_2_weeks)\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_ae_hos_resp_cov_plus_test_2_weeks))} ({code_plus_test[6]}%)')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Covid in primary care with recent positive test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 296,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with suspected or probable covid in primary care and recent positive test (2 weeks) before admission: 380 (47.5%)\n",
      "False positives (FDR): 75 (83.33%)\n"
     ]
    }
   ],
   "source": [
    "ae_hosp_prim_care_before_plus_2_week_test = (df.loc[\n",
    "    \n",
    "    \n",
    "    (\n",
    "        (df[\"ae_before_with_prim_care_suspected\"]==1) | \n",
    "        (df[\"ae_before_with_prim_care_probable\"]==1) \n",
    "    )\n",
    "&\n",
    "    (df[\"ae_before_with_positive_covid_test_2_weeks\"]==1)\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with suspected or probable covid in primary care and recent positive test (2 weeks) before admission: {round_to_5(len(ae_hosp_prim_care_before_plus_2_week_test))} ({round((round_to_5(len(ae_hosp_prim_care_before_plus_2_week_test))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "        )&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    ~(\n",
    "        (\n",
    "            (\n",
    "                (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "                (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "            )&\n",
    "            (df[\"pos_test_2_weeks\"]==1)\n",
    "        )& \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    ((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "        )&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )\n",
    "    ) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    ~((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "        )&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "\n",
    "primary_care_plus_test = performance_metrics(tp_ae_hosp_prim_care_before_plus_2_week_test, fp_ae_hosp_prim_care_before_plus_2_week_test, tn_ae_hosp_prim_care_before_plus_2_week_test, fn_ae_hosp_prim_care_before_plus_2_week_test)\n",
    "\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_ae_hosp_prim_care_before_plus_2_week_test))} ({primary_care_plus_test[6]}%)')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Recent positive test and either covid in primary care or diagnosis code (resp or cov)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 297,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with recent positive test and either suspected or probable covid in primary care or a covid or respiratory diagnosis code on attendance: 470 (58.75%)\n",
      "False positives (FDR): 95 (82.61%)\n"
     ]
    }
   ],
   "source": [
    "combined_before = (df.loc[\n",
    "        (df[\"ae_before_with_positive_covid_test_2_weeks\"]==1) &\n",
    "        (\n",
    "            (df[\"ae_before_with_prim_care_suspected\"]==1) | \n",
    "            (df[\"ae_before_with_prim_care_probable\"]==1) |\n",
    "            (df[\"ae_before_with_covid\"]==1) | \n",
    "            (df[\"ae_before_with_resp\"]==1)   \n",
    "        )\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with recent positive test and either suspected or probable covid in primary care or a covid or respiratory diagnosis code on attendance: {round_to_5(len(combined_before))} ({round((round_to_5(len(combined_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_combined_before = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_combined_before = df[\n",
    "    ~(\n",
    "        (\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "        )& \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_combined_before = df[\n",
    "    ((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )\n",
    "    ) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_combined_before = df[\n",
    "    ~((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "\n",
    "pos_test_prim_care_or_diagnosis_code = performance_metrics(tp_combined_before, fp_combined_before, tn_combined_before, fn_combined_before)\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_combined_before))} ({pos_test_prim_care_or_diagnosis_code[6]}%)')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 298,
   "metadata": {},
   "outputs": [],
   "source": [
    "identification_table_index = [\"tp\", \"fp\", \"tn\", \"fn\",\"sensitivity\", \"specificity\", \"fdr\"]\n",
    "identification_table = pd.DataFrame({\n",
    "    \"discharge_destination\": pd.Series(discharge_destination,index=identification_table_index),\n",
    "    \"covid_diagnosis_code\": pd.Series(diagnosis_code_cov,index=identification_table_index),\n",
    "    \"resp_diagnosis_code\": pd.Series(diagnosis_code_resp,index=identification_table_index),\n",
    "    \"primary_care_suspected\": pd.Series(prim_care_suspected,index=identification_table_index),\n",
    "    \"primary_care_probable\": pd.Series(prim_care_probable,index=identification_table_index),\n",
    "    \"positive_test_2_weeks\": pd.Series(positive_test_2_weeks,index=identification_table_index),\n",
    "    \"positive_test_4_weeks\": pd.Series(positive_test_4_weeks,index=identification_table_index),\n",
    "    \"diagnosis_code_with_positive_test\": pd.Series(code_plus_test,index=identification_table_index),\n",
    "    \"primary_care_with_positive_test\": pd.Series(primary_care_plus_test,index=identification_table_index),\n",
    "    \"pos_test_prim_care_or_diagnosis\": pd.Series(pos_test_prim_care_or_diagnosis_code,index=identification_table_index)\n",
    "    \n",
    "})\n",
    "identification_table.to_csv(\"output/jun/identification_table_no_hosp_discharge.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 299,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "produce_matrix(df_list = [ae_stats, tp_ae_hosp_discharge, tp_ae_cov, tp_ae_resp, tp_ae_suspected, tp_ae_probable, tp_ae_test, tp_ae_test_2_weeks, tp_ae_hos_resp_cov_plus_test_2_weeks, tp_ae_hosp_prim_care_before_plus_2_week_test], names=[\"AE\",\"Hospital discharge\", \"Covid diagnosis code\", \"Respiratory diagnosis code\", \"Probable covid code\", \"Suspected covid code\", \"Recent positive test (2 weeks)\", \"Recent positive test (4 weeks)\", \"Diagnosis codes (resp or cov) with recent positive test\", \"Covid in primary care with recent positive test\", \"Recent positive test and either covid in primary care or diagnosis code (resp or cov)\"], filename=\"cm_no_hosp_discharge\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Restrict to A&E attendance with hospital discharge"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of A&E attendances with hospital discharge: 400\n"
     ]
    }
   ],
   "source": [
    "ae_with_hosp_discharge = round_to_5(len(df[\n",
    "    (df[\"ae_attendance_hosp_discharge_date\"].notnull())\n",
    "]))\n",
    "print(f'Number of A&E attendances with hospital discharge: {ae_with_hosp_discharge}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 1 - Discharge destination\n",
    "Restrict to those that attended A&E and had a hospital discharge destination (hospital ward, emergency short stay ward, icu)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with hospital discharge before admission: 500 (62.5%)\n"
     ]
    },
    {
     "ename": "ZeroDivisionError",
     "evalue": "division by zero",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/1777310896.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     39\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     40\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 41\u001b[0;31m \u001b[0mdischarge_destination\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_hosp_discharge\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_hosp_discharge\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_hosp_discharge\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_hosp_discharge\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     42\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf'False positives (FDR): {round_to_5(len(fp_ae_hosp_discharge))} ({discharge_destination[6]}%)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/3350781174.py\u001b[0m in \u001b[0;36mperformance_metrics\u001b[0;34m(tp, fp, tn, fn)\u001b[0m\n\u001b[1;32m     38\u001b[0m     \u001b[0msensitivity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     39\u001b[0m     \u001b[0mspecificity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 40\u001b[0;31m     \u001b[0mfdr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     41\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msensitivity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mspecificity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfdr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     42\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
     ]
    }
   ],
   "source": [
    "ae_hosp_before = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_date\"].notnull()\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with hospital discharge before admission: {round_to_5(len(ae_hosp_before))} ({round((round_to_5(len(ae_hosp_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_hosp_discharge = df[\n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_hosp_discharge = df[\n",
    "    ~(\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_hosp_discharge = df[\n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_hosp_discharge = df[\n",
    "    ~(df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "discharge_destination = performance_metrics(tp_ae_hosp_discharge, fp_ae_hosp_discharge, tn_ae_hosp_discharge, fn_ae_hosp_discharge)\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_ae_hosp_discharge))} ({discharge_destination[6]}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 2 - Diagnosis code\n",
    "Restrict to those with a COVID-19 diagnosis code or respiratory-related diagnosis code from the [emergency care diagnosis refset](https://termbrowser.nhs.uk/?perspective=full&conceptId1=991411000000109&edition=uk-edition&release=v20220216&server=https://termbrowser.nhs.uk/sct-browser-api/snomed&langRefset=900000000000508004)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance and covid diagnosis code before admission: 500 (62.5%)\n",
      "Number with A&E attendance and respiratory code before admission: 500 (62.5%)\n"
     ]
    },
    {
     "ename": "ZeroDivisionError",
     "evalue": "division by zero",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/2422805883.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     89\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     90\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 91\u001b[0;31m \u001b[0mdiagnosis_code_cov\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_cov\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_cov\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_cov\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_cov\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     92\u001b[0m \u001b[0mdiagnosis_code_resp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_resp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_resp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_resp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_resp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     93\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/3350781174.py\u001b[0m in \u001b[0;36mperformance_metrics\u001b[0;34m(tp, fp, tn, fn)\u001b[0m\n\u001b[1;32m     38\u001b[0m     \u001b[0msensitivity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     39\u001b[0m     \u001b[0mspecificity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 40\u001b[0;31m     \u001b[0mfdr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     41\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msensitivity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mspecificity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfdr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     42\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
     ]
    }
   ],
   "source": [
    "ae_cov_before = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_covid\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "ae_resp_before = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_resp\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance and covid diagnosis code before admission: {round_to_5(len(ae_cov_before))} ({round((round_to_5(len(ae_cov_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "print(f'Number with A&E attendance and respiratory code before admission: {round_to_5(len(ae_resp_before))} ({round((round_to_5(len(ae_resp_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_cov = df[\n",
    "    (df[\"ae_cov\"]==1) &\n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list)) &\n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_cov = df[\n",
    "    ~(\n",
    "        (df[\"ae_cov\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_cov = df[\n",
    "    (df[\"ae_cov\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_cov = df[\n",
    "    ~(\n",
    "        (df[\"ae_cov\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))\n",
    "    ) &\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "tp_ae_resp = df[\n",
    "    (df[\"ae_resp\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_resp = df[\n",
    "    ~(\n",
    "        (df[\"ae_resp\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_resp = df[\n",
    "    (df[\"ae_resp\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_resp = df[\n",
    "    ~(\n",
    "        (df[\"ae_resp\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "diagnosis_code_cov = performance_metrics(tp_ae_cov, fp_ae_cov, tn_ae_cov, fn_ae_cov)\n",
    "diagnosis_code_resp = performance_metrics(tp_ae_resp, fp_ae_resp, tn_ae_resp, fn_ae_resp)\n",
    "\n",
    "print(f'False positives (FDR) - A&E attendance and covid diagnosis code: {round_to_5(len(fp_ae_cov))} ({diagnosis_code_cov[6]}%)')\n",
    "print(f'False positives (FDR) - A&E attendance and respiratory diagnosis code: {round_to_5(len(fp_ae_resp))} ({diagnosis_code_resp[6]}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 3 - Identification in primary care\n",
    "\n",
    "Restrict to those with a probable or suspected COVID-19 code recorded in primary care in the 2 weeks prior to A&E attendance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance and recent suspected covid code in primary care before admission: 500 (62.5%)\n",
      "Number with A&E attendance and recent probable covid code in primary care before admission: 500 (62.5%)\n"
     ]
    },
    {
     "ename": "ZeroDivisionError",
     "evalue": "division by zero",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/3869214867.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     90\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     91\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 92\u001b[0;31m \u001b[0mprim_care_suspected\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_suspected\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_suspected\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_suspected\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_suspected\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     93\u001b[0m \u001b[0mprim_care_probable\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_probable\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_probable\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_probable\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_probable\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     94\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/3350781174.py\u001b[0m in \u001b[0;36mperformance_metrics\u001b[0;34m(tp, fp, tn, fn)\u001b[0m\n\u001b[1;32m     38\u001b[0m     \u001b[0msensitivity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     39\u001b[0m     \u001b[0mspecificity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 40\u001b[0;31m     \u001b[0mfdr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     41\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msensitivity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mspecificity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfdr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     42\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
     ]
    }
   ],
   "source": [
    "ae_prim_care_probable_before = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_prim_care_probable\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "ae_prim_care_suspected_before = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_prim_care_suspected\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance and recent suspected covid code in primary care before admission: {round_to_5(len(ae_prim_care_suspected_before))} ({round((round_to_5(len(ae_prim_care_suspected_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "print(f'Number with A&E attendance and recent probable covid code in primary care before admission: {round_to_5(len(ae_prim_care_probable_before))} ({round((round_to_5(len(ae_prim_care_probable_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_suspected = df[\n",
    "    (df[\"suspected_covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_suspected = df[\n",
    "    ~(\n",
    "        (df[\"suspected_covid_primary_care_before_ae\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_suspected = df[\n",
    "    (df[\"suspected_covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_suspected = df[\n",
    "    ~(\n",
    "        (df[\"suspected_covid_primary_care_before_ae\"]==1)& \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))\n",
    "        \n",
    "    )& \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "tp_ae_probable = df[\n",
    "    (df[\"covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_probable = df[\n",
    "    ~(\n",
    "        (df[\"covid_primary_care_before_ae\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_probable = df[\n",
    "    (df[\"covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_probable = df[\n",
    "    ~(\n",
    "        (df[\"covid_primary_care_before_ae\"]==1) &\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "prim_care_suspected = performance_metrics(tp_ae_suspected, fp_ae_suspected, tn_ae_suspected, fn_ae_suspected)\n",
    "prim_care_probable = performance_metrics(tp_ae_probable, fp_ae_probable, tn_ae_probable, fn_ae_probable)\n",
    "\n",
    "print(f'False positives (FDR) - A&E attendance and probable covid code in primary care: {round_to_5(len(fp_ae_probable))} ({prim_care_probable[6]}%)')\n",
    "print(f'False positives (FDR) - A&E attendance and suspected covid code in primary care: {round_to_5(len(fp_ae_suspected))} ({prim_care_suspected[6]}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 4 - Positive test result\n",
    "\n",
    "Restrict to those with a positive COVID-19 test result recorded up to 4 weeks prior to A&E attendance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance and recent positive test (2 weeks) before admission: 500 (62.5%)\n",
      "Number with A&E attendance and recent positive test (4 weeks) before admissio: 500 (62.5%)\n"
     ]
    },
    {
     "ename": "ZeroDivisionError",
     "evalue": "division by zero",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/4249659114.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     92\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     93\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 94\u001b[0;31m \u001b[0mpositive_test_2_weeks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_test_2_weeks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_test_2_weeks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_test_2_weeks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_test_2_weeks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     95\u001b[0m \u001b[0mpositive_test_4_weeks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     96\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/3350781174.py\u001b[0m in \u001b[0;36mperformance_metrics\u001b[0;34m(tp, fp, tn, fn)\u001b[0m\n\u001b[1;32m     38\u001b[0m     \u001b[0msensitivity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     39\u001b[0m     \u001b[0mspecificity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 40\u001b[0;31m     \u001b[0mfdr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     41\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msensitivity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mspecificity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfdr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     42\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
     ]
    }
   ],
   "source": [
    "ae_pos_test_before_2_weeks = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_positive_covid_test_2_weeks\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "ae_pos_test_before = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_positive_covid_test\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "\n",
    "print(f'Number with A&E attendance and recent positive test (2 weeks) before admission: {round_to_5(len(ae_pos_test_before_2_weeks))} ({round((round_to_5(len(ae_pos_test_before_2_weeks))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "print(f'Number with A&E attendance and recent positive test (4 weeks) before admissio: {round_to_5(len(ae_pos_test_before))} ({round((round_to_5(len(ae_pos_test_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_test = df[\n",
    "    (df[\"pos_test\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_test = df[\n",
    "    ~(\n",
    "        (df[\"pos_test\"]==1) &\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_test = df[\n",
    "    (df[\"pos_test\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_test = df[\n",
    "    ~(\n",
    "        (df[\"pos_test\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tp_ae_test_2_weeks = df[\n",
    "    (df[\"pos_test_2_weeks\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_test_2_weeks = df[\n",
    "    ~(\n",
    "        (df[\"pos_test_2_weeks\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_test_2_weeks = df[\n",
    "    (df[\"pos_test_2_weeks\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_test_2_weeks = df[\n",
    "    ~(\n",
    "        (df[\"pos_test_2_weeks\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "\n",
    "positive_test_2_weeks = performance_metrics(tp_ae_test_2_weeks, fp_ae_test_2_weeks, tn_ae_test_2_weeks, fn_ae_test_2_weeks)\n",
    "positive_test_4_weeks = performance_metrics(tp_ae_test, fp_ae_test, tn_ae_test, fn_ae_test)\n",
    "\n",
    "print(f'False positives (FDR) - A&E attendance and recent positive test (2 weeks): {round_to_5(len(fp_ae_test_2_weeks))} ({positive_test_2_weeks[6]}%)')\n",
    "print(f'False positives (FDR) - A&E attendance and recent positive test (4 weeks): {round_to_5(len(fp_ae_test))} ({positive_test_4_weeks[6]}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "####  Combining approaches"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Diagnosis codes (resp or cov) with recent positive test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with either covid or respiratory diagnosis code and recent positive test (2 weeks) before admission: 375 (46.88%)\n"
     ]
    },
    {
     "ename": "ZeroDivisionError",
     "evalue": "division by zero",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/1160108914.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     70\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     71\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 72\u001b[0;31m \u001b[0mcode_plus_test\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_hos_resp_cov_plus_test_2_weeks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_hos_resp_cov_plus_test_2_weeks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_hos_resp_cov_plus_test_2_weeks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_hos_resp_cov_plus_test_2_weeks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     73\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf'False positives (FDR): {round_to_5(len(fp_ae_hos_resp_cov_plus_test_2_weeks))} ({code_plus_test[6]}%)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/3350781174.py\u001b[0m in \u001b[0;36mperformance_metrics\u001b[0;34m(tp, fp, tn, fn)\u001b[0m\n\u001b[1;32m     38\u001b[0m     \u001b[0msensitivity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     39\u001b[0m     \u001b[0mspecificity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 40\u001b[0;31m     \u001b[0mfdr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     41\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msensitivity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mspecificity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfdr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     42\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
     ]
    }
   ],
   "source": [
    "ae_hos_resp_cov_plus_test_2_weeks = (df.loc[(\n",
    "        (df[\"ae_before_with_hospital_discharge_covid\"]==1) | \n",
    "        (df[\"ae_before_with_hospital_discharge_resp\"]==1) \n",
    "    )&\n",
    "    (df[\"ae_before_with_hospital_discharge_positive_covid_test_2_weeks\"]==1)\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with either covid or respiratory diagnosis code and recent positive test (2 weeks) before admission: {round_to_5(len(ae_hos_resp_cov_plus_test_2_weeks))} ({round((round_to_5(len(ae_hos_resp_cov_plus_test_2_weeks))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)    \n",
    "        )&\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    ~(\n",
    "        (\n",
    "            (\n",
    "                (df[\"ae_cov\"]==1) |\n",
    "                (df[\"ae_resp\"]==1)    \n",
    "            )&\n",
    "            (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "            (df[\"pos_test_2_weeks\"]==1)\n",
    "        ) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)    \n",
    "        )&\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    ~((\n",
    "        (\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)    \n",
    "        )&\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "code_plus_test = performance_metrics(tp_ae_hos_resp_cov_plus_test_2_weeks, fp_ae_hos_resp_cov_plus_test_2_weeks, tn_ae_hos_resp_cov_plus_test_2_weeks, fn_ae_hos_resp_cov_plus_test_2_weeks)\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_ae_hos_resp_cov_plus_test_2_weeks))} ({code_plus_test[6]}%)')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Covid in primary care with recent positive test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with suspected or probable covid in primary care and recent positive test (2 weeks) before admission: 375 (46.88%)\n"
     ]
    },
    {
     "ename": "ZeroDivisionError",
     "evalue": "division by zero",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/1052203179.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     73\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     74\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 75\u001b[0;31m \u001b[0mprimary_care_plus_test\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_hosp_prim_care_before_plus_2_week_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_hosp_prim_care_before_plus_2_week_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_hosp_prim_care_before_plus_2_week_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_hosp_prim_care_before_plus_2_week_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     76\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf'False positives (FDR): {round_to_5(len(fp_ae_hosp_prim_care_before_plus_2_week_test))} ({primary_care_plus_test[6]}%)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/3350781174.py\u001b[0m in \u001b[0;36mperformance_metrics\u001b[0;34m(tp, fp, tn, fn)\u001b[0m\n\u001b[1;32m     38\u001b[0m     \u001b[0msensitivity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     39\u001b[0m     \u001b[0mspecificity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 40\u001b[0;31m     \u001b[0mfdr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     41\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msensitivity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mspecificity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfdr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     42\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
     ]
    }
   ],
   "source": [
    "ae_hosp_prim_care_before_plus_2_week_test = (df.loc[(\n",
    "        (df[\"ae_before_with_hospital_discharge_prim_care_suspected\"]==1) | \n",
    "        (df[\"ae_before_with_hospital_discharge_prim_care_probable\"]==1) \n",
    "    )\n",
    "&\n",
    "    (df[\"ae_before_with_hospital_discharge_positive_covid_test_2_weeks\"]==1)\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with suspected or probable covid in primary care and recent positive test (2 weeks) before admission: {round_to_5(len(ae_hosp_prim_care_before_plus_2_week_test))} ({round((round_to_5(len(ae_hosp_prim_care_before_plus_2_week_test))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "        )&\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    ~(\n",
    "        (\n",
    "            (\n",
    "                (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "                (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "            )&\n",
    "            (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "            (df[\"pos_test_2_weeks\"]==1)\n",
    "        )& \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    ((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "        )&\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )\n",
    "    ) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    ~((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "        )&\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "\n",
    "primary_care_plus_test = performance_metrics(tp_ae_hosp_prim_care_before_plus_2_week_test, fp_ae_hosp_prim_care_before_plus_2_week_test, tn_ae_hosp_prim_care_before_plus_2_week_test, fn_ae_hosp_prim_care_before_plus_2_week_test)\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_ae_hosp_prim_care_before_plus_2_week_test))} ({primary_care_plus_test[6]}%)')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Recent positive test and either covid in primary care or diagnosis code (resp or cov)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with recent positive test and either suspected or probable covid in primary care or a covid or respiratory diagnosis code on attendance: 475 (59.38%)\n"
     ]
    },
    {
     "ename": "ZeroDivisionError",
     "evalue": "division by zero",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/2452013306.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     90\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     91\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 92\u001b[0;31m \u001b[0mpos_test_prim_care_or_diagnosis_code\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_combined_before\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_combined_before\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_combined_before\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_combined_before\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     93\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf'False positives (FDR): {round_to_5(len(fp_combined_before))} ({pos_test_prim_care_or_diagnosis_code[6]}%)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/3350781174.py\u001b[0m in \u001b[0;36mperformance_metrics\u001b[0;34m(tp, fp, tn, fn)\u001b[0m\n\u001b[1;32m     38\u001b[0m     \u001b[0msensitivity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     39\u001b[0m     \u001b[0mspecificity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 40\u001b[0;31m     \u001b[0mfdr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     41\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msensitivity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mspecificity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfdr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     42\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
     ]
    }
   ],
   "source": [
    "combined_before = (df.loc[\n",
    "        (df[\"ae_before_with_hospital_discharge_positive_covid_test_2_weeks\"]==1) &\n",
    "        (\n",
    "            (df[\"ae_before_with_hospital_discharge_prim_care_suspected\"]==1) | \n",
    "            (df[\"ae_before_with_hospital_discharge_prim_care_probable\"]==1) |\n",
    "            (df[\"ae_before_with_hospital_discharge_covid\"]==1) | \n",
    "            (df[\"ae_before_with_hospital_discharge_resp\"]==1)   \n",
    "        )\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with recent positive test and either suspected or probable covid in primary care or a covid or respiratory diagnosis code on attendance: {round_to_5(len(combined_before))} ({round((round_to_5(len(combined_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_combined_before = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "\n",
    "]\n",
    "\n",
    "fn_combined_before = df[\n",
    "    ~(\n",
    "        (\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "        )& \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_combined_before = df[\n",
    "    ((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )\n",
    "    ) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_combined_before = df[\n",
    "    ~((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "pos_test_prim_care_or_diagnosis_code = performance_metrics(tp_combined_before, fp_combined_before, tn_combined_before, fn_combined_before)\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_combined_before))} ({pos_test_prim_care_or_diagnosis_code[6]}%)')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'discharge_destination' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/3823603461.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0midentification_table_index\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m\"tp\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"fp\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"tn\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"fn\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"sensitivity\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"specificity\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"fdr\"\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m identification_table = pd.DataFrame({\n\u001b[0;32m----> 3\u001b[0;31m     \u001b[0;34m\"discharge_destination\"\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mSeries\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdischarge_destination\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0midentification_table_index\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      4\u001b[0m     \u001b[0;34m\"covid_diagnosis_code\"\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mSeries\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdiagnosis_code_cov\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0midentification_table_index\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m     \u001b[0;34m\"resp_diagnosis_code\"\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mSeries\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdiagnosis_code_resp\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0midentification_table_index\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'discharge_destination' is not defined"
     ]
    }
   ],
   "source": [
    "identification_table_index = [\"tp\", \"fp\", \"tn\", \"fn\",\"sensitivity\", \"specificity\", \"fdr\"]\n",
    "identification_table = pd.DataFrame({\n",
    "    \"discharge_destination\": pd.Series(discharge_destination,index=identification_table_index),\n",
    "    \"covid_diagnosis_code\": pd.Series(diagnosis_code_cov,index=identification_table_index),\n",
    "    \"resp_diagnosis_code\": pd.Series(diagnosis_code_resp,index=identification_table_index),\n",
    "    \"primary_care_suspected\": pd.Series(prim_care_suspected,index=identification_table_index),\n",
    "    \"primary_care_probable\": pd.Series(prim_care_probable,index=identification_table_index),\n",
    "    \"positive_test_2_weeks\": pd.Series(positive_test_2_weeks,index=identification_table_index),\n",
    "    \"positive_test_4_weeks\": pd.Series(positive_test_4_weeks,index=identification_table_index),\n",
    "    \"diagnosis_code_with_positive_test\": pd.Series(code_plus_test,index=identification_table_index),\n",
    "    \"primary_care_with_positive_test\": pd.Series(primary_care_plus_test,index=identification_table_index),\n",
    "    \"pos_test_prim_care_or_diagnosis\": pd.Series(pos_test_prim_care_or_diagnosis_code,index=identification_table_index)\n",
    "    \n",
    "})\n",
    "identification_table.to_csv(\"output/jun/identification_table_hosp_discharge.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "produce_matrix(df_list = [tp_ae_hosp_discharge, tp_ae_cov, tp_ae_resp, tp_ae_suspected, tp_ae_probable, tp_ae_test, tp_ae_test_2_weeks, tp_ae_hos_resp_cov_plus_test_2_weeks, tp_ae_hosp_prim_care_before_plus_2_week_test], names=[\"Hospital discharge\", \"Covid diagnosis code\", \"Respiratory diagnosis code\", \"Probable covid code\", \"Suspected covid code\", \"Recent positive test (2 weeks)\", \"Recent positive test (4 weeks)\", \"Diagnosis codes (resp or cov) with recent positive test\", \"Covid in primary care with recent positive test\", \"Recent positive test and either covid in primary care or diagnosis code (resp or cov)\"], filename=\"cm_hosp_discharge\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Number of emergency primary covid hospital admissions and those that can be identified in A&E attendances\n",
    "### Any A&E attendance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of emergency primary covid hospital admissions: 300\n",
      "Number with A&E attendance before admission: 500 (166.67%)\n",
      "---\n",
      "Number of A&E attendances: 400\n"
     ]
    }
   ],
   "source": [
    "df = pd.read_csv('output/input_prim_cov_2021-06-01.csv.gz', parse_dates=[\"ae_attendance_date\", \"ae_attendance_hosp_discharge_date\", \"emergency_covid_hospital_admission_date\", \"ae_before_date\", \"ae_before_with_hospital_discharge_date\"])\n",
    "\n",
    "\n",
    "\n",
    "num_emergency_cov_admissions = df['emergency_covid_hospital_admission_date'].count()\n",
    "ae_before = (df.loc[\n",
    "    df[\"ae_before_date\"].notnull()\n",
    ", \"patient_id\"])\n",
    "print(f'Number of emergency primary covid hospital admissions: {num_emergency_cov_admissions}')\n",
    "print(f'Number with A&E attendance before admission: {round_to_5(len(ae_before))} ({round((round_to_5(len(ae_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "ae = round_to_5(len(df[\n",
    "    (df[\"ae_attendance_date\"].notnull())\n",
    "]))\n",
    "\n",
    "ae_with_hosp_discharge = round_to_5(len(df[\n",
    "    (df[\"ae_attendance_hosp_discharge_date\"].notnull())\n",
    "]))\n",
    "\n",
    "summary_table = pd.DataFrame({\n",
    "    \"emergency covid hospital admission\": num_emergency_cov_admissions,\n",
    "    \"ae attendance before admission\": round_to_5(len(ae_before)),\n",
    "    \"ae_attendance\": ae, \n",
    "    \"ae attendance with hospital discharge\": ae_with_hosp_discharge,\n",
    "    \n",
    "}, index=[\"count\"])\n",
    "\n",
    "summary_table.to_csv(\"output/jun/summary_table_primary.csv\")\n",
    "\n",
    "print('---')\n",
    "print(f'Number of A&E attendances: {ae}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of emergency covid hospital admissions: 300\n",
      "Number with A&E attendance before admission: 500 (166.67%)\n",
      "---\n",
      "Number of A&E attendances: 400\n"
     ]
    }
   ],
   "source": [
    "num_emergency_cov_admissions = round_to_5(df['emergency_covid_hospital_admission_date'].count())\n",
    "ae_before = (df.loc[\n",
    "    df[\"ae_before_date\"].notnull()\n",
    ", \"patient_id\"])\n",
    "print(f'Number of emergency covid hospital admissions: {num_emergency_cov_admissions}')\n",
    "print(f'Number with A&E attendance before admission: {round_to_5(len(ae_before))} ({round((round_to_5(len(ae_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "ae = round_to_5(len(df[\n",
    "    (df[\"ae_attendance_date\"].notnull())\n",
    "]))\n",
    "\n",
    "ae_with_hosp_discharge = round_to_5(len(df[\n",
    "    (df[\"ae_attendance_hosp_discharge_date\"].notnull())\n",
    "]))\n",
    "\n",
    "summary_table = pd.DataFrame({\n",
    "    \"emergency covid hospital admission\": num_emergency_cov_admissions,\n",
    "    \"ae attendance before admission\": round_to_5(len(ae_before)),\n",
    "    \"ae_attendance\": ae, \n",
    "    \"ae attendance with hospital discharge\": ae_with_hosp_discharge,\n",
    "    \n",
    "}, index=[\"count\"])\n",
    "\n",
    "summary_table.to_csv(\"output/jun/summary_table.csv\")\n",
    "\n",
    "print('---')\n",
    "print(f'Number of A&E attendances: {ae}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 1 - Discharge destination\n",
    "Restrict to those that attended A&E and had a hospital discharge destination (hospital ward, emergency short stay ward, icu)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with hospital discharge before admission: 500 (166.67%)\n"
     ]
    },
    {
     "ename": "ZeroDivisionError",
     "evalue": "division by zero",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/2108371917.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     39\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     40\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 41\u001b[0;31m \u001b[0mdischarge_destination\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_hosp_discharge\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_hosp_discharge\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_hosp_discharge\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_hosp_discharge\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     42\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     43\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf'False positives (FDR): {round_to_5(len(fp_ae_hosp_discharge))} ({discharge_destination[6]}%)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/3350781174.py\u001b[0m in \u001b[0;36mperformance_metrics\u001b[0;34m(tp, fp, tn, fn)\u001b[0m\n\u001b[1;32m     38\u001b[0m     \u001b[0msensitivity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     39\u001b[0m     \u001b[0mspecificity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 40\u001b[0;31m     \u001b[0mfdr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     41\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msensitivity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mspecificity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfdr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     42\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
     ]
    }
   ],
   "source": [
    "ae_hosp_before = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_date\"].notnull()\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with hospital discharge before admission: {round_to_5(len(ae_hosp_before))} ({round((round_to_5(len(ae_hosp_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_hosp_discharge = df[\n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_hosp_discharge = df[\n",
    "    ~(\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_hosp_discharge = df[\n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_hosp_discharge = df[\n",
    "    ~(df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "discharge_destination = performance_metrics(tp_ae_hosp_discharge, fp_ae_hosp_discharge, tn_ae_hosp_discharge, fn_ae_hosp_discharge)\n",
    "\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_ae_hosp_discharge))} ({discharge_destination[6]}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 2 - Diagnosis code\n",
    "Restrict to those with a COVID-19 diagnosis code or respiratory-related diagnosis code from the [emergency care diagnosis refset](https://termbrowser.nhs.uk/?perspective=full&conceptId1=991411000000109&edition=uk-edition&release=v20220216&server=https://termbrowser.nhs.uk/sct-browser-api/snomed&langRefset=900000000000508004)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance and covid diagnosis code before admission: 500 (166.67%)\n",
      "Number with A&E attendance and respiratory code before admission: 500 (166.67%)\n",
      "False positives (FDR) - A&E attendance and covid diagnosis code: 345 (98.57%)\n",
      "False positives (FDR) - A&E attendance and respiratory diagnosis code: 350 (97.22%)\n"
     ]
    }
   ],
   "source": [
    "ae_cov_before = (df.loc[\n",
    "    df[\"ae_before_with_covid\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "ae_resp_before = (df.loc[\n",
    "    df[\"ae_before_with_resp\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance and covid diagnosis code before admission: {round_to_5(len(ae_cov_before))} ({round((round_to_5(len(ae_cov_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "print(f'Number with A&E attendance and respiratory code before admission: {round_to_5(len(ae_resp_before))} ({round((round_to_5(len(ae_resp_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_cov = df[\n",
    "    (df[\"ae_cov\"]==1) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_cov = df[\n",
    "    ~(\n",
    "        (df[\"ae_cov\"]==1) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_cov = df[\n",
    "    (df[\"ae_cov\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_cov = df[\n",
    "    ~(df[\"ae_cov\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "tp_ae_resp = df[\n",
    "    (df[\"ae_resp\"]==1) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_resp = df[\n",
    "    ~(\n",
    "        (df[\"ae_resp\"]==1) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_resp = df[\n",
    "    (df[\"ae_resp\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_resp = df[\n",
    "    ~(df[\"ae_resp\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "diagnosis_code_cov = performance_metrics(tp_ae_cov, fp_ae_cov, tn_ae_cov, fn_ae_cov)\n",
    "diagnosis_code_resp = performance_metrics(tp_ae_resp, fp_ae_resp, tn_ae_resp, fn_ae_resp)\n",
    "\n",
    "print(f'False positives (FDR) - A&E attendance and covid diagnosis code: {round_to_5(len(fp_ae_cov))} ({diagnosis_code_cov[6]}%)')\n",
    "print(f'False positives (FDR) - A&E attendance and respiratory diagnosis code: {round_to_5(len(fp_ae_resp))} ({diagnosis_code_resp[6]}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 3 - Identification in primary care\n",
    "\n",
    "Restrict to those with a probable or suspected COVID-19 code recorded in primary care in the 2 weeks prior to A&E attendance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance and recent suspected covid code in primary care before admission: 500 (166.67%)\n",
      "Number with A&E attendance and recent probable covid code in primary care before admission: 500 (166.67%)\n",
      "False positives (FDR) - A&E attendance and probable covid code in primary care: 97.22%)\n",
      "False positives (FDR) - A&E attendance and suspected covid code in primary care: 97.1%)\n"
     ]
    }
   ],
   "source": [
    "ae_prim_care_probable_before = (df.loc[\n",
    "    df[\"ae_before_with_prim_care_probable\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "ae_prim_care_suspected_before = (df.loc[\n",
    "    df[\"ae_before_with_prim_care_suspected\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance and recent suspected covid code in primary care before admission: {round_to_5(len(ae_prim_care_suspected_before))} ({round((round_to_5(len(ae_prim_care_suspected_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "print(f'Number with A&E attendance and recent probable covid code in primary care before admission: {round_to_5(len(ae_prim_care_probable_before))} ({round((round_to_5(len(ae_prim_care_probable_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_suspected = df[\n",
    "    (df[\"suspected_covid_primary_care_before_ae\"]==1) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_suspected = df[\n",
    "    ~(\n",
    "        (df[\"suspected_covid_primary_care_before_ae\"]==1) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_suspected = df[\n",
    "    (df[\"suspected_covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_suspected = df[\n",
    "    ~(df[\"suspected_covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "tp_ae_probable = df[\n",
    "    (df[\"covid_primary_care_before_ae\"]==1) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_probable = df[\n",
    "    ~(\n",
    "        (df[\"covid_primary_care_before_ae\"]==1) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_probable = df[\n",
    "    (df[\"covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_probable = df[\n",
    "    ~(df[\"covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "prim_care_suspected = performance_metrics(tp_ae_suspected, fp_ae_suspected, tn_ae_suspected, fn_ae_suspected)\n",
    "prim_care_probable = performance_metrics(tp_ae_probable, fp_ae_probable, tn_ae_probable, fn_ae_probable)\n",
    "print(f'False positives (FDR) - A&E attendance and probable covid code in primary care: {prim_care_probable[6]}%)')\n",
    "print(f'False positives (FDR) - A&E attendance and suspected covid code in primary care: {prim_care_suspected[6]}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 4 - Positive test result\n",
    "\n",
    "Restrict to those with a positive COVID-19 test result recorded up to 4 weeks prior to A&E attendance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance and recent positive test (2 weeks) before admission: 500 (166.67%)\n",
      "Number with A&E attendance and recent positive test (4 weeks) before admissio: 500 (166.67%)\n",
      "False positives (FDR) - A&E attendance and recent positive test (2 weeks): 350 (98.59%)\n",
      "False positives (FDR) - A&E attendance and recent positive test (4 weeks): 345 (97.18%)\n"
     ]
    }
   ],
   "source": [
    "ae_pos_test_before_2_weeks = (df.loc[\n",
    "    df[\"ae_before_with_positive_covid_test_2_weeks\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "ae_pos_test_before = (df.loc[\n",
    "    df[\"ae_before_with_positive_covid_test\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "\n",
    "print(f'Number with A&E attendance and recent positive test (2 weeks) before admission: {round_to_5(len(ae_pos_test_before_2_weeks))} ({round((round_to_5(len(ae_pos_test_before_2_weeks))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "print(f'Number with A&E attendance and recent positive test (4 weeks) before admissio: {round_to_5(len(ae_pos_test_before))} ({round((round_to_5(len(ae_pos_test_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_test = df[\n",
    "    (df[\"pos_test\"]==1) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_test = df[\n",
    "    ~(\n",
    "        (df[\"pos_test\"]==1) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_test = df[\n",
    "    (df[\"pos_test\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_test = df[\n",
    "    ~(df[\"pos_test\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tp_ae_test_2_weeks = df[\n",
    "    (df[\"pos_test_2_weeks\"]==1) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_test_2_weeks = df[\n",
    "    ~(\n",
    "        (df[\"pos_test_2_weeks\"]==1) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_test_2_weeks = df[\n",
    "    (df[\"pos_test_2_weeks\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_test_2_weeks = df[\n",
    "    ~(df[\"pos_test_2_weeks\"]==1) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "positive_test_2_weeks = performance_metrics(tp_ae_test_2_weeks, fp_ae_test_2_weeks, tn_ae_test_2_weeks, fn_ae_test_2_weeks)\n",
    "positive_test_4_weeks = performance_metrics(tp_ae_test, fp_ae_test, tn_ae_test, fn_ae_test)\n",
    "\n",
    "print(f'False positives (FDR) - A&E attendance and recent positive test (2 weeks): {round_to_5(len(fp_ae_test_2_weeks))} ({positive_test_2_weeks[6]}%)')\n",
    "print(f'False positives (FDR) - A&E attendance and recent positive test (4 weeks): {round_to_5(len(fp_ae_test))} ({positive_test_4_weeks[6]}%)')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "####  Combining approaches"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Diagnosis codes (resp or cov) with recent positive test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with either covid or respiratory diagnosis code and recent positive test (2 weeks) before admission: 385 (128.33%)\n",
      "False positives (FDR): 255 (98.08%)\n"
     ]
    }
   ],
   "source": [
    "ae_hos_resp_cov_plus_test_2_weeks = (df.loc[(\n",
    "        (df[\"ae_before_with_covid\"]==1) | \n",
    "        (df[\"ae_before_with_resp\"]==1) \n",
    "    )&\n",
    "    (df[\"ae_before_with_positive_covid_test_2_weeks\"]==1)\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with either covid or respiratory diagnosis code and recent positive test (2 weeks) before admission: {round_to_5(len(ae_hos_resp_cov_plus_test_2_weeks))} ({round((round_to_5(len(ae_hos_resp_cov_plus_test_2_weeks))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)    \n",
    "        )&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    ~(\n",
    "        (\n",
    "            (\n",
    "                (df[\"ae_cov\"]==1) |\n",
    "                (df[\"ae_resp\"]==1)    \n",
    "            )&\n",
    "            (df[\"pos_test_2_weeks\"]==1)\n",
    "        ) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)    \n",
    "        )&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    ~((\n",
    "        (\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)    \n",
    "        )&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "code_plus_test = performance_metrics(tp_ae_hos_resp_cov_plus_test_2_weeks, fp_ae_hos_resp_cov_plus_test_2_weeks, tn_ae_hos_resp_cov_plus_test_2_weeks, fn_ae_hos_resp_cov_plus_test_2_weeks)\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_ae_hos_resp_cov_plus_test_2_weeks))} ({code_plus_test[6]}%)')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Covid in primary care with recent positive test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with suspected or probable covid in primary care and recent positive test (2 weeks) before admission: 375 (125.0%)\n",
      "False positives (FDR): 255 (98.08%)\n"
     ]
    }
   ],
   "source": [
    "ae_hosp_prim_care_before_plus_2_week_test = (df.loc[\n",
    "    \n",
    "    \n",
    "    (\n",
    "        (df[\"ae_before_with_prim_care_suspected\"]==1) | \n",
    "        (df[\"ae_before_with_prim_care_probable\"]==1) \n",
    "    )\n",
    "&\n",
    "    (df[\"ae_before_with_positive_covid_test_2_weeks\"]==1)\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with suspected or probable covid in primary care and recent positive test (2 weeks) before admission: {round_to_5(len(ae_hosp_prim_care_before_plus_2_week_test))} ({round((round_to_5(len(ae_hosp_prim_care_before_plus_2_week_test))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "        )&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    ~(\n",
    "        (\n",
    "            (\n",
    "                (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "                (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "            )&\n",
    "            (df[\"pos_test_2_weeks\"]==1)\n",
    "        )& \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    ((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "        )&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )\n",
    "    ) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    ~((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "        )&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "\n",
    "primary_care_plus_test = performance_metrics(tp_ae_hosp_prim_care_before_plus_2_week_test, fp_ae_hosp_prim_care_before_plus_2_week_test, tn_ae_hosp_prim_care_before_plus_2_week_test, fn_ae_hosp_prim_care_before_plus_2_week_test)\n",
    "\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_ae_hosp_prim_care_before_plus_2_week_test))} ({primary_care_plus_test[6]}%)')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Recent positive test and either covid in primary care or diagnosis code (resp or cov)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with recent positive test and either suspected or probable covid in primary care or a covid or respiratory diagnosis code on attendance: 470 (156.67%)\n",
      "False positives (FDR): 325 (98.48%)\n"
     ]
    }
   ],
   "source": [
    "combined_before = (df.loc[\n",
    "        (df[\"ae_before_with_positive_covid_test_2_weeks\"]==1) &\n",
    "        (\n",
    "            (df[\"ae_before_with_prim_care_suspected\"]==1) | \n",
    "            (df[\"ae_before_with_prim_care_probable\"]==1) |\n",
    "            (df[\"ae_before_with_covid\"]==1) | \n",
    "            (df[\"ae_before_with_resp\"]==1)   \n",
    "        )\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with recent positive test and either suspected or probable covid in primary care or a covid or respiratory diagnosis code on attendance: {round_to_5(len(combined_before))} ({round((round_to_5(len(combined_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_combined_before = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_combined_before = df[\n",
    "    ~(\n",
    "        (\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "        )& \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_combined_before = df[\n",
    "    ((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )\n",
    "    ) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_combined_before = df[\n",
    "    ~((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "\n",
    "pos_test_prim_care_or_diagnosis_code = performance_metrics(tp_combined_before, fp_combined_before, tn_combined_before, fn_combined_before)\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_combined_before))} ({pos_test_prim_care_or_diagnosis_code[6]}%)')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'discharge_destination' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/2749801917.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0midentification_table_index\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m\"tp\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"fp\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"tn\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"fn\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"sensitivity\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"specificity\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"fdr\"\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m identification_table = pd.DataFrame({\n\u001b[0;32m----> 3\u001b[0;31m     \u001b[0;34m\"discharge_destination\"\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mSeries\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdischarge_destination\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0midentification_table_index\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      4\u001b[0m     \u001b[0;34m\"covid_diagnosis_code\"\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mSeries\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdiagnosis_code_cov\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0midentification_table_index\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m     \u001b[0;34m\"resp_diagnosis_code\"\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mSeries\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdiagnosis_code_resp\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0midentification_table_index\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'discharge_destination' is not defined"
     ]
    }
   ],
   "source": [
    "identification_table_index = [\"tp\", \"fp\", \"tn\", \"fn\",\"sensitivity\", \"specificity\", \"fdr\"]\n",
    "identification_table = pd.DataFrame({\n",
    "    \"discharge_destination\": pd.Series(discharge_destination,index=identification_table_index),\n",
    "    \"covid_diagnosis_code\": pd.Series(diagnosis_code_cov,index=identification_table_index),\n",
    "    \"resp_diagnosis_code\": pd.Series(diagnosis_code_resp,index=identification_table_index),\n",
    "    \"primary_care_suspected\": pd.Series(prim_care_suspected,index=identification_table_index),\n",
    "    \"primary_care_probable\": pd.Series(prim_care_probable,index=identification_table_index),\n",
    "    \"positive_test_2_weeks\": pd.Series(positive_test_2_weeks,index=identification_table_index),\n",
    "    \"positive_test_4_weeks\": pd.Series(positive_test_4_weeks,index=identification_table_index),\n",
    "    \"diagnosis_code_with_positive_test\": pd.Series(code_plus_test,index=identification_table_index),\n",
    "    \"primary_care_with_positive_test\": pd.Series(primary_care_plus_test,index=identification_table_index),\n",
    "    \"pos_test_prim_care_or_diagnosis\": pd.Series(pos_test_prim_care_or_diagnosis_code,index=identification_table_index)\n",
    "    \n",
    "})\n",
    "identification_table.to_csv(\"output/jun/identification_table_no_hosp_discharge_primary.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAw0AAAMICAYAAACKLAlTAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8rg+JYAAAACXBIWXMAAAsTAAALEwEAmpwYAACp6klEQVR4nOzdd5xcVf3G8c8m9BBABEQIWCgPoIj0IiUgKGAhdkR+SDOgoIIVKQIqxQIKIiUIIiIiFlAREKV3jKjUPPSqCEgLLSHZ/f1xzsK47C6TsDuz2X3er9e+dubeO+d8753Z2XPuaR1dXV1ERERERET0ZVS7A4iIiIiIiKEtlYaIiIiIiOhXKg0REREREdGvVBoiIiIiIqJfqTRERERERES/UmmIiIiIiIh+zdXuACIiIiIi4iWSRgHHAqsB04Bdbd/R45jFgSuBt9l+XtL8wGnAEsBU4JO2H5H0PuDrwAzgZNsnzk5MaWmIiIiIiBhaJgDz2V4f2Ac4onGnpHcDFwBLNmz+NHCj7Y2AU4H9Jc0NfB94F7AJMFHS62YnoFQaIiIiIiKGlg2B8wFsXwOs1WN/J7A58FhvrwHOq/tXBu6w/bjt6cAVwMazE1C6J0VEjHCTx03oamf+az1wdjuzj4jhrWMgE3vh0bsG7Pvyre/YajdgYsOmSbYn1ccLAU827JspaS7bMwBs/xlAUmOSja+ZCizcSzrd22dZKg0REREREc3onDlgSdUKwqQ+dj8FjG14Pqq7wtCPxteMBZ7oJZ3u7bMs3ZMiIiIiIoaWK4GtASStB9w4K68BtgIuB24FVpC0qKR5KF2Trp6dgNLSEBERERHRjK7OVuV0FrCFpKsoXax2kvQFyviE3/fxmuOAn0q6ApgObGf7hfq6P1EaC062/eDsBNTR1dXWrqwREdFmGdMQEcPYwI5p+PetA/Z9OffrVx7Q2AZbuidFRERERES/0j0pIiIiIqIJXa3rnjTkpNIQEREREdGMzpFbaUj3pIiIiIiI6FdaGiIiIiIimpHuSRERERER0a8BXNxtTpPuSRERERER0a+0NERERERENCPdkyIiIiIiol+ZPSkiIiIiIqJ3aWmIiIiIiGhCFneLiIiYVR0dLHvobiywyhvpmj6De758DNPueajdUUVEDJ50T4qIGHokjZd0Ro9th0vacYDS/239vaqkjWcljl6Oeaj+3kfSOrMQwyumPVQtsuW6jJp3HqZssw8PHHYq4w7Yqd0hRUTEIElLQ0SMWLY/WB9+CHgIuGwA0jz81aYxp1hw7ZV58pLrAXjm+tsYs9rybY4oImKQpXtSRMScR9IRwIb16em2j5L0QeCrwAvAv4Btga8DKwFLAK8BPmv7ito6sCawIzBd0vXAssAewNxAF/CBPvIeDUwC3gLcCcxbt58CnAHcBfwEmEFp1d0OeAD4IbAOMA9wIPAksIKk82p8f7B9kKRN6v5RwIL19dOBPwD/Bc4FLgF+BEwFHgaet72jpM/W47uAM2wfPcsXtwmjxy7AzKnPvvi8a2YnjB4FM0fuP9WIGOayuFtExJC1maRLun8ohWEkvRd4E7AepeKwnaRVgY8D37W9IXAOsFBN51nbmwHbUwraANh+EDgFONL2dcCKwHvq628B3t1HXB8A5rO9HvA1YIEe+7cArgM2pxT+FwYmAIvZXgfYFFirHjtf3bcRsGfd9hZge9vjgd8CH6nblwTeZfs7wPHAjvW87qzXZRXgY/WabARMkKQ+zuFVmTn1WUaPmf/F5x2jOlJhiIgYplJpiIih7iLb47t/gNPr9pWBy2132X4BuAZYBfgCpaJxKbAB0F2KvQjA9s2UgndfHgZ+KuknwNsoLQ69WZFSKcD2fcD9PfafBDwBnE+pCMwABFxdX/O47QPqsTfZnmb72XocwIPA0bXlYtOGOO62Pb0+XqqeD8Dl9fdbgTcAF9af1wIr9HO+s+3pyVNYeLM1ARizxoo8N+XewcgmImLo6OocuJ85TCoNETGnupXaNUnS3JQKwu3AROAg25sAHbzUvWjNeuxbKQXyRp3AKEkLAwdTujTtCjxX0+jNLcD6Nc2lgKV77N+GUql5J/ArSpepW4G162sWlvSnemxXL+mfCOxke0dKN6vuOBr/09xfWxagtLgAGLgZ2LRWsk4BbujjHF6VJ867hs5p01np7MNZ5sCduf+gkwcjm4iIoaOzc+B+5jAZ0xARcyTb59SZh66mjA840/b1kpYGzpE0FXia0kXps8Dqki4ExgCf6pHc34DvUgr1V1JaA2YAjwNLAXf3EsLvgC0kXQvcCzzaY/9kSovF/sBoYG/g78Dmkq6gfP8e3M8pngZcLukZ4D81jp4+A5ws6WnKeIcHbf+znucVkualtIb0rCQNjK4u7vva8YOSdEREDC0dXV293eCKiBg+JB0EPGR7WJVwJe1BqSw9IulbwHTb35jVdCaPm9DWfwRrPXB2O7OPiOGtr9bi2TLtpj8P2PflvG/dYkBjG2xpaYiImHP9B7igtjQ8CXyyzfFERAxvc2C3ooGSSkNEDHu2D2p3DIPB9q+BX7c7joiIGP5SaYiIiIiIaEJX18hdpyGVhoiIiIiIZsyBU6UOlEy5GhERERER/UpLQ0REREREMzIQOiIiIiIi+pXuSREREREREb1LS0NERERERDM6M3tSRERERET0J92TIiIiIiIiepeWhoiIiIiIZmT2pIiIiIiI6Fe6J0VERERERPQuLQ0REREREc1I96SIiIj2mDxuQrtDYK0Hzm53CBExJxjBlYZ0T4qIiIiIiH6lpSEiIiIiogldXVncLSIiIiIi+jOCuyel0hARERER0YwRPOVqKg0REREREUOIpFHAscBqwDRgV9t3NOz/FLAbMAP4lu1zJP0AeHs9ZEngCdvrSToK2BCYWvdtY/vJWY0plYaIiIiIiGa0rnvSBGA+2+tLWg84AtgGQNKSwOeAtYD5gCsk/dn2XnX/3MAVwKdqWmsC77b96KsJKJWGiIiIiIhmDGD3JEkTgYkNmybZnlQfbwicD2D7GklrNRy3DnCl7WnANEl3AG8D/lr3fxa4wPaNtcViBWCSpNcBJ9k+eXbiTaUhIiIiIqLFagVhUh+7FwIauxDNlDSX7Rm97JsKLAwgaR5Kt6V16r4xwA+BI4HRwMWSJtu+YVbjTaUhIiIiIqIZreue9BQwtuH5qFph6G3fWOCJ+nhz4LKGMQvPAkfZfhZA0kWUcRKpNEREREREDIrWzZ50JfA+4Mw6puHGhn3XAYdImg+YF1gZuKnu2xw4r+HYFYFfSlqdsqjzhsBPZyegrAgdERERETG0nAU8L+kq4PvA3pK+IOn9th8CjgYuBy4C9rP9fH2dgLu6E7F9K/Az4BrgUuBU2zfPTkAdXV1ds302EREx55s8bsKI/0ew1gNntzuEiBgcHQOZ2HPnHT1g35fzb/W5AY1tsKV7UkREREREM0bwitDpnhQREREREf1KS0NERERERDNaNxB6yEmlISIiZk9HB8seuhsLrPJGuqbP4J4vH8O0ex4aeTFExMgxgrsnpdIQES0j6S3Ad4AFgAWBc4GDbDc9sEzSPsBFtq9r2DYfMMX2G/t53RnA8cB8wLINq262hKQdgcds/36Q0t8dWNL2QYORfm8W2XJdRs07D1O22Ycxa6zIuAN24s5dDmtV9kMmhoiIkSCVhohoCUmLAGcAH7R9u6TRwK8oK1ce32w6tg9/NXHYPv/VvP5V5HtKO/IdTAuuvTJPXnI9AM9cfxtjVlt+RMYQESNIuidFRAy6bSgtBLcD2J4paQdgOoCkIyiLzgCcDhwL3AqsZvsZSV8CZlJWsjwDuAL4OfAa4I7eMpS0B7Ar8G9gibptR2Al2/tIOgxYC3gt8E/bO0larOY/L2BgM9vLS7qBMsf124AuYBvbT/aM2/ZRkj4IfBV4AfgXsC3wdeAh4DfALykTUcwH7G77Hw0xzw/8BHgDMA+wJzC5bnszMBo40vYvJW0IHAU8DsygzMONpM8C29U4z7B9dP9vzewZPXYBZk599sXnXTM7YfQomNm6f6pDIYaIGEFGcPekzJ4UEa2yFA0LzgDYftr2dEnvBd4ErEcpgG8HrEQpYH+oHr4dcGrDy3cHbrK9MXBCz8wkvQ74fE1zG0oBvHH/QsDjtregVBzWk7Q0sB9wtu1NKC0h3TdXFgJ+Ubc/CGzVW9ySVgU+DnzX9obAOfW13dYB/gtsBewBjOkR+u7APbbXp1Q21qW0xjxiewPKap/fqpWb44CP294cuLue1yrAx2o8GwETJKnn9RkIM6c+y+gx87/4vGNUR8sL60MhhoiIkSCVhoholXuBZRo3SHqTpI2BlYHLbXfZfoFyx3wV4MfADpLWAWz7vw0vXxG4jrLjWspd/UbLATfbnlbTvK7H/ueAJST9glLpWBCYu8ZyVT3m8h6v+Xv9fT+llaCvuL8AbCbpUmADoLEUex5wJfA74Bs99kFZzfPqel632/5Bzeeyum0qcEs9v9fZvq2+7sr6+62UVooL689rgRUYBE9PnsLCm60JwJg1VuS5KfcORjZDPoaIGEG6OgfuZw6TSkNEtMo5wJaSlgOQNDdwJKWQeyu1i0/dvgFwe+3K1AF8GTixR3q3AOvX16xOKfA3uh14i6T56/iJ1Xvs3wpYxvbHgX2B+WteN3WnS2lBaNRzwHavcQMTKQO8N6lpfqDhNeOBf9t+F/At4NBe0ly7pvlmSafXbRvVbWOBVSktCw9KWrm+bu3628DNwKa2xwOnADcwCJ447xo6p01npbMPZ5kDd+b+g04ejGyGfAwRMYJ0dg7czxwmYxoioiVsPyXpk8CJkkYBY4E/AMfZ7pI0XtLVlG5EZ9q+vr70JMod+Yt7JHk8cKqkK4ApwLQe+T0i6XBKq8EjwDM9Xn8dcICkyyiVgbsoXagOB34m6aOU8Qg9WzAa8zint7hrN6dzJE0FnqZUmD5bX/ZP4AxJn6Z8B3+jR7InACfXVorRwF6UQv+J9VznBw62/bCk3eo1eAqYSulu9U9JFwJXSJq3nueDfZ3Dq9LVxX1fa3oM++AYCjFERIwAHV1dTc90GBEx7EnamjJ+4K+SNgf2tb1Zu+MaTJPHTRjx/wjWeuDsdocQEYOjYyATe+7MbwzY9+X8H/36gMY22NLSEBHxv+6m3OmfQbnT/7k2xxMREUPFCL7ZnkpDREQD27fy0piGiIiIIJWGiIiIiIjmzIEDmAdKKg0REREREc0YwZWGTLkaERERERH9SktDREREREQz5sBF2QZKKg0REREREc1I96SIiIiIiIjepaUhIiIiIqIZWachIiIiIiL6le5JERERERERvUtLQ0REREREM0ZwS0MqDRERERERzRjBU66me1JERERERPQrLQ0REREREU3o6szsSRERESPW5HET2h0Caz1wdrtDiIhXMoLHNKR7UkRERERE9CstDRERERERzRjBA6FTaYiIiIiIaMYIHtOQ7kkREREREdGvtDRERERERDRjBA+ETqUhIiIiIqIZI7jSkO5JERERERHRr7Q0REREREQ0o2vkDoROpSEiIiIiohnpnhQREREREdG7tDRERERERDRjBK/TkEpDREREREQzWrQitKRRwLHAasA0YFfbdzTs/xSwGzAD+JbtcyQtCtwG3FQPO8v2Ub0dOzsxpdIQERERETG0TADms72+pPWAI4BtACQtCXwOWAuYD7hC0p+BNYBf2P5sdyJ9HWt72qwGlEpDREREREQzWtc9aUPgfADb10haq2HfOsCVteA/TdIdwNuANYE1JV0KPEypLKzdx7F/ndWAUmmIiIiIiGhC1wDOniRpIjCxYdMk25Pq44WAJxv2zZQ0l+0ZveybCiwMTAH+Zvsvkj4B/BA4u49jZ1kqDRERMXs6Olj20N1YYJU30jV9Bvd8+Rim3fPQyIqh3flHxByrVhAm9bH7KWBsw/NRtcLQ276xwBPAtcCzddtZwDeAU/s4dpZlytWIEU7SeEkPS7pE0sWS/ibpV5LmGYC0l5R07Cwcv7Gkt73afPtI+6H6ex9J6wxGHq+Q/w8kLTuI6Z8hafxgpd+bRbZcl1HzzsOUbfbhgcNOZdwBO7Uy+yERQ7vzj4gW6+wauJ/+XQlsDVDHNNzYsO86YCNJ80laGFiZMvj5x8CH6jHvBP7Wz7GzLC0NEQFwke1tu59IOh14P/DrV5Oo7YeAz8zCS3YGzgBueDX59sf24YOV9ivku1c78h1MC669Mk9ecj0Az1x/G2NWW37ExdDu/COixVo0exKlpWALSVcBHcBOkr4A3GH795KOBi6nNADsZ/t5SfsAJ0v6DPAMZcalh3o7dnYCSqUhIv5HbWF4PfB4fX4YsBEwGjjS9q/qF9IngU7gr7Y/J+kUyhfbMsCCwA7A88AZtteTdBNlKrjpwJeA4ygzObwe2B+4H9gSWEPSLTXPvShTzd1O6ff5CUrFYhTwTcoX4kdqnFcCH7H9r/p8NKXZ9y3AncC8dfsplIrJVZS7MosASwE/sn1cbYX4EaXf58P1HA4CflFjXA64zvanJS0CnEbpXzoXsL/tiyQdAmxat/3G9rclXQLsDryWMgvGC5Rm5A/bntpw/Veocc1T928LjAFOrul1AZ+z/U9JewC7Av8Glqivnxs4HlihXqf9bV/Sx9v9qoweuwAzpz774vOumZ0wehTMbN2Kqe2Ood35R0SLtWggtO1Oyv+MRlMa9p8InNjjNXdT/vf0TOtlx86OdE+KCIDNavekW4DrKXM7XyhpK+BNtjekfBHtVwvKOwF72l4fuFVS9w2IO21vRilkf6dHHgsC36wtGisBR9jeglIZ2MP23ygzRXyFcofkYGCzmvcTlDmmAR6v2y4AVpX0GklvAR7trjBUH6BMV7ce8DVggR7xLE+p0LwLeBfwhbr9eGDHeh53Nhy/IrALZdaKres0dvsDf7a9MfAR4CRJHZTKzXaUis8TPfKdAJwJbEKpOL2mx/7vAYfVa3sUsHrddlTN5/M1n9fVx+tRpuHr7k62a70WG9ftP2KQzJz6LKPHzP/i845RHS0vLLc7hnbnHxHRKqk0RASU7knjKYXc6cDddfuqlOnbLqEU6OcG3kipNOxRp3V7A6WFAeCi+vsqQL3k4/r738Bukn5GuZMyd4/j3gzc3HAH/jJKi8GLadjuotzl/3iN56QeaaxI6cuJ7fsorQSN/gNMkHQapfDfHcNStm+ujy9vOP4O21Ntz6zxz0fpG3pZzeNByuC0JSiVhsOBP1FaMhodSmnZuBD4MKXFoZGAq2uav7d9QY98/kFpzVmuXqNptl/oPlfKe7Z1fc9+A8wlaTEGwdOTp7DwZmsCMGaNFXluyr2Dkc2QjqHd+UdEi3V2DtzPHCaVhoh4ke3/AtsDP5b0ekpT6MW1QrEZ5Q75ncCngN1tb0K5E75BTWLN+vsdwM28XPe35DeBU23/H3AxL1U6OinfS3cDq0gaU7dvQuna1JgGwE8od/g3Bs7tkdctwPoAkpYClu6x/4vA1ba3B37VEMP9klapj9drOL63NulbKRUtJC1NaTV4osb0cUrrzI6S3tDwmu2BU2xvSrlGjdPtdae5dk3zE5I+2yOftwMPUbpsvUXS/LUr1ur19VMoi/uMB7aq5/ZYL7G/ak+cdw2d06az0tmHs8yBO3P/QScPRjZDOoZ25x8RLda6gdBDTsY0RMT/sH1LHTR1NPBRYLykyyndi86yPVXSjcDlkqYCD1KmedsJ2ErSNpTxDzv2k82vgO9J+hrwANB9J/xayh36jwEHAhdL6gTuAPah9O9vjPXBGsM1DVPRdfsdZRDZtcC9wKM99v8B+KGkbSkF/RmS5qUM3D5Z0tOUVpcH+zmPQ+uxHwbmBybanibpMeAa4DlKN6r7Gl5zHaVS9gylAtSz0vBl4ARJ+1PGNGxfYz1R0pcoLSK72H5E0uGUVp1HKF26AE6ox15KGWtxbO0bO/C6urjva8cPStJzTAztzj8iokU6urrmvJpORAw93QOMbZ/f4nzPAfayfccApbcHcGYtlH8LmG77GwOR9lA1edyE/CMYAtZ64Ox2hxAxHHW88iHNe+aAjw7Y9+WYb545oLENtrQ0RMQcSdL8wBWU8RgDUmGo/gNcUFsanqTMEhURETFHdisaKKk0RMSAsL1ji/N7jpfGUAxkur/mVa5PERERMdyk0hARERER0YSuOXDWo4GSSkNERERERDNGcPekTLkaERERERH9SktDREREREQzRnBLQyoNERERERHN6Bq5YxrSPSkiIiIiIvqVloaIiIiIiGake1JERERERPSnawRXGtI9KSIiIiIi+pWWhoiIiIiIZozgloZUGiIiIiIimjGCV4RO96SIiIiIiOhXWhoiIiIiIpqR7kkRERHRTpPHTWh3CKz1wNntDiFiaBvBlYZ0T4qIiIiIiH6lpSEiIiIiogldXSO3pSGVhoiIiIiIZqR7UkRERERERO/S0hARERER0YwR3NKQSkNERERERBO6RnClId2TIiIiIiKiX2lpiIiIiIhoxghuaUilISIiIiKiGZ3tDqB90j0pIiIiIiL6lZaGiIiIiIgmjOSB0Kk0REREREQ0YwRXGtI9KSIiIiIi+pWWhoiIiIiIZozggdCpNERERERENCFjGiIiImZVRwfLHrobC6zyRrqmz+CeLx/DtHseGlkxtDv/oRJDRAx7GdMQEW0jabykhyVdIuliSddI+uwsvP4SSSs1eewpkrbsse2Nkq6Z1bhnh6Tf9rJtd0kHDVD6A5ZWsxbZcl1GzTsPU7bZhwcOO5VxB+zUyuyHRAztzn+oxBAxYnQO4M8cJi0NEdFuF9neFkDSvIAl/cz2E+0Na2DZ/mC7YxhoC669Mk9ecj0Az1x/G2NWW37ExdDu/IdKDBEjRbonRUQMDWOBmcAMSZcADwOLAu8BTgLeDIwGjrT9y/qab0haDJgG7AA8BpwALAO8Hvi97f3rsZ+R9GXKd98uwIzujCVtAhxS878T2M32Cw37VwB+DMwDPAtsC4wBTq7pdQGfA94AfMD2TvV11wNbAjfYXlLShsBRwOM1//9p6ZA0P/CTms48wJ7A5Lrtf86/r7Rqa812NaYzbB/d1NWfRaPHLsDMqc+++LxrZieMHgUzW3cLrd0xtDv/oRJDRAx/6Z4UEe22We1mdBHwc+Cztp+u+35he3PgU8AjtjcANge+VSsKAL+1vRnwB+BrlMrCNbbfDawD7N6Q11W23wl8G/hO90ZJHcCJwAdtbwI8COzYI87vAYfZXp9SUF+9bjvK9sbA5ykVmz8C60saI2lt4C7bDzekcxzw8Xped/dyPXYH7qn5bAusC+zWx/m/LC1JqwAfAzYENgImSFIv+bxqM6c+y+gx87/4vGNUR8sLqu2Ood35D5UYIkaMFnVPkjRK0vGSrq7/I5fvsf9TkibXbr3vrduWlfSXevyl3d/9kvaWdHPdfsns/k9IpSEi2u0i2+Ntb2b73bbPbdjn+ntl4DIA21OBW4Dl6r7L6u+rAFFaGtaW9HPg+8C8Den1PLbb4pRWiTNrC8e7KHf6Gwm4usbwe9sX9IjrH8AytmcCvwY+COxEqYw0ep3t2+rjK3u5Ho353G77B/2cf29pvbXGfmH9eS2wQi/5vGpPT57CwputCcCYNVbkuSn3DkY2QzqGduc/VGKIGCm6Ogfu5xVMAOarN5D2AY7o3iFpSUrL9juAdwOH1e693wSOsT0eOBQ4rL5kTWCH+r92vG0zG1JpiIihrPtr9VbKXXMkjQVW5aW79OvU3xsBN1FaCJ6w/QnKl+wCtSWht2O7PQo8AGxTv2wPAS7qEcutwNo1hk/ULkCNcb0d6J6y5iTg/yitBH/ukc6Dklauj9fu5Zwb83mzpNP7Of/e0jJwM7BpPZdTgBt6yedVe+K8a+icNp2Vzj6cZQ7cmfsPOnkwshnSMbQ7/6ESQ0QMuA2B8wFsXwOs1bBvHeBK29NsPwncAbwN+CKltRtKt9nn6+M1ga9JukLS12Y3oIxpiIg5wSTgRElXAPMDB9t+uLawTpC0F/AU8ElgKeB0SetTxjncXrcBrFe7QXUBOwMdALY7JX0e+KOkUTWtHXrE8GXgBEn7U8Y0bE/pEnWipC8Bc1PGSWD77hrb72z3vJ+0G3CqpKeAqZTxCI1OAE6WdCll/MJelEJ/b+f/srRs/1PShcAV9c7TdZTuVgOvq4v7vnb8oCQ9x8TQ7vyHSgwRI8UA9vyTNBGY2LBpku1J9fFCwJMN+2ZKmsv2jF72TQUWtv1oTVeU7rMT6v4zgB9R/redJem9ts+Z1Xg7urpG7ijwiIiAyeMm5B9BALDWA2e3O4SIgdbxyoc079GtNhmw78vFzru0z9gkHUkZn3dmff6A7XH18fuBLW1/pj4/CzjE9mRJmwLHAv9Xn3cAC9UWCSR9Bnit7W/OarxpaYiIiIiIGFquBN5HGWu3HnBjw77rgEMkzUcZt7cycFOtMBxFqVB0D25aqO5bGXgG2Iwy698sS6UhIiIiIqIZrZuY7CxgC0lXUVpLdpL0BeAO27+XdDRwOWV88n62n5f0A8pU3T+tXWRtezdJ+wIXU7rsXthjwpGmpXtSRMQIl+5J0S3dk2IYGtDuSY9sMXDdkxb/c9/dk4aizJ4UERERERH9SvekiIiIiIgmNLG+wrCVSkNERERERBNGcqUh3ZMiIiIiIqJfaWmIiIiIiGhG1xw1dnlApdIQEREREdGEkdw9KZWGiIiIiIgmdHWO3JaGjGmIiIiIiIh+paUhIiIiIqIJ6Z4UERERERH96hrBA6HTPSkiIiIiIvqVloaIiIiIiCake1JERESMeJPHTWh3CKz1wNntDiGiT5k9KSIiIiIiog9paYiIiIiIaEJXV7sjaJ9UGiIiIiIimpDuSREREREREX1IS0NERERERBNGcktDKg0REREREU0YyWMa0j0pIiIiIiL6lZaGiIiIiIgmpHtSRERERET0q6tr5FYa0j0pIiIiIiL6lZaGiIiIiIgmdHW2O4L2SaUhIiIiIqIJnemeFBERERER0bu0NERERERENGEkD4ROpSEiIiIiogmZcjUiImJWdXSw7KG7scAqb6Rr+gzu+fIxTLvnoZEVQ7vzTwwR0SIZ0xARg0rSPpL+IulSSRdLWrMNMXxA0lJNHrulpFMGOP99JK3TY9t8ku4ZoPQHLK1ZsciW6zJq3nmYss0+PHDYqYw7YKdWh9D2GNqdf2KIaK2uroH7mdOkpSEiBo2kVYD3A++w3SXp7cBPgdVaHMrngd2Bf7U4XwBsH96OfAfbgmuvzJOXXA/AM9ffxpjVlh9xMbQ7/8QQ0VrpnhQRMTieBJYFdpZ0vu1/dN9xl3QJsLvtKZJ2B5YEDgfOBBYGFgD2s32BpLuAa4HlgJuAXYGxwEnAa2ten7N9o6RdgE8Do4HfA9cBbwdOlbQhsBuwHdAFnGH7aEkrAycDz9SfxxtPQlIH8ENgHWAe4EDbv5N0BLBhPex04FjgVmA1289I+hIwk1JJOgO4Avg58Brgjt4uWM/4bR8o6RPAXsA04HZgIjBvb2lJWhU4GugA/gvsbPvJXt+dV2n02AWYOfXZF593zeyE0aNgZusmMm93DO3OPzFERKuke1JEDBrbD1JbGoCrJU0B3tvPS5YDFgPeB3ycl25sjAMOsL0OsCAwAdgXuND2ppRC9HGSlgD2ATYC1qAUrC8F/gHsACwPfIxS0N8ImCBJwHeBr9veHLiql7gmAIvV/DcF1pL0XuBNwHo1ve2AlYDfAB+qr9sOOLUhnd2Bm2xvDJzQM5Pe4pf0BuBgYDPbGwJPUCo+faV1IrCH7fHAucBXejmfATFz6rOMHjP/i887RnW0vJDY7hjanX9iiGitzq6OAfuZ06TSEBGDRtLywFO2d7a9LLA9cLykRXsc2gFg+2ZKAfgXlLv23d9R99nuvpt+FSBgVUoLxiWUgvKiwJspBennbHfZ3sf20w35vBV4A3Bh/XktsAKwIqVFAuDK3k4FuLrG+LjtA4CVgctrPi8A1wCrAD8GdqgtKrb934Z0XszH9rXACz3yeVn8wBLAzban1mMuA97ST1orA8fW67IzsHQv5zMgnp48hYU3K0NUxqyxIs9NuXewshqyMbQ7/8QQ0VpdXR0D9jOnSaUhIgbT24BjJM1Tn99GuVM+E3geeH3dvga82LVmrO33AJ+kdAkCWFrSkvXxO4CbgSnA9+sd9Y8CpwF3AitJmrem92tJSwOdlO8719duWl93CnADcAuwfk1/7V7O49bu7ZIWlvSnum3Dum1uYAPgdtu3UypBX6ZUZhq9mI+k1YG5e+x/WfzAw8AqksbUYzap17GvtAzsUM/vK8A5vZzPgHjivGvonDadlc4+nGUO3Jn7Dzp5sLIasjG0O//EEBGt0tE1Jw7fjog5hqT9KIX6pykF92/bPlvS1sARwH3Ag/X34ZTC/xL12BNs/0zSE8BfgGUod/T3orQsnAQsAiwEHGT795J2pHTd6QL+YPtQSd8CtgTeBexC6W40L+VO/WeBN1IGaE8DHgGet71jwzl0UMYJrE7pMnWw7fMkfY9SiZkHONP2t+vx2wHfAFaoA8BPoYxpuITSXWkpSqVnI9vqcb16i3+7es6dlPELu9bDX5ZWnZ3qiBpnF7CL7dv6e48mj5uQfwQxZKz1wNntDiGGlwG9pX/DG983YN+Xb7vnD3NUc0MqDREx5El6yPaSr3xkzI5UGmIoSaUhBtiAFsz/8Yb3D9j35dvv/f0cVWlI96SIiIiIiOhXplyNiCEvrQwRETEUzIkDmAdKKg0REREREU1oVa9+SaMoswiuRhlvt2vDLIJI+hRl+u0ZwLdsnyNpMcqaQfNTFjPdyfazvR07OzGle1JERERExNAyAZjP9vqU9XuO6N5RZxP8HGUijncDh9VZ974OnG57I+DvwG79HDvLUmmIiIiIiGhCCxd32xA4H8D2NcBaDfvWAa60Pc32k5RZ9d7W+BrgPGDzfo6dZemeFBERERHRhIEc0yBpIjCxYdMk25Pq44WAJxv2zZQ0l+0ZveybCizcY3tv2xq3z7JUGiIiIiIiWqxWECb1sfspYGzD81G1wtDbvrGUhVO7tz/Xy7aex86yVBoiIiIiIprQRLeigXIl8D7gTEnrATc27LsOOETSfJSFSlcGbqqv2Ro4BdgKuLyfY2dZxjRERERERDShawB/XsFZwPOSrgK+D+wt6QuS3m/7IeBoSqXgImA/288D3wK2lXQlsD5wTD/HzrKsCB0RMcJlRegYSrIidAywAW0auGapDw7Y9+V6//rtHLXoQ7onRUREREQ0oYXdk4acVBoiIiIiIpowkleEzpiGiIiIiIjoV1oaIiIiIiKa0NnuANoolYaIiIiIiCZ0Dey46jlKKg0RERExZEweN6HdIWQGp4hepNIQEREREdGEzhE8QXUqDRERERERTegcwd2TMntSRERERET0Ky0NERERERFNyEDoiIiIiIjo10iecjXdkyIiIiIiol9paYiIiIiIaEK6J0VERERERL9GcvekVBoiIiIiIpowkisNGdMQERERERH9SktDREREREQTMqYhIiIiIiL61Tly6wzpnhQREREREf1LS0NERERERBM60z0pIiIiIiL609XuANoo3ZMiIiIiIqJfaWmIiIjZ09HBsofuxgKrvJGu6TO458vHMO2eh0ZWDO3OPzEMrRhi2BvJ6zSk0hAxzEgaD5wJ3EJpSV0IuAv4hO3pg5TnssBqtv8wwOn+ADgSeBrY0vbpkvYBLrJ93WymORH4ie0XZuE1fZ6fpKOA71Jabk+mfK92ABNte3ZibCKeh2wv2cRxBwO/tH3LYMSxyJbrMmreeZiyzT6MWWNFxh2wE3fucthgZDVkY2h3/olhaMUQw19nx8gd05DuSRHD00W2x9ve1PaawAvA+wcxv82Adwx0orb3sn0f8DZq/LYPn90KQ7UvMHoWX9Pr+UlaD5hh+wHgm8AxtscDhwJDobTyfeB7g5X4gmuvzJOXXA/AM9ffxpjVlh+srIZsDO3OPzEMrRgihrO0NEQMc5LmAV4PPF6fHwZsRCk4H2n7V5LWBX5AuZHwIPAJYHngaMpd8/8COwOrA18FpgNvBs4ADgf2ARaQdJXt3zfkfRdwLbAccBOwK6Xl47T6ey5gf9sXSToE2LRu+43tb0u6BNgd2A9YrbYSbFDznQgcZftSSWsBBwAfBo4HVqjnsr/tSxri2QVYsr5+Qh/X4jPAJymt0H8F9u7r/IDPAUfUx18EnqyP5wKe7/E+nAUcYnuypCnAvrZ/K+kCYKd6Xl8AZgJX2N5H0sLAScBru/OzfWNDmocCCwN7Ulo5lgfmr9flZ7afkPScpLfZvoEBNnrsAsyc+uyLz7tmdsLoUTCzdQ347Y6h3fknhqEVQwx/GQgdEcPNZpIukXQLcD1wlu0LJW0FvMn2hpQC+n6SFgFOAHa2vS7wR2Bl4ERgj3rn/FzgKzXtNwAfAtYDvmJ7JqXicHqPAjXAOOAA2+sACwITgP2BP9veGPgIcJKkDkpFZTtKIf6JHukcQmk9mdSw7URK4R5KoftESqXk0Zr2NsCPGhOxfRLwELBtP9diJ2BP2+sDt1IqTX2d3ybAjTXtR22/IEmUu/sH9zj2LGArSW8CpgGb10rBfMBz9fh31niWlrQFpVXkQtubUipJx3UnJul7wFy296jXdmPgg8CWlIpHtxuA8QyCmVOfZfSY+V983jGqo+UFtHbH0O78E8PQiiGGv84B/JnTpNIQMTxdVAv7G1FaBe6u21cF1qx38M8H5gbeCCxp+1YoBWvb11MqDsfWY3cGlq5p3Gh7hu1nKIXd/txn+476+CpANd3Lal4PAk8BS1AqDYcDfwIWaeIc/wSsI2nRep7n1fPbusb8G2AuSYv18fq+rsVOwB6SLqVUkPrrwDq6cZyIpE2Bs4H/62U8wx+ALSiF+m8D6wBb1e3LA4sD59Z4VqG0zqwK7Fy3nQgsWtN6HaXL1oIAtqcCewGTgF8C8zbk+29eaqkYUE9PnsLCm60JwJg1VuS5KfcORjZDOoZ2558YhlYMEcNZKg0Rw5jt/wLbAz+W9HpgCnBxrVBsRhkwfSfwL0krAEj6qqQPAAZ2qMd+BTinJttb62wnvX+fLC2pe8DuO4CbKXfvN6p5LQ28htKy8BHg45S7/jtKekN/6dvuBH5Fuft+dm3xmAL8osa8Vd3/WB+x9nUtPgXsbnsTSnesDfo5v+ckja7nsilwFGXA9uSeB9p+HHgW+BilknIf8Hngt5RK3f3AFjWeHwLX1Bi/X7d9lNKtC+A/wLuBt0jasr63a9r+APAe4DuSurufvgZ4uJfYX7UnzruGzmnTWensw1nmwJ25/6CTByObIR1Du/NPDEMrhhj+OjsG7mdOkzENEcOc7VskHU0Zn/BRYLykyyl3qc+yPVXSbsDJkjopd6Z/QCnUnloLn13ALsBSfWRzI6V7z/W2z2jYPg04RtIylELwH4Aral4fpvS/n2h7mqTH6jHPARfU/LvdCawqaa8e+Z5MmRlqhfr8BODE2kqwEHBsrVw0upzS3WrTPq7FjcDlkqZSxndcS2kN6e38rgTWoIx9+AEwD/DT0kMJ296tR96/A3ay/ZikPwGfsX0ngKQjgUtrJeQeSiXmEEr3rYn1fA7qTsh2Vx2jcT6wLrCkpKsoXZO+Z3tGPXRdSjengdfVxX1fO35Qkp5jYmh3/olhaMUQw95IXhG6o6trJA/piIjB1OzUoHMqSesD29r+fLtj6U3tuvVT2+/r77jJ4ybkH0FEg7UeOLvdIcTAGdBS/s+X2n7Avi8/8a/T5qgaSLonRUTMJttXU8ZNjGt3LH3Ym8FqZYiIGIG6BvBnTpPuSRExaIZzK0O3OnvRkGT7gHbHEBExnMyJYxEGSloaIiIiIiKiX2lpiIiIiIhowpy4vsJASaUhIiIiIqIJc+JYhIGS7kkREREREdGvtDRERERERDRhJA+ETqUhIiIiIqIJI3lMQ7onRUREREREv9LSEBERERHRhHa2NEiaHzgNWAKYCnzS9iM9jjkQeA8wA9jL9nWS3g78EJgJTAN2sP0fSUcBG9a0ALax/WRf+aelISIiIiKiCV0dA/czGz4N3Gh7I+BUYP/GnZLWADYB1gW2BX5Udx0FfNb2eOC3wFfr9jWBd9seX3/6rDBAKg0REREREXOCDYHz6+PzgM172X+B7S7b9wFzSVoc2Nb2P+oxcwHPSxoFrABMknSlpJ1fKfN0T4qIiIiIaMJAdk+SNBGY2LBpku1Jdd8uwN49XvIfoLs1YCqwcI/9CwH/bXg+FVjY9h01zQ2APYGNgTGULktHAqOBiyVNtn1DX/Gm0hARERER0YSBrDTUCsKkPvadBJzUuE3Sb4Gx9elY4IkeL3uqYf//HCPpY8B+wHtsPyJpNHCU7Wfr/ouA1YBUGiIiIiKaMXnchHaHwFoPnN3uEGLouRLYGrgO2Aq4vJf935H0PWAcMMr2o5K2B3YDxtt+rB67IvBLSatThitsCPy0v8xTaYiIiIiIaEJXe7M/DvippCuA6cB2AJK+A/y6zpR0OXA1pSKwR21ROBq4D/itJIBLbR8o6WfANcALwKm2b+4v846urjaffkREtNXkcRPyjyBiiElLw4AZ0DWcj1p2+wH7vvz8fafNUetLZ/akiIiIiIjoV7onRUREREQ0oZ2Lu7VbKg0REREREU0YyZWGdE+KiIiIiIh+paUhIiIiIqIJI3nWiFQaIiIiIiKa0DlHzXc0sNI9KSIiIiIi+pWWhoiIiIiIJozkgdCpNERERERENGEkj2lI96SIiIiIiOhXWhoiIiIiIprQOYLbGlJpiIiIiIhowkge05DuSRERERER0a+0NERERERENGHkdk5KpSEiImZXRwfLHrobC6zyRrqmz+CeLx/DtHseGlkxtDv/xJAYoqXSPSkihg1J4yU9LOkSSRdL+pukX0maZxDzXFbS+wYh3R/UtBeVtF3dto+kdV5FmhMlzT2Lr+nz/CQdJWlcw/NNJN0/u/E1GU9TpRBJB0taZbDiWGTLdRk17zxM2WYfHjjsVMYdsNNgZTVkY2h3/okhMUS0SioNEcPTRbbH297U9prAC8D7BzG/zYB3DHSitveyfR/wNmr8tg+3fd2rSHZfYPQsvqbX85O0HjDD9gP1+TLAF4BZqpQMou8D3xusxBdce2WevOR6AJ65/jbGrLb8YGU1ZGNod/6JITFEa3V2DNzPnCbdkyKGudrC8Hrg8fr8MGAjSsH5SNu/krQu8APKjYQHgU8AywNHAx3Af4GdgdWBrwLTgTcDZwCHA/sAC0i6yvbvG/K+C7gWWA64CdgVWAg4rf6eC9jf9kWSDgE2rdt+Y/vbki4Bdgf2A1aTNBHYoOY7ETjK9qWS1gIOAD4MHA+sUM9lf9uXNMSzC7Bkff2EPq7FZ4BPUlqh/wrs3df5AZ8Djqhpz1fzngj8rZf34SzgENuTJU0B9rX9W0kXADvV8/oCMBO4wvY+khYGTgJe252f7Rsb0jwUWBjYEzi5vmfz1+vyM9tPSHpO0tts39Azpldr9NgFmDn12Refd83shNGjYGbrGvDbHUO7808MiSFaayRPuZqWhojhabPaPekW4HrgLNsXStoKeJPtDSkF9P0kLQKcAOxse13gj8DKwInAHrbHA+cCX6lpvwH4ELAe8BXbMykVh9N7FKgBxgEH2F4HWBCYAOwP/Nn2xsBHgJMkdVAqKttRCvFP9EjnEErryaSGbSdSCvdQCt0nUiolj9a0twF+1JiI7ZOAh4Bt+7kWOwF72l4fuJVSaerr/DYBugvxxwDfs/0gvTsL2ErSm4BpwOa1UjAf8BxwMPDOGs/SkragtIpcaHtTSmXkuO7EJH0PmMv2HvXabgx8ENiSUvHodgMwvo+YXpWZU59l9Jj5X3zeMaqj5YWjdsfQ7vwTQ2KIaJVUGiKGp4tqYX8jSqvA3XX7qsCa9Q7++ZRuNG8ElrR9K5SCte3rKRWHY+uxOwNL1zRutD3D9jOUwm5/7rN9R318FaCa7mU1rweBp4AlKJWGw4E/AYs0cY5/AtaRtGg9z/Pq+W1dY/4NMJekxfp4fV/XYidgD0mXUipI/TUij7Y9XdJSNYYDa3qLSjqjx7F/ALagFOq/DawDbFW3Lw8sDpxbX78KpXVmVWDnuu1EYNGa1usoXbYWBLA9FdgLmAT8Epi3Id9/81JLxYB6evIUFt5sTQDGrLEiz025dzCyGdIxtDv/xJAYorW6BvBnTpPuSRHDmO3/StoeuFjS24EpwMW2J0oaRenScyfwL0kr2L5d0leB2wADO9i+T9I7KF2coPfvuk56vwmxtKQlbT9EGRPwM0oBdiPg75KWBl5DaVn4CPDx+rpbehS6X5a+7U5Jv6LcfT/b9sza7ecB24dKmp/SremxPmLt61p8C9jd9vOS/kTpNtTX+T0nabTtf1EqREAZqGx72x7xPi7pWeBjlBaBDwOfB7anVJzuB7aw/YKkHYF/ACsCp9k+XdISlJYUgP8A7wYukbQl8E9gTdsfqN2k7pf0M9sz6vV9uJfYX7UnzruGhTZajZXOPhw64J4v/HAwshnSMbQ7/8SQGKK1RnK7USoNEcOc7VskHU0Zn/BRYLykyyl3qc+yPVXSbsDJkjopd6Z/ANwHnCppLkpFYRdgqT6yuZHSved6242F/WnAMXWA8DWUu+pX1Lw+TOl/P9H2NEmP1WOeAy6o+Xe7E1hV0l498j0ZuIsyhgFKN6sTayvBQsCxtnt+x19O6W61aR/X4kbgcklTKeM7rqUU6ns7vyuBNShjH5rxO2An24/VCslnbN8JIOlI4FJJo4F7gDMp3bJOqmM5FgIO6k7Idlcdo3E+sC6wpKSrKF2TvlcrDNR9+zYZ36zp6uK+rx0/KEnPMTG0O//EkBiipUbymIaOrq6Re/IRMbjqHfcl2x3HYJG0PrCt7c+3O5be1K5bP7Xd73S4k8dNyD+CiCFmrQfObncIw8WAzlP01Td+fMC+L799zy/mqDmUMqYhImI22b6aMm5i3Cse3B57M1itDBERI1DGNEREDILh3MrQrc5eNCTZPqDdMUREDCcjeUxDWhoiIiIiIqJfaWmIiIiIiGjCSB4InUpDREREREQTRm6VId2TIiIiIiLiFaSlISIiIiKiCSN5IHQqDRERERERTegawR2U0j0pIiIiIiL6lZaGiIiIiIgmpHtSRERERET0ayRPuZruSRERERER0a+0NERERERENGHktjOk0hARERER0ZSR3D0plYaIiIiIIWbyuAntDoG1Hji73SHEEJJKQ0REREREEzJ7UkRERERE9GskL+6WSkNERERExBAnaX7gNGAJYCrwSduP9DjmQOA9wAxgL9vXSVodOAe4vR52nO1f9nZsf/mn0hARERER0YQ2d0/6NHCj7YMkbQvsD3y+e6ekNYBNgHWBZYDfAGsDawJH2j6iiWP7lEpDREREREQTBrJ7kqSJwMSGTZNsT+rnJRsC36mPzwMO6GX/Bba7gPskzSVpcUqlQZK2obQ27NXXsT1bLhql0hARERER0WK1gtBrJUHSLsDePTb/B3iyPp4KLNxj/0LAfxuedx9zHfBj23+TtB9wIPBEH8em0hARERER8Wq0qnuS7ZOAkxq3SfotMLY+HUsp+Dd6qmF/4zFn2e4+9izgh8Dv+ji2T6OajD0iIiIiYkTr7OoasJ/ZcCWwdX28FXB5L/vfLWmUpGWBUbYfBf4kaZ16zDuBv/VzbJ/S0hARERERMfQdB/xU0hXAdGA7AEnfAX5dZ0q6HLia0jCwR33dp4EfSnoBeAiYaPupPo7tU0fX7NV0IiJimJg8bkL+EUTEywyTFaE7BjKx7d/wwQH7vjzt3t8OaGyDLS0NERERERFN6BzBi7tlTENERERERPQrLQ0REREREU0YyHUa5jSpNERERERENKHNK0K3VbonRUREREREv9LSEBERERHRhJE8EDqVhoiImD0dHSx76G4ssMob6Zo+g3u+fAzT7nloZMXQ7vwTQ2IYijEMYyN5TEO6J0VTJI2X9LCkSyRdKukqSR+t+94u6ettiGlLSRMHOf1TXmUaG0n6/ACF1HaStpK0yyy+5sXPh6QPSFpK0hslXTM4Uc5SbBtLetsgpPvb+ntVSRvXx2dImmc205tP0q6z8bpBOb9ui2y5LqPmnYcp2+zDA4edyrgDdhqsrIZsDO3OPzEkhqEYQwxPaWmIWXGR7W0BJC0IXCrpNtv/AP7R6mBsn9/qPGeFpA7gIMpS78OC7fMknSfpV7afavI1/+Clz8fngd2B5wcnwlm2M3AGcMNAJmr7g/Xhhyirb17W/bczm5YEdgV+PIuvG5Tz67bg2ivz5CXXA/DM9bcxZrXlByObIR1Du/NPDIlhKMYwnI3kgdCpNMRssf20pBOAD0taBNjd9raS9gQ+CIwBHgU+AIwGTgWWAu4HNra9lKRLKIXJtwILAR+xfa+kLwLbAjMoha2vSnoHcATwAvAs8GFKgWwlSsH8TGBhYAFgP9sXdMdaC+8/BNYB5gEOtP07SUcAG9bDTrd9lKSVgZOBZ+rP4zWNjwBfAGYCV9jep7eYbE9tuExbALfYni5pPPBtyrLvk4D7gENqencCuwFvAn5Sz3sUZXn45YD9KN9TSwKTbP+o8b2Q9AlgL2AacDswEfgEpcA4qp7vhbNyLYBjgVuB1Ww/I+lLwEzb3wfOBXYEjq5pjgZc34vFgQeAJYCnKcvTf4FSUfgZ8HbKZ2F7YHFJZwOvB26w/ake53UTcFu9ZrsBJwGvrbs/Z/vG2urxacpn7Pe2D+zjvTqoXt8lgDcAe1M+n1sCa0i6xfZ9Nd/xvV1zSavXazeTUun5FPAwvXz2JD0ErFmv03RJ19fjVgX+3vO6Ar+mfC7mB54DJtq+v+Fy7AesUltsjurjWvwEWL6mcRRwS2/nN5BGj12AmVOfffF518xOGD0KZrbu32q7Y2h3/okhMQzFGIazrq50T4qYHf8BFut+ImkUpSCzue11KZXStSmF2Lttv4NSwH9dQxrX2d4c+DPwcUmrAh8FNqg/K0h6LzCBUujaBDgOeE1DGsvVON4HfJyXV4YnAIvZXgfYFFirpvkmYD1KYXm7mvd3ga/XmK6q57UocDDwTtsbAktL2uIVYgIYz//e4Z3P9kbAacCJwAdtbwI8SClcbgFcB2wOHEgpiAIsDby/xrq3pCUarvlra2yb1dieoBSwAR63vWF3hWFWrgWlAvAbSsWMuu3U+viGem4A2J4JXAasTymk3gS8s/5c0HDcHymVxB0oFYGFgJ3q697ZeF7VgsA36x36fYELbW9K+TwdV4/fB9gIWAOYV9Ky9P5eAUyzvRWltWNv238Dzge+0kuBurdrfiKwZ33PjgWOpJ/Pnu0HgVOAI21fVze/0Md1/R5wtO3x9fHhPeI5hFIB/UYf12IssDGlwr4lpYLX3/kNiJlTn2X0mPlffN4xqqPlBZN2x9Du/BNDYhiKMcTwlEpDvBpvoNxVBsB2J6Uw+AtJJwHjgLmBlakFcNtTgEca0vh7/X0/MB+lsHqN7RdsdwGXA28BDqW0VFxIaWV4oSHfm4ETgF9QCnM9P9ei3PHG9uO2D6gxXW67y/YLwDXAKsCKlII7wJX19/KUO+jn1taRVSiFxT5jqhajVKxeDLX+Xpxyd/3Mmt676rU8iVLoPx/Yk9LiAHCV7Wm2n6MUyJdrSPPNwM0NLRyX1evVmN/sXosfAztIWqcc7v/WNP7NS3e5u/0W2Bp4N+Wu+BaUQvdveomh2101hk7KHfsFejmm+xxWBXau1+tEYNF67jfZfq7Gvg+lZaC39wpe/lnrT2/XfKna1QrqdW7is9eb3q7rqsC+Neav878V655edi3q+78XpbXil8C8TcTxqj09eQoLb7YmAGPWWJHnptzbimyHVAztzj8xJIahGMNw1knXgP3MaVJpiNkiaSFK94xfNWx7GzDB9seAz1I+Xx2UQtf69ZjuO7Pdev7VTAHWlTRX7UqzMaWLyvbAKfXu6s2UO6zd+a4KjLX9HuCTlC4kjW6ltHggaWFJf6rbNqzb5qa0atxO6dKxfn3d2vX33ZSC5hb1TvAPKQXrPmOqHgYWaXjefavnUUpla5ua3iHARcA2lML7OynX9av1+LdLGi1pAUqF4PaGNO+mdFsZU59vUq9XY36zdS1s3055/75MKZx2e009t0Z/rnkvRum+tCbwdtt/7XFcJy997zTzjdl9DlOA79fr9VFKa82dwEqS5q2x/5pSSevtveorv8Z4GvV2zf/VMKh4E+C2Jj57L0u/j+s6BfhqjXk3Gv6ueknnZddC0uuBNW1/AHgP8B1Jc/VzfgPiifOuoXPadFY6+3CWOXBn7j/o5MHKasjG0O78E0NiGIoxDGedA/gzp8mYhpgVm9W7mzMpn50DbbsWWADuAJ6R1H2H/t+UO/EnAadIugy4l34Gwda+2WdS7vKPAq4Azqb0wf+xpGcof2sTKQU3KAW6A+tsTqMod2ob/R7YXNIVNe6D64De8ZKupvTtP9P29XU8xU8lfZnSIvK87UckHUkZ+D0auIfSLWneXmJqdAllTMepjRttd9YZlf5Yu3Q9RemyM7bmvT+lj/7elC48cwPnUe7uf8v2ow1pPSrpQOBiSZ31PdiHMiakN01fi3r8ScA3gIsb0liX0rrSeE7TJN0P3FvPz7y8YgGlxenUXq7VKzkEOElltqyFgIPq+/JtyvvSBfyhjonp7b3qy7XA4ZLutn1rw/aXXXNJnwKOqZXZGcAuwL/o/7P3N+C7km7tsb3ndf0SpZvRfJQxCT1n3HoYmKee78uuBWWw9ZKSrqL8fX7P9gxJfZ3fwOjq4r6vHT/gyc5RMbQ7/8SQGIZiDDEsdYzkAR3RGpI2ABasA0RXAM63vdwrvW5OVysEFwHvsj19NtMYTx1kPpCxvRqSzgc+6iZnT5rTDMVrPtgmj5uQfwQR8TJrPXB2u0MYCB0Dmdh7l33PgH1fnnPfHwc0tsGW7knRCncBX6stED8H9mhzPC1R++ofDHym3bEMFEnvAX4zXCsMERER/RnJYxrS0hARMcKlpSEiepOWhpfbetmtB+z78tz7zp2jWhoypiEiIiIiogkj+WZ7Kg0REREREU2YE2c9GigZ0xAREREREf1KS0NERERERBO65sABzAMllYaIiIiIiCbMibMeDZR0T4qIiIiIiH6lpSEiIiIiogmZPSkiIiIiIvqV7kkRERERERF9SEtDREREREQTMntSRERERET0qzNjGiIiIiIioj8jt8qQSkNERERE9GLyuAntDoG1Hji73SFElUpDREREREQTRvLsSak0REREREQ0YSRXGjLlakRERERE9CstDRERERERTciK0BERERER0a90T4qIiIiIiOhDWhoiIiIiIpqQFaEjIiIiIqJfI3lMQ7onRUREREREv9LSEBERERHRhHYOhJY0P3AasAQwFfik7Ud6HHMg8B5gBrCX7esknQEsWQ95I3CN7W0l/Q5YDHgBeM72Vv3ln0pDREREREQT2tw96dPAjbYPkrQtsD/w+e6dktYANgHWBZYBfgOsbXvbuv81wMXA3vUlKwBvsd3USaV7UkRERETE0LchcH59fB6weS/7L7DdZfs+YC5JizfsPxj4oe1/S3odsAjwB0lXSHrvK2WeloaIiIiIiCYMZPckSROBiQ2bJtmeVPftwkstAt3+AzxZH08FFu6xfyHgvw3Pu495RNISwDsb0pwHOAI4ClgUuFLSdbYf7iveVBoiIiIiIpowkFOu1grCpD72nQSc1LhN0m+BsfXpWOCJHi97qmF/z2M+DJxue2Z9/hBwvO0ZwMOS/g4I6LPSkO5JERERERFD35XA1vXxVsDlvex/t6RRkpYFRtl+tO7bnNKliYbnvwKQtCDwVuDW/jJPS0NERERERBM62zsQ+jjgp5KuAKYD2wFI+g7w6zpT0uXA1ZSGgT0aXivgru4nts+T9G5J1wCdwL4NFYxedYzkRSoiIgImj5swe/8IOjpY9tDdWGCVN9I1fQb3fPkYpt3z0ABHN8RjaHf+iSExDPMY1nrg7FcdzatNoNFbXrfugBWcb/7PtQMa22BL96QYFiS9RdIfJV0s6a+SDpY0S3+MkvaRtE6PbfNJuqeXY3/7KkNuTGtHSe8fqPSGou5rW6/nrnXbQZJ2HwKx7TkIaW5ZB7ghaaKkuSW9XdLXX0Waq0raeDZeN+Dn122RLddl1LzzMGWbfXjgsFMZd8BOg5XVkI2h3fknhsSQGKJV0j0p5niSFgHOAD5o+3ZJoyn99HYDjm82HduHz8KxH5zVOPtJ65SBSmuo6r62kt4I7Ar8uK0B/a/9gWMGMkHb5zc83Rc41fY/gH+8imQ/RBm4dtksvm7Az6/bgmuvzJOXXA/AM9ffxpjVlh+MbIZ0DO3OPzEkhsTQWm3untRWqTTEcLANcJHt2wFsz5S0A6W/H5KOoMxdDHA6cCxlsM9qtp+R9CVgJrAapfJxBfBz4DXAHb1lKOkh20tKuoRSEHwrZaqzj9i+t+G4HYEJlBkMFgO+Yfs3km4CbqsxTqEUBqcAXwOmURZlOR7YrMZ1lO3jJH2Y0kdxbqAL+EDN+9s1rb8A77G9Ts3/l8ARtq9riGn/GtNcwHG2T5B0GLAW8Frgn7Z3knQQsAGwILALZdDUdjXfM2wf3ZDmNsAWtveUtA+wge33S/oE8AZgxXptPwSs0nDHfRtJH6n5HmD7Dw1pjm84r0nAfcAh9b26k1IpnAv4Sc1jHmBPYHK9ditQWlP3t32JpBuAS4G31XPYph6/qKRjbX+mIe9TKE3ay9Tz38H2FElfBLalrLR5me2vSnoHZdq6F4BnKTNUfAhYCbidsgrnGZJ+AOxO+Qx+wPZONa/rgS0pC/J8oZ7fFbb3aYhnaWBHYHo9fv5ersWb6rWYUc97O2CH3s5voIweuwAzpz774vOumZ0wehTM7BzorIZsDO3OPzEkhsTQWgM5e9KcJt2TYjhYiobBPQC2n7Y9vS5W8iZgPUrFYTtKYe43lIIdddupDS/fHbjJ9sbACU3kf53tzYE/Ax/vZf8YYAvgXcCRkuaiFES/2b1KY4NxNa5PU+4Q/x9lhoTd6v4VKZWCDYFbgHfX7fPZ3sj2wcBzklaRtCjwph4VhtVreusC6wArSloYeNz2FpSKw3q1kApwq+0NKAXoj1Gu4UbABElqiPsCoLvrzMbA0vU83w80duU6BLjF9jfq8wdtvxPYq55zT/PZ3gg4DTiR0pq0CfAgpRC9O3CP7fUphfl1KS0Zj9b3bxvgRzWthYBfNLx+K9uHAI/1UaC+0/ZmwEHAdyStCnyUUpHaAFihfr4mAGdSCv3HUSqbwItT5j1UY+v2R2B9SWMkrU357M6gLLrzzvreLi1pi4Z0HgROAY4E/trHtdgCuI5SuTsQWPgVzu9Vmzn1WUaPmf/F5x2jOlpeKGh3DO3OPzEkhsQQrZJKQwwH91LuCL9I0ptq/++Vgcvr6ogvANcAq1C6x+xQxzDYduNiKCtSCl/YvpZyB7k/f6+/7wfm62X/pbY7bf8HeBzoXp3RvRx7U43zCUqhdXp9TXe6D1NmTvgJ5Y753L2kdSKlELkdpbDdSJRKzkzb021/kXJ3fAlJv6BUkhbsJd23Uu7mX1h/Xku5k18Osp8DbquF4O7rvDGwrO0pvZxnt7/V3w8BC/Syvzv/xYHXA2fW1p131XhEmSUC27fb/gGwKrB1Pe43lBUxF6vpvNJ71eii+vuqms9KwDW2X7DdRZnq7i3AoZSK64WUVoZ+Py91juxfAx8EdqK8X8vXczy3xr0KsFwfSfR1LU6ifG7Op7SgzHiF83vVnp48hYU3WxOAMWusyHNT7n2FVwy/GNqdf2JIDImhtTq7ugbsZ06T7kkxHJwD7CvpONt3Spqbckf2z5RuSDsB36/bNwB+Wsc+dABfptwdbnQLsD7wu3pnfm7690p/+WsC1CXbF+KlhVN6u+3SZ1q1ReBgYNm66c+8NCtEY1q/Br5EWRXyIz2SmQJ8WtIoYDRwLqW/+zK2P1aXm/9AL+kauJlyd75L0t7ADT3SPgv4LnA25e75oTXGRp38782KV7p23fk/CjwAbGP7yTpw/GlKBWFtynv1ZuBblArLA7YPlTQ/sB/wWD/59TVgfk1KV7V3UM59CvDF2oIyk1IpOhXYHjjF9pckfY2yumfjf8ie5wylgH8CpfK1Z/19P6WL1wu1W9s/erkWo/q5FttQKsgHS/o48FXKZ3/QZud44rxrWGij1Vjp7MOhA+75wg8HK6shG0O7808MiSExtNZI7p6USkPM8Ww/JemTwIm1MDwW+AOlv36XpPGSrqb0eT/T9vX1pScB3wAu7pHk8cCpdR7kKZQxBq/GkpIupCzl/pk65mJ20nmKsnDL1ZS7yI9T7nDf3XiQ7eclXQYsbvuxHvv+Ien8ms4oSoXpWmD/+pouSoF/qR6v+2c9hyskzUtpiXmwR3znACcDn6EUgH/Ny7scPQzMI+nbwHPNnrjtTkmfB/5Y3+OnKP31rwJOlnQppRK0F3Aj5bNwKaWSdmx9fV/J3yLpNNvb99i+VR2rMRrY0fbdks7kpWt3BaWCtA7wY0nPUAr2EyldlbpdTqmcHdxwPnfXeH5nuxN4RNKRwKV1IP89lC5Pjf5GqZTdCvR2LcZSWqH2rzHv/Qrn9+p1dXHf15qea2BwtDuGduefGBJDYogWyToNEYOo3jFeqXFQa4vy/RHwG9sXveLB8TJ1IPQZPWZBGrZme52GiIhBNtTWaVhusTUG7Pvyzkevn6PWaUhLQ8QwI+kCykDgVBgiIiIG0EjunpSWhoiIES4tDRExVA21loY3L7b6gH1f3vXo39PSEBEREREx3HR1jdypY1NpiIiIiIhoQucI7p6UdRoiIiIiIqJfaWmIiIiIiGjCSB4LnEpDREREREQT0j0pIiIiIiKiD2lpiIiIiIhoQronRUREREREvzpHcKUh3ZMiIiIiIqJfaWmIiIiIiGhC1wgeCJ1KQ0REREREE0bymIZ0T4qIiIiIiH51jOQaU0REAIzg9vaIGO46BjKxxRfWgH1fPvKkBzS2wZbuSRERERERTRjJN9vTPSkiIiIiIvqVloaIiIiIiCaM5HUaUmmIiIiIiGhCuidFRERERET0IS0NERERERFN6BzBk82l0hARERER0YR0T4qIiIiIiOhDWhoiIiIiIpqQ2ZMiIiIiIqJfXSN4TEO6J0VERERERL/S0hARERER0YR0T4qIiIiIiH6N5NmTUmmIiIiIiGhCxjRERERERET0IS0NERERERFNSPekiIiIiIjo10iuNPTbPUnSeEkPS7pE0sWS/ibpV5LmGayAJC0r6X2DkO4PatqLStqubttH0jqvIs2JkuaexdcM6PlJ2l3SQU0ee4akeRpjqO/tSgMVz+yQ9Ntetr3svCRtKWniAOZ7xmB+loc6SStJuqQ+bsu16O29H+D0H5qFY18r6YTBjKeVJL1O0jHtjiMiIgaGpPkl/UbS5ZLOlbR4H8ctL+nGhueLSbqgvu6Xkhao2z8labKkayS995Xyb2ZMw0W2x9ve1PaawAvA+5s8v9mxGfCOgU7U9l627wPeRo3f9uG2r3sVye4LjJ7F1wzK+TXD9ra2p7czht7Y/mCTx51ve9IA5tt9PUa8dl2LZt/7FvkW8KN2BzFQbP8HmCppk3bHEhExXHQN4M9s+DRwo+2NgFOB/XseIOn/gDOAxgrF14HT6+v+DuwmaUngc5Ty4LuBwyTN21/ms9Q9qd6JfD3weH1+GLARpeB8pO1fSVoX+AGlQvIg8AlgeeBooAP4L7AzsDrwVWA68OZ6gocD+wALSLrK9u8b8r4LuBZYDrgJ2BVYCDit/p4L2N/2RZIOATat235j+9v1juruwH7AavWO9QY134nAUbYvlbQWcADwYeB4YIV6LvvbvqQhnl2AJevrJ/RxLT4DfBLoBP4K7N3P+e0JfBAYAzwKfADYDtgaWKCe97dtnyJpQ+Co+j7MAK7p8T7N3Vvsku4B3tIYQ33JgZJeV/P+uO27+jifS4CHgUWBd9ueWfNbAfgxMA/wLLBtTevk+h50UT6YbwA+YHun+rrrgS2BG2wv2cR57QisVM/tF8D99bpcZ/vTPY69BJhSj+8APlYff5vymZsEfLMhvRdqfPPW9/R9wLLANsA9wAnAMpTP/+9t7y/pFOC19eefwE22fyTpNcBfaiW7O57FgZ8Ci9R4dgCeA44D5qvp7m/7bEk3AbfVOHcDTqp5AHzO9ot3D2raRwAb1qen2z6qR2zvsd39N/t64Oc1hoca0rinXovlgSMp7/tiwKdtX1U/73sCj9W4fllf2tvnc3Xgh8BM4HngU5TPzZnAwvX4/WxfIOmh+t7/z9+K7c/1OMf3AgfWuK+n/C2/k1LQf56XvlemUt7btwB31vcTScvU7fPX6z7R9v0N6S8ErN39OZJ0L+Xzc0u9Hv/zWuCRPs7nZd9Ttjsb8unt2owC/lDP4Vzb32k4fhfKP4nRlM/dgZI+AewFTANur/H8kh7fYba3AU4HDgYupX8dr7A/IiKAGdMfHLDvy1oWbexBMekVbo5uCHT/jziPUl7t6XFgE8r/wMbXHdrwukPr/ittTwOmSbqDcmP9r31l3kxLw2a1C8stlH/WZ9m+UNJWwJtsb0gpoO8naRFK4Wpn2+sCfwRWBk4E9rA9HjgX+EpN+w3Ah4D1gK/UQujhlILPiwXqahzlH+E6wILABEoN68+2NwY+ApwkqYNSUdmOUuh9okc6h1BaTxrflBMpBRaAnerzXYFHa9rb0OMOpO2TKIWubfu5FjsBe9peH7iV8o/5ZecnaRSlcLd5vW5zAWvX3Qvbfi+ldWSfuu04SuF+c+BuXq6/2Hu7xn+0vRnlg/Thfs4H4Be2N++uMFTfAw6r53kUpUL4PUohZmPg85SC7x+B9SWNkbQ2cJfthxvSeaXzarQisAuwDrB1rTH3dFX9zP2S0ioEMJ/tjWz/rMex99h+F+V9epPtrYHfUCoPywDX2H53zW/3htddZHsD4LuUigCUz97Pe6S/P6XQtwHwxZrOSsARtregfGnsUY9dEPim7W1r3Bfa3rQec1xjorUw/SbK39CGwHaSVm2MrbvCUO1HeQ83Bc7u5Zq9Bfii7XdSKlg7SVqMUsF/B/AuSoWwW2+fzxMpn/tNgGMphe7lKJWQ9wEf5+U3LP7nb0XSi/vr42MolZ+1gDso78kk4IM1n0vrNf4A5T1eD/gapUAP5fN4dP08fI/yN9BoPcANz5cBtrO9dx+v7et8evueatTbtYFyA+JdPSoMS9RruhGwBjCvpDdQKgGb1b/PJygVy96+w6BUerorlBERMYTYnmR7rYafF8umknaRdFPjD+VG1ZP1kKn1ec80z7H9TI/NC/XyusZtfabXqOnuSZR/XNN5qTC3KrBmvaN7PjA38EZgSdu31sBPsn09peJwbD12Z2DpmsaNtmfUk3vuFeK4z/Yd9fFVgGq6l9W8HgSeApagVBoOB/5EubP7Sv4ErCNp0Xqe59Xz27rG/Btgrlp46k1f12InYA9Jl1IqSL3WTuudyOnALySdRCl4dI+V+Ef9fT/ljjTA62zfVh9f2Uc8zcYO8Lf6+yFKIauv84H/LVh1E3B1PZff276A/31v/gEsUysav6a0qDQWbLq90nk1usP21Jrmv3np2jS6qP7u/rz0FT+UCjGUQtgt9fHjNd3HgLUl/Rz4PvXudWN6tu+idAVZhfL5O7VH+o3X6CrbP69x7ybpZ5SKSOP4mO44VwV2ru/FiZRWnkYrA5fb7rL9AqV1ZpV+znVFoLtLXm/X+EHgAEk/pbS2zU1pfbjF9rP1el/VcPw/6u/Gz+dS9T2H8hl4i+2bKTcUfkEpLPf87unvb2Ux4PHuCmYtWD8LPFX/7l/Mp/H8XLojdrcmrArsW6/j14HX9ch/MeA/Dc8ftf3fvl7bz/n09j3V6GXXpj6+2y/vHvZmSuvVc/X93Yfy/Xaz7ak90ujtO4z6fr1Qb0xERMQcopah39r4Qynkj62HjOXlN8b78lQvr2vc1lR6Tf8jqf9Atwd+XLs4TAEurhWKzShN9XcC/6rdVZD0VUkfoBRedqjHfgU4pybbW5euzj7iWrrhbvI7gJspd4U3qnktDbyGcsIfodz92xTYsd6d6zP9Wmj/FeUu7tn1H+0Uyh3Z8cBWdf9jfcTa17X4FLB7vau4OqU71Mvyl/Q2YILtjwGfrfu7C029XaMHJa1cH6/dy/5Xir1nDD3z6Ot8ul/b063dcUj6hKTP8r/vzdt5qSvMScD/AesCf57F82rUTHfA7u5B3Z+XvuJ/pfR2BJ6w/QngCErXru73pzG9EylNhQ/YfrRHGo3XaGNJ36Z0jzrV9v8BF/O/BeXudKcA36/vxUcp3fF6prthTXduymfs9n7O9RZg/fq4t2t8NHCg7U8CN9aY7gBWUhmANYrSStKtt+v2r/qZhtJEeltt/Rhr+z2UO+I/7PGa3v5Wuj0MLFILxEg6mlKgXqh+F72YT+P5SVqKl25QTAG+Wq/jbpS/iUYP8783GBqv3cte28/59PY91e+16SW/bndSrnt3F6tf1zhXkdTd2rMJcFsf32HUz+kMN3SRioiIOdaVlG7BUMp3l7+K110HbCRpPkkLU25C3tRfIrN098n2LZRCxdGUPrhPS7qccqe6q9792g04ud4xXJ3SHenTwKmSrqC0ANzQTzY3AttI2rbH9mnAMZKuBf5V8z+U0n3qMkpXi4kufbMeo9xxvRi4ALivIZ07gVUl7dUj/ZMpd8BPrs9PoPzDvpRyx/DeXv7xXl7Pr69rcSNwuaSLKP/sr+3j/O4AnpF0JaUg/W9gqX6u0W6U63kh5a5sT68Ue1/XuFtf59OXLwNfq3diP0HpmvMl4LP1vTmO0pUI290tVb/r5Xq+0nnNqh3rNXgPpVva7LoQ2LLhXG6n9/fnLGBzSsWop0Mp1/wSSveSEyiFvO/VdLeg3O3u6RDgow2tPv/zB237HOBuSVdTPvO/rq17ffkW8IGaXm8TGpxGKRRfTrlrv1StAH2b8nk/n9K3/4V+8vgU5W/1ckrXtL0p12x8PddfUe7YN+rtb6X7HDuBzwB/rN8hHZQvu08Bv61/N5tTKmG/A/5bvyd+QBkfBOXzeGD9PJzKy7+DrgFW6+N8enttX+fT2/fUK12bXtl+hHLdL63v7/W276WM7bhY0jWUz0x3l7We32FQWkmu7iuPiIiYoxwHvKX+L5xIKU8g6TvqfzbQb1G6019JubF2jO2HKOX5yyk9M/az/Xx/mXfMKfPNqg6YbHccMWeoheLdbU9pYZ4LUPrWrzuc7uzWMQVftX1IvXN9GeXL5bI2hzagJB0PnGD7768ijSH1PSXpO5SxNFe0O5aIiJizpZ9rxACQtAHl7vi3h1OFAcD2DGCMymxXV1PGfzTbJDon+TqlRWNYqN2kFkqFISIiBsIc09IQERERERHtkZaGiIiIiIjo1ywt7hYREdGtzpS3AmVw+IO2W9p03e78G+JY1HbP2fVamX+734exlLVclqLMjnhDw9TDrYrhrcB4yppHD1PWt7mt3xcNzxjG2X6g4bls9zXV+GDFcIztPRuen2p7h/5eE3OGVBoiImKWqaxi/wHK2iE/paznsWe/LxpG+dcYNqEsnjla0q8oM9X1NnvaYMbQ9utAmbHrPMoUwA9RZpDbpBUZ12m6v0dZu+VGyqxlrwEOq5M47FvXVRnuMbyVMsX0tyV1L6A7GjgMePtg5t0Qwx6URTYXlfRBykx3Hbx86unBjmMjYC/KdOTTgRmU8XjH2L6qn5fGK0ilISIiZse2wMaUu6k/kPTXEZY/lGl+N6YsonkoZS70llYaGBrX4bW2T5a0ve2rWryY4Mcoq7c/2XOHpNdQpjXuOcXzcIzhNZTPwuso61RBWf/l2EHO90W2fwT8SNK+tg9tVb6NJP2QsmjZgZRFSTvr9lWB7etndNhMeNFqqTRERMTsGEVZ3K+7K8y0EZY/QKftxyR12X5eUn/r2QyWoXAdkLRS/T2Ocme3JWwf1Esso2x32n6cwS+sD5UYLqesdbNG91o93TEMdt69+KmkVSifg68CR9v+Z4vy/qbth3tutH0j8FVJr2tRHMNSBkJHRMTs+AVlzY7lJZ1LWWCzlU5vc/4Ad0g6DHitpH2Ae9sQw1C4Dp8HfgKsAfwa+GKrA5D0CUnbSvok8JCkL43EGICVG2L4d5tiOJ3S4nEoZcHaH7Qq4+4Kg6SzJW1d1xZq3P+fVsUyHGXK1YiImC21L/dbgSn1Tl678rftnqt8tyL/uYBdKStv3wpMsj29DXG09ToMBZKuA7YCzgDeC1xguyXjKhLDy2K4GNgc+JPtzSVdaPudLY5hTWAnyriGs4GTbd/XyhiGo3RPioiIWSbp5IanW0l6Abgf+FHtkjFY+fbW1WNlSRNsf2Ow8u0Rw8YNT2+pPwDrUe76tyKGoXAd7ualblEALwBzA9Nsr9yKGBo8V39PtT2tVuhabSjE8PwQiGFu4DvAZZI2BeZpdQC2/wb8rY4pOQ64HZi31XEMN+meFBERs2N+yiwxv6R0y1ma8k/5p4Oc73/qz/rAksCdlJmD3j7I+Tb6dP35HnA08H/AkUBLCuvVULgOKwGrABcD29oW8CGgHauQ3wVcA5ws6UDK9LMjMYY7h0AMO9U4DgcWBz7Z6gAkbSTpJOBSSqV+uVbHMByle1JERMwySX+xvXnD8wtsv0vSZbY37u+1A5T/Bbbf1fD8z7a3GOx8e8TwR2Ab2zMkjQb+aHvLFscwFK7DJbbHNzxvyWeglzgWtP20pCUprR2D1uI1p8Rg+6E25D+aUnFYFrgIuMn2oy2O4TfAJEr3rBR0B0haGiIiYnYs1DBjzkrAWEmvBRZsUf6LSlqu5i9g4Rbl2+j1DY/nApZoQwxD4To8Iembkt5XB4b/u9UBSPqh7afr09WAa0doDG8BzpN0E7CjpPe2OgbgBEqFYQtgLHBqG2L4GGXNkmMkfV5Sy7tIDUcZ0xAREbNjD+DnkpYC7qvPPwYc0qL89wLOkrQE8CCwe4vybXQScHMtoL2F0h2j1fai/dfhEzXf91C6ghzUhhieknQ4pdL6Vspg4JEYw9GUu/wnUj6f51FW6W6l5WzvKmlD23+oM4u12gnAE5TZmzYBfgxkVepXKZWGiIiYHZvYXrPHtsmtytz2FXWQ5XLAXa3u/lBj+JHKStBvBm63/d82xND260AZfPs8ZTGxDv53cHRL2N5P0neB5Ru7So20GGocd9S1Qx5p09ohc0laDEDSWMrnotVWaOgid7akrAQ9ANI9KSIiZsfWte9yW0j6KHAV8DXgGknbtyGGVYE/UqZ0/Iuk1dsQQ9uvA6Xv+JuBC4A3Uu7qtoSkf0v6l6R/AdsD72p4PmJiaPCYpN2AMZK2pdxtb7X9KKujr0UZlH1wG2KYT9ICAJLmB9r2XTWcpKUhIiJmx+LAvxqm3eyyvUEL898bWLMO+BxLGXB5Wgvzh9IVZFfb/5T0duBHwDtaHMNQuA5tu6tr+/WvfNTwj6HBLsC+wKOUQvvObYjhWduStHiNo+WD4oGjgH/WroOrAAe2IYZhJ5WGiIiYHe0YYNmos3vQqe2pkp5/pRcMgg7b/6wx/EPSjDbEMBSuw3ySFrD9bLvu6tYBwMcDr6FUmm6y3dK+/EMhBuBztl8cQ1AHpn+tFRlL2ohSQN9b0pF18yhgT8oYj5ax/XNJl1MmJ3g4C7sNjFQaIiJidswNfKT+7gCWAnZrYf53STqCspjaxpR54VttZp2d5vIaw7Q2xDAUrsNQuKs7FAYAty0GSbtQVidfWdLWdfMoysJqLak0AI9T1gyZl5dmFusEvtKi/F9U16iY1/a+kn4labLtb7c6juEmlYaIiJgdpwNnARtSFnlr1VSr3XaiVFK2oMzY044ZWnamLPB2eI3hU22Ioe3XYajc1R0CA4DbGcNpwIWUrkndM5h1Ag+3KgDbNwE3STrRdjvGczR6f/dEDbY/IulKIJWGVykDoSMiYnY8bfsw4AHbOwKva3H+awKjbe9JGUfQ0u4P1dPAJNtvpdzpf7INMbT9OtS7urvbngwcIemrrY6BoTEAuG0x2J5m+x7bE23fW3/ut93y1q8hUGEA6Oxem0HS3KS8OyByESMiYnZ01VVvx0oaQ+tbGo6hzFwEcACli0yrnUHpigHwGK0fgAxD4zq83/a+UO7qAu9vQwy7AG/ipQHAu4zQGKI4ntLq8RvgH8Bx7Q1neEilISIiZsfBwAeAnwF3UbpGtNILtu8EsH0X7ZkLfkz3QFfbpwMLtCGGoXAd2n5X1/ZTwF+A31FWIH52JMYwVEh6p6SJkt4mab5W52/7JErL23coa8q0Y1XqYSdjGiIiYpbZvozSJQfg920I4V5JhwJXA+tQVkNutemStqDMRb8O7SmwD4Xr0H1X90ZgJdrQd7xeg3HAypQB6V8DPj4CY1iacv2XAH4F3GD72hbH0PbrAGD7EeCRVuc7nKWlISIiZpmkAyU90r2IVRsWstqJMshza0rBoB3z0e8K7AFcB3yG1s4e1a3t12GI3NXd0PYOlLE2P6V0ExqJMUwCTqbManYZ7emuNhSuQwyCtDRERMTseC+wrO3n2pG57eeBH7Qj74YY7gAmtDmGtl+HGke77+rOVbvBdNWVymeO0Bjmt32RpP1dVlhrx7odbb8Okr5k+3utzne4S0tDRETMjoeBF9odRET1feBvlNmjrqWszj0SY3he0ruB0ZLWA9pRaRgK12HrWmGJAZSWhoiIaJqkXwBdlClW/14X9ALosr1d+yKLEe4KypohywN32350hMYwkbJ2yGLAl4BPtyGGiykDwtt5HRYD/iXpbsr3VZftDdoQx7CSSkNERMyK49sdAICktwFjKIOPDwUOtd3SGZzqVLOvobS4TAROtX1vi/L+el/7bH+jRTF0F8i6vUDpSz/N9sqtiKHBKZTpb/9AmfK0HQXVoRDDh4BP2368DXl3+xNlZfIT21RhAHhfm/Id1tI9KSIimmb7UtuXAgsBm9XHXwNaPa3i8ZSZWfYH9gMObHH+AL+mLK72XUqBeVIL8/5P/VkfWJJSSFsUeHsLY1gJWIVyZ3lb26IUWq9oYQwA2H43ZX2IB4DTJP19JMZAuRn8F0k/lzS+DflTV2L+DjBB0l8l7deGMGZSWlzOpYz56WhDDMNOKg0RETE7DgaOrI8/RusL7c8DNwPz2L6G9gw6XYAy3ew424cDLetDbfsE2ydQVoP+jO2f294LGNvCGKbVgdjL2b6ubvs7oFbF0E3SBEqL05eAh2jDYl5DIQbbR9RC+w+Az0i6rdUxVDdTpgF+DNioDfmfSFlD5h3AT4GT2hDDsJPuSRERMTtesP0kgO0nJbW60N5FWUDrXEkfpT2DsucBPg/8TdIqlO5SrbaopOVs3ylJwMJtiOEJSd+kTD27AfDvNsRwGKXl6XDgfNtPjMQYJM1Pae35JOXuestb4CSdDKxHaYnbzfY9rY4BmM929/oxZ0v6QhtiGHbS0hAREbPjOkmnS/qspJ8Bre6K8THgp7aPokz1uW2L8wf4IrAUcAiwGaUC0Wp7AWdJ+g9wOmXdiFb7BPAEZRrefwM7tDqAOoZiAqXS9FtJ14zEGIAbgLdQxjVsbvsXbYjhLOCttr/epgoDlGlfVwWov7te4fhoQloaIiJiltn+bO2OIeBM239oRb6S3mv7HOCD9fnEumsFWjSmQNI42w9QBrr+mLL67gWtyLsn21cAb2tH3pLWsj2Z0gXkxvoDsAktvh6S1gC2At4FPAuc2cr82x2DpLlszwBWB6bXbfMA2J7eohj2t/0tYDvg46Xhq2jDzGqfBU6W9HrgX8CnWpz/sJRKQ0REzDJJSwO3AbcCX5F0v+1/tCDr19bfr29BXn35Qv05gXIHs3uQZRelxWHQSfq17Q9L+jc97qLaXqoVMQDvBCYDH++xvYvWV6L2B34LbNOmrkntjuFUSmH9Rl7+mXxzi2LovnHQ9hnW6nfR2u2OY7jp6OpKi01ERMwaSZcCB1G6w3T3Xd60xTEsTJlydQJwTjunmZS0jO3725Dv4nU15raRtBelq1g7p/kMQNLatv/a8Hy87UtalPdoymQAZ1C6D3bU53+03ZLKdAyutDRERMTs6AQuA/azfYakljb/SzoDOIcy8HYUpbvSB1ocw5cpffkXAXaSdL7tVg+4PF9S95z4f25x3t1GU6b5nFLjuKRNcYxYkjakjGXYW1L3rGajgD0pKzO3ws7AvpQpgE2pNMykDVPwxuDIQOiIiJgdc1PmYr9M0qaUmYRaaSnbpwEr296dFk412uBDlOkct7K9CqU/eUsNhTnxh9A0nyPZE5TC+ryUrnuvBxYHvtKqAGyfaPtNwB6232z7TbaXt71jq2JoJGmCpK9Kem878h+O0tIQERGzYydgC8r859tQpnhspXkkfRC4RdJitKfSMJNSUPtPfT5/G2KAl+bEX542zInfzmk+Ja1AmeL0OeBg27fX7cfZ/nSr4ugR0+mtHvhr+ybgJkmTbLdjylsk7Wr7x8AKkg7tEd++LY7lx5TvhKuAHSS90/berYxhOEqlISIimtYwY86bgDsoM+U8QSmw3tXCUL5D6Tf9ReBzwDdbmHe3S+rP9pK+D/yx1QEMkTnxb6j5f9r2HS3OexJlfYS5KfPxb18XmFupVQFIuo+XylMdlLUz/g10tWpQevfAeOB6Sd2DVTtaGQPQPaZnSovy68+qttetj49q0/S3w04qDRERMSuGxIw5tn9bu8GsD/y2RTM39YxhP2A/SYsDX7HdjgXmzgJ2sd3OWU1Wpky9u7qkBVv9Xti+AEDSHZT1EbaktfPy/x+wN6XS9G9JF7d6UoBaYcB222YVs/2n+vDXwGuAGZSpTk9tQzh3SHqT7bslLQHc14YYhp3MnhQREXMcSZ+jVFyupQyGPtP291ocw3jgZOApymDoT7V6MLKkt1G6iI0DHqJUIK5vcQyfpUz32fL3QtKfgaOBc23PrONrfgDMXceZtISkccBxwPeAA9s1W5CkzSk3hEcBPwQOsH16i2M4j3ItPgzcAmxq+90tjuEuYGlKZWFpykrdz9PalpdhJy0NERHRtIZ1AeYFFqB0SVgaeMT2G1sYyseBjWzPkDQ3pe9ySysNwLeADW3/q65b8Vug1TMYHQXsavufkt4O/Iiy2ForbUf73oudKV3TrgQes31xnQL2+y3KHwDbD0j6MHAM7V1D5BDK+9H9OTiTslJ4Ky1AWbNhL9s71IpMS9lu1doUI0pmT4qIiKbZfn29U3cesKLtFSmrMV/b4lA66gq41G5B7egaNNP2v2oMD1LuZLbaKNv/rDH8g9IlpNXa+V7MtL2j7ce6N9i+2PbbASQN+l3l7jxsT7P9KXos8NeKGBo8SxmYP8P2Q7S2m1a3eYDPA3+TtAowptUBSNpc0paStpZ0p6RWr0g9LKWlISIiZsebuxczq3fal21x/ldI+jVwOWXGoCtbnD/AU7VrzmXAxsBjr3D8YJhRp5S8vMYwrQ0xtPO9+IqkFyh302+srR0dwBqUsQajgc+2OIZ/tyGGbk8B5wOTJO0BPNyifBt9kbLg4rco5//5NsQwFFpchp2MaYiIiFlWpzScF7iOMhj5Mdt7tjiG91AG4d5qux0zFy0M7F9juAU4rNWrIkt6A6UrUHcMX7Z9bytjqHF0vxe32D63xXmvS5lBaxNKD4rnKBWXY223ZNacoRBDjWNeYDnbt0h6C3CH7ZZXJOvnYRXgNtu/a0P+FwPvA86w/V5Jl9repNVxDDepNERExCyTNIqyAvOKlIJiSwsGksYCWwHzdW+z3fJZWurMLI0xtHSWFkmvBdaw/WdJewKn2X6ixTG8D1jL9oGSzgeO7J7RKFqrDsj+PrXADuzd6ml4JR1G6bJ4BaX1627bX2xxDL8DXkuZkncsMN72R1oZw3CU7kkRETHLbHcCv2ljCL8D/sVLc8O3/A6YpGMpFZd/U+fEp8we1EpnUAZDQ+kedRrQ6hVwDwa6pxj9GGW8SyoN7XEiZeaiy4DxlJm13tniGDa2/Q4ASUcB7Vgj4aO81OLyVuDHbYhh2EmlISIi5kSjbG/f5hjWoRRMOtsYwxjb5wDYPl3Srm2I4QXbT9YYnpQ0sw0xRDGf7d/Xx2dL+kIbYphb0qj6d9FdmW61xYB9a0vgryiDsVs9WcOwk9mTIiJiTnSDpHUlzStpHknztCGGO2jomtQm0yVtIWmspHcC7ajAXCfpdEmflXQq8PdWB1AHgzc+/+hIjAGYS9KqNf9VaU+B/ZfAlXWV9CsorWGtNomyhsrclFaXo/o/PJqRloaIiGiapF/QR0HEdiunNdyEMtCxWxfQ6rnZlwXurSsRQ1k4qtXdk3alDIQ+mjIQercW54/tz0qaQFkV+tcNd7oHXS2ovwP4uKTuaz8aeD9lxpwREUODzwEnS3o9pfvexBbnj+0jJP2J8nn4se2bWx0DML/tiyTtb9uS2jEd8rCTSkNERMyK49sdAIDt1dodA2WBubayfQdlest2x3F2m7L+J2XA63OA67ZO4BcjLAYAbP9d0tbAcpSZi1o+DbCkFSlTngq4SdIX2zCj1/OS3g2MlrQe7VlDZdjJ7EkRETHLJC0KvJvS/N8BLGX7sBbmfzvlbm63FyiDor9i+/oWxXByj03dMfyo1VOvjnR1Ni8o3a7XB661PX0ExvAZYC/gJuAtwDdtn9biGK6hDI6/CtgQ+JLtTft/1YDHMI7SArcqcCtlKuK7WxnDcJSWhoiImB1nUf4Zr0q5i/dsi/O/iDLA8XJKAW1X4CeUbjobtiiG+YE7awzrAWtTFtP6KaVrSrTOkZTP4xsoi6r9B/jkCIzhU8DbbD8vaQHgUsqMWq30jO3z6uM/tmkw9t62t21DvsNaKg0RETE7OmzvXu+270opOLfSirb/Uh9fIukA2xdKOrCFMSxuu7uL0p8kXWD7AEmXtSoASaOBHSkF1YuAm2w/2qK8f0Lf41t2bkUMDda2vZeki21vKunCFuc/VGL4DzCjPn4O+G8bYrhf0v6Uz+OawDRJ7wJo4fodq0hapNVrlgx3qTRERMTsmCFpPspUhl20/v/JdEm7U7pAbEApmKzZ4jgWkrSS7SmSVgLG1sXWFmxhDCdQBrxuAfwVOBXYukV5d8+K82nK+3AlpbVlnRbl32h0ff/vqTNpjR2hMYwC/iHpKmB1yvSnp0NLJyroooypWK4+/w9l/E8XrVu/YxXgUUmP1ny7bC/VoryHrVQaIiJidvyI0nf6Ako//itanP92wH7ANsCNwP9RCqutvMO9J/BzSUsB9wF7UBY3O6SFMSxne1dJG9r+g6R9WpWx7T8B1IGu36mbr5T051bF0OBU4FjK+/8dSmVqJMbQ+Nn7eRvyx/ZO7ci3RwxvaHcMw1EGQkdExCyTtJbtyfXxQsDqti9tcQyv538HYl/dyvx7iWdu2y+0OM/LgQ9QpvbcBjjX9kYtjuEqYH9KS8cGwP6tjqHGsTClm9adtp9pdf5DJYaRTNLxwDG2b+pl39uBT9tu+bTEw0VaGiIiommSNqI0/e8t6ci6eRTlrvtbWxjHSZQB0GMoA5LvogxGbhlJuwFf4KWKywxghVbGQCmsXwm8HriG0vrTajsD36Wc+y20fvAvkj5EuRZzAWdK6rL9rZEWQ7Av8C1Ja1Gmv/0P8BpgNeA6yvsTsykrQkdExKx4HFgSmJdSUH09sDjwlRbHsRplSsk/USox7ZiHfQ9gPHAesBPQ8kWsbF9qW5T+42+13fKuQbanAF+sP58D2jG15RcolcZHgW9RWl9GXAx1qtHG52pDDMf0eH5qq/K2/ZjtzwDvpHQX+xtwOrCJ7c/YfqRVsQxHaWmIiIim1Wb/myRNsv3vNobyX9tdksbYfrQNZSOAf9n+t6Sxti9p5cxNkq6ml5mLJNHqVakl7UkpIC8KnEJpcdizlTEAM21Pq3f3uyS1o2tQ22KQ9FZgaeDbkror8KOBw4C3tyiGPSh38heV9EFK61sH7alMTwXaMbZmWEulISIimibp17Y/DFwvqbvQ2kHrZyf5m6QvAf+SdAawQAvz7vakpAlAV+2qtFgL8x5Kc9BvC2wMXGj7KEl/bUMMV9RZgsbVfu0jLYbXUN6H1/HSSuWdlIHZLWH7R8CPJO1r+9BW5Rutk4HQERExR5K0IKVb0laU1XcfbnH+Y4HlKf2mvwj8wfYlLY7hzZTxBCtSVgH+iu37WxzDVcA7KJWGzSRdYbtVC+w1xrElZbHBKbb/0Or8h0IMktZo1Yro/cQwlvI3OV/3Ntst66IUgyctDRERMcskbUC5i/k64EFgV9v/aEG+u9r+saTD+N/uOetTBkEOuoaZo9avmxanjK2YpxX593ASZXrPqyh3+0+mrNnQSr8ALgPeIOlc4OxWZVwXtxtNWTPiY5QFxUZLusj2ZiMlhgavre9BY4G91TH8jrJ2SHflteV3pyUtDXwbWIKycvwNtq9tdRzDTSoNERExO34IbGf7ltqfehJlus3B1l0QmdKCvPryTmAyL3UD6dbKxau6zbR9Xn38B0l7tTh/gOOAv1BmzzJlzYpW2ZlSWVyy5g2lW04r1w0ZCjF0+z5lBq2Wtjb1MMr29m3MH8r30RH/396dx/1ez/kff5zSQgwqWsRQ9FLJIJGlyJKGNEwGNcaSVBprzC9D9oxlGNMYVBQ1CpU1S/atUkiyVM+0aKKFhFRaz/X74/O5dDmTZjrn9Hmf63097rfb9/b9fD9X3d7P23EdfV7f9/v9egOvZihoD2Xi7mo9smiQJC2N3yY5DYbN0VV15UTj/qGqtqFNhx4Akrx1fG92iFVVbTdeXjFufP0mw+F2F0+YYV3gLxi61PwDcCrDN+5fZKJToZO8D3hfVe2a5JApxlwRM8zx30m+3DjDD6vqwcAPGGcZklwzcYZbJ/lqVe2bJFXVortadywaJElL45dV9X6GpRhbACtV1e4ASQ66Bcd9/vi+EcNyoO8C9wcuZ2h/eourqgsZHoZWY9iAfT5D55pfJbn7FBm4YZbjUmCT8QVw9UTjw/DN7YuBYjj9eBHDN+xfmDADACvAw/oKkYHh7+UBwCnc8MB+S/59vDGPAJ445/MMsOHEGa6qqscxLBPbijYtmbtj0SBJWhqzy4PuBVwGfIPhzIZbdP1ykp0BquqzwN8kuW5cU/7ZW3LcJTKsN2b4EPDPSc6vqvUZloZMleFGZznGU7KnyvBJ4JNV9Xjg60murKr1k1wwVQb9D7MzcOu2CpDkrwCqai3g0iQtOu7sDrydoaPZy7nhywYtA4sGSdLNluT1VfUXDEXCk4DPJPnNhBHmPhzfimHD49Q2nO1UlOSCqrrb1AGq6g0MD0SrMsx6nMlw6N2UtgQezrCuf/+q+t7sEq6pVdWaSS5tMfaKkGH8e/kYhm/2T2T4fZjUuHzwPQxL1Y6qqvOSHDxlhiQ/Z8VqS9wFiwZJ0s02no3wGYbNzysBf8u0J+AeDPykqn7M8JD8lgnHnnVaVf0X8B2GTkonN8iwI7ABwyzHvzFhX/65GZJsAZDk76rqeIbONZOpqkcA72ZYjnIUMPmD6gqS4V8Yfh82YViq9s/8zw37t7T9GDp5fQz4F+B4hr+vt7gVZOlgt1ZqHUCSNC+tn+RDwCZJ9gRuN+Xg40FSWzMsQXh4kg9OOf5od4b2orcFPppk6lOQAS5McjVwuyRn0abt6+KqWhWgqlahzbPFGxkeVC9ieFDda4FmeHiSZwKXJzkUuEeDDIvHmZaZJFcBv59q4CTrjYdMfh7YOMnGDEsobbe6HDjTIElaGqtW1d8yfNu+NhMXDQDjYW6THui2xPiLGb5NbennVbUrQxelNwN3aJDhAODHVfUj4N5MPMswWpzk0qqaSXJVVU32oLqCZbhVVa3OcEr5ysD1DTKcNf4urlVVrwDOa5Ch+dLBHlk0SJKWxtsYDrJ6GfAihm9ZNb09GJajHAU8G9hl6gBJDq6qTzOsoz87ySVTZ2DFeFBdETK8k2GZ3J0Yvl2fbHP+HHsCuzGcU3E58LwGGeYuHXwobZYOdsflSZKkmy3JxxkKhYcAH0/ymcaRJjeeU9DaHRnanh4BrM9wOvekqmoz4OMM69Z3q6odps7A8KB6HsOD6hW0eVBtniHJUcD2wBOAv01y+NQZGFowr5rkHxn+/2Hqjflww9LBNYAPN1o62B1nGiRJN1tVvYhhg+VJwD9V1ZFJ3j7BuOfyp21drwVWAa5OssmN/1u3mKOr6lcMD8ufG5crTe0whg3phzHs8TiUoZvVlP4DeA7wPoY/i8+PmaZ01Dj+gY1afK4QGarqtcBqSV5ZVUc16mT1n9zQuejVwAcZ9npM6bYM54hsBqxTVce17qrVA2caJElLY2dg6yQvAR7GsFRpCvcGNgW+Bjw9SQE7MXy7O6kkDwdexXCY1QlV9aaqmvoQq9WTvDfJqUn+E7j9xOMDMG7CnknyKybc+DrHfsDjgVOq6nWN1rCvCBl2TPJKGDpZMXTXmtq1Sc4eM5zDcODf1A4B/puhDfDPGAoXLSNnGiRJS2NRkusAklxbVddOMejYKYiq2ijJd8Z7p1RVTTH+jfgFcA7Dkoz7MJxT8JMkr7glB62qjcfLS6rq74BvAQ/ihsO9pnRpVe0BrFFVTwd+O3WAJCcDJ1fVHYH3Aj9laLu5oDIwdrJKck3DTlbnja1fv83wOzn5kjlgrSTvGq9/UFVPaZChOxYNkqSlcVxVHc3wsLo1Qy/2Kf22qt7IDRsdL5x4fKrqSIZC4UPAM2ZPQq6q700w/IFzrvdiOOBtEbfwidx/xnMZvtG9BHjg+HlSVbU1w0bwLRmWCb18IWZgxehk9RyG/R1/DZzOMAMztVtX1bpJLqqqdRgOmtMysmiQJN1sSV5eVU9gOETqA0k+O3GEv2d4MNkB+AnwuonHB3h/ki/eyP2H39IDJ9l29rqq1gI2As5p0bkoyWVVdRLDDMOPG60dfwlwELBbwz0NzTOsIJ2srmXomnQJ8COGdsxXT5zh1QxLBn8H/AXDxmgtI/c0SJL+z6rqmbMvYC2GcxLWGj9PMf4Dx8uHMTyQfAw4g2FfwdT2vbGb44FWkxiXJp3A8E3/iVX1jKnGnpPh/Qx7Wv4APLOqJmvzOef34SCGWZbHVtV2VbXdQsowJ8uK0MnqQOBuwGMZCobDpg6Q5EvAI8cMOyf5ytQZeuRMgyTp5pjboWhnhlafUy6LeTTwvXHsuWaAG/vW/5Y0U1WfAMK42XN2E+qE9ga2SHJ5Vd0O+CrDcqkpbZ7kweP1/lV14oRjz/4+PH2J+1P+PqwIGWatCJ2sNkqyW1VtneSY8cyKSVXVAcBZSd5eVa+uqr9P8uKpc/TGokGS9H+W5J9nr6tqq6kfkue0jzwVODTJb6YcfwmHNBx71uIklwMk+X1VTTbLMcdZVXWPJOdW1Z0ZutZMYs7vw/FJ3j97f2wJvGAyLJHnrPFU6l81PJV6bYai+na06Z70gCR7AiR5cVV9s0GG7lg0SJKWVqu14zBsbPxyVZ0BvC/J1xtkOJzhROZNgTMZOuZM7ZyqegfwTYZe+Gc3yPAQ4Iyq+m/gLsDVVXUhQwvW9W/JgatqZ4a2ottW1aPG2ysBmzN8636LWxEyzNG8kxVDG+LjgfWAExkOH5xcVa2V5NdVdQd83l0u/EOUJM07Sd4BvKOqtmQ4XO6gJBv/b//ecnYgw0PZlxj2VLwfmGRvxxzPZTh5+LEMnWomXwqSZOqzKeY6lqFz1lrc0FFqMdMWTytChlnNO1kBd01SVXUn4JJGm8LfAHyvqi4F7sDQYUzLyKJBkvR/VlUfZphhWARsVlVHzP4syS4T5rg1w6FuzxqzvHaqsee4V5LZk24/WVUnNMjwmSSTb7idq6oew/A8sRLwLuDVSY646X9ruVkjyder6oIl7t92ovFXlAyzDpjy7+GfsTtw+HjQXxNJPlNVnwfWBn7ZsKNWVywaJEk3xwF/5npqPwSOBp4/nkbcwupVdZskV45FTIte8L+pqh0ZlkfNbsY+c+IMbwJ2Ad7N0NXqSIYN8lPYe3wduMT9GeBR//Mf7zbDrNWq6r786e/DNQ0ynMKfNgiYvJBJcj1w8dTj9syiQZL0f5bkG60zjDYBCrh/Vd02yQ8aZNgfOLWqfsywr6HFbMedgZfO+dziQfVKhoez68bDtCb7VjfJ3uP73HMr7prk/IWUYY6NgU9xw2zgDMOZDVPaZ+LxNJFFMzPO2EiS5peqeiHDt9snMZwIfWSSt0+cYQOGB+YNgXOBtZNkwvH/guFB/cqpxvwzOT4NrMlwTsHtgEcm+buJM/wTw/6SOzC0HD129mF+IWXQDapqJYbC6aHASQ1mXLrj4W6SpPloF2DrJC9hWBLztKkGrqr7VNXjGPrfP5DhgXlL4KMTZngBQ9vZU8csLf0dsHuSw4BvMJzWPbWdgEOBv06yKXD/BZpBQFX9O0ODgDcydHN6X9NAnXB5kiRpPlqU5DqAJNdW1bUTjn1HhoO81uGGQ+YWA++ZMMMuDMuz/gL4L+ALE469pH0AqmruvTdMnOF6YF1uWMN+64nHX1EyaLBlkpdU1deSbFtVngi9HFg0SJLmo+Oq6mjgW8DWDH3hJ5HkW8C3quoBSb4/1bhLuGpcbnFJVa3aKMOs2YfkRcADaLOK4evj6xlV9U7gsws0A+PBareZ/ZxkssP2xvGfwNDi9I9FU5Kp99msXFVbAD8b/37cbuLxu2TRIEmad5K8fHw42QT4QJIWD2gbVNWbgVUYHpjXTrJ5gxyLGoz5R0n+pGvQ2Opy6gyvAl41ng3w/5JMOfO0wmSoqoOARzMUcrMboR86cYw3MmzOv2jicec6jGHmb1fgbfzPzlZaChYNkqR5pap2Bw5J8tmqugzYrFGU/RhOhN4T+BrwmAnHnj0jo+l5GQBVNfdQvfWAv5xy/DHDI4FDgMuAO1TV85J8aaFlAO4L3LPxuQSXtu6yluQ93LBc8CUNo3TFokGSNG9U1euA+wAfAq4DzgdeWlV3TjL1OvoLk3y7qvZM8sGqevaEYz91znXL8zLgT7/FvQp4WYMM+wEPT3JBVd0F+DjDSd0LLcMFDEtxLpt43NliHuCaccbjZIaZDpIcNFGGo5M8paounB2bccYlyfpTZOiZRYMkaT75a2Cr2W9Sk/ysqp4GnMD0m2+vrqptgFXGDkZrTzVw629y55o9n6Cq7gBcn+T3DWJcn+SCMc8vquqqhZShqr7N8JB8Z+CnVXXO+KOZJFMtT1pvfD9pfF93NsNE45PkKeP7ev/bP6ubz6JBkjSfXL7k0ouxe1KLB9XnA/dm+Ib5jeP7glFVDwAOBh4E7MAw4/Cbqnp5kmMmjnPZeHbHN4FtgEsnHr91hqeP76sCc88jWHOqAEleD1BV+yb549+Fcd/PpKrqMQzPuCsB7wJenWSqU8q7ZdEgSZpP/lBVGyaZ/SaVqtqQCb/NnONf5+wf2KnB+K39K/CssWh7E8Ms0E+BzwNTFw3PAPZlKNxOZ9gAO7WWGa5maL97GPAPDEtyVmIo5B40RYCqei6wG7BJVT1+vL0SQyHzz1NkmONNDG2J381wjsuRgEXDMrJokCTNJ/sAnxz7rp8D3A14HPCsBllWq6r7AmcynNPAAjt1duUkP6yq9YE1kpwMUFWLpw6S5HdVdQLwa+DHSX6zwDJsBbyY4eyO2f0Di5n2/I4PAV8BXsnw0D6b4ZcTZph1JUMHqeuSXFRVLTeGd8OiQZI0byT5SVVtDfwNsD7wfeANjdbRF/CpOZ9ngA0b5GhltqXo9sCXAapqFRr0xK+q9wO3Bb4NPLOqHp3kpQslQ5JPMhTTj0/yuSnGvBGbJ/leVX2M4e/GrE2AL06c5TLgWOCgqvpH2hQu3bFokCTNK0l+x7AMo3WO+7TO0NiXq+p44K7AjlW1EfCfwEcbZNk8yYPH6/2r6sSFlGHOPoJ/qKpnzP3ZhC14Hw18jxv2V8yaYfqi4anARklOq6rNgPdPPH6XLBokSboZquo/k7xgTseaP5qwU01zSd5aVZ8Gfje2Gd0IOCjJJxrEOauq7pHk3Kq6MzDpKcgrQIbZPSTN2u8meet4eSbwsSRntsoC3Al4fVVtOuZ5KfCzhnm6sGhmxmVekiT9X1XVOkkurqr/cYhZkvNaZFroqups4C4MD+obMGwMvooJ+/OvIBlOZtjH8LHZPSZTq6pdgB0ZZqC+BHw8yQ8nzvB54L0MnaweCbwwyaOnzNAjiwZJkpbCeBLyvwIbAz8BXmbRoJaqalWGZUI7An8FnJhk7wY5VmZoO/svwP2TrD7x+F+bPT9k/PzNJNtMmaFHK7UOIEnSPHUYw3KQrYAPAB9smkaCNcbXysBqwDpTB6iqTwEnAn/L0EVpskMP57hVVW0+5tmcNi2Zu+OeBkmSls4VST4/Xn+2qib/Rldawq+AHwGvSrJ7owzfBrZmWJ60IcPZHZk4w4uAQ8Z2wL8AWv1ZdMXlSZIkLYWq+iBwFvBVYAuGw83+HSDJ1N1iFrSq2iHJZ+Z8fmqSIxdghvUYzi3ZjuEb/pOTTH2w2myWBzIs39sqya1bZNDy5UyDJElLZwbYaHzBcJjUzrRpMbkgVdUODCf+7lxVs52rVmZY0z/JA/uKkGGOixm+2d8Y+Evg7hOPT1W9i2Gm4UzgfQxnqkyd4ZnAK4A/7qVIspDOULlFWDRIkrQUkjxn3PC5CHgIcNICOxF6RXAqsBbwB25YArMY+PACyzArwDeAjwOva/T7+CXgn5Jc1WDsWfswFG3nN8zQHZcnSZK0FKrq34HTGb7RfQBwUZJnt8y0UFXVSkkWm6F9hhVBVR2T5Imtc/TGmQZJkpbOlkleMtvesaq+0jrQArZPVe0DXMkw8zPZ2QgrUgYLhj+6cjyr4QeMnZOSvLJpog5YNEiStHRWrqotgJ+N/fFv1zrQAvZ0YP0kVy7wDCuEqroXcC/gh8Avkky9rOVzE4+3IFg0SJK0dA4D3gPsCrwNOLBtnAXtXIY9BQs2Q1XdFngO8AiGPRa/BL4CHJHk8glzvAB4MrAmcChwT+AFU40/WrJIubaqHp7kuIlzdMU9DZIkLaWquj3Dnoazk1zROs9CVVWfA+7GcEbB7HKUXRZKhqraFXgKwzfsPwQuBO4IPBh4PHB0koMnynIcw2nQXxmX7X03yZZTjD0nw+eA2zCcGfEghi5K1zO0oH3plFl64kyDJElLoap2AvZl+G/pkVU1k2S/xrEWqre2DkDbDBckefyN3P8O8K6qurGf3VJWYiiaZr+VvnrCsWetAjwqyeKqWgn4XJLtq+qEBlm6sVLrAJIkzVN7A1sBlwD7MSzJUBvfBx4LPIthac4vFlKGJMcCVNWLq+qON/LzKdf4HwF8E7jn+I3/Jycce9ZaDIUD4/ua4/VqDbJ0w6JBkqSlc32Sqxm65MwALk9q5xDgHIbNtxcBkyzFWQEz3Ar4clUdXlWPbDA+DOc07A68DHhFkrc3yPBu4IdV9XHgFOA9VfVK4NgGWbph0SBJ0tI5rqqOADaoqgOA77YOtICtleQQ4NokJ9Dm+aZ5hiTvSLIF8O/AXlV15tQZgIOTnJ7kqCQ/bDA+4/6NhzIsGds6yQeBtyZ5VYs8vXBPgyRJSyHJK6tqe4ZvMs9IckzrTAtZVd17fN8AuG4hZqiqWwM7MSyRWgS8duoMwBVV9U6G06kXAyQ5aMoAVbUVQyepVYBFVbV+ksdNmaFHFg2SJN0MVXUrYEfgN+Na8mOrat2q+miSpzWOt1C9CPgAsAlwNLDXAs3ww3Hs5yc5q8H4ALObjdcZ31u06XwvQxvkpzB0s1q1QYbuuDxJkqSb53CGb3NfXVUvqKonMJw8+4OWoRa4nwJ7JbkD8BaGB8WFmGET4EPA/avqfg3GJ8nrgYMYCqgPMuxxmNolST4MXJbkdcAGDTJ0x6JBkqSbZ6Mkfw9sD+wBvB7YNsmb28Za0A4H7jdeb8xwqNhCzPB84P3Aw4CDqurlUweoqoMZDpX7FsM+n3dOnQFYXFWbAbepquKG7klaBhYNkiTdPJcBJLmG4b+j2yU5vW2kBe8uST4AkORtwHoLNMMuDBt/X8JQOLRYLvdXwGbAF4BNgasaZNh7zPAfDC1gD2mQoTsWDZIkLb2Lk1zaOoSYqaqNAapqI2DlBZphUZLrAJJcC1zbIMOvxxbEayS5ZMqBx/1GMCwV+yTDTMdDGFqwahm5EVqSpJtns7HV6qI51wAk2aVdrAXtpcBHq2od4AJgzwWa4biqOpphadDWwPENMpw8Lou6oKo+Atx6wrEPY5htCcMG7EXj/RlgwwlzdGnRzEyLTe2SJM1PVfWIP/ezJN+YMotuUFW3B+4OnJ3k8gWc4QkMG6JPm/gk6LkZbgf8AXg8cFKSi1vk0PJl0SBJkua1qtoJ2JdhBcWRDKd077cAMzwReGCS11bVscC/JfnixBmeB2yc5J+q6ovAfyX5r4nG/ho33uJ1Jsmjp8jQM/c0SJKk+W5vYCvgEmA/4MkLNMPrgX8br58GvK5BhucD/zxeP4Fpz6vYcxz/IuAA4B+AdwE/mzBDtywaJEnSfHd9kqsZvlGeAa5YoBmuTfI7gPH9+gYZrl9iM/ZkS1oyAtZJcmSSXyT5BHCPqTL0zI3QkiRpvjtu3JC+QVUdwNA1ZyFm+M6Y4dvAlsApDTJ8qqq+BXwHeADwqQYZqKrnjhkeClzTIkNv3NMgSZLmvaraHtgcOCPJMQs4w5OAAk5P8ulGGe43ZjgjyakNxl8XeBXDIXs/Ad6U5NdT5+iNRYMkSZqXxr78OwK/SfK18d66wP5JJjnYbEXIIE3B5UmSJGm+Ohy4DlivqjYDzgUOBvZfYBmkW5wboSVJ0ny1UZK/B7YH9mDoHrRtkjcvsAwAVNW2U495IxkeuAJkuH3rDD2yaJAkSfPVZQBJrmF4ptkuyekLMMOs1zcad66XV9WJVfWCqrpDowyfbTRu11yeJEmSenBxkksXeIaZqvoEEGAxQJJXThkgydOr6o7ALsBRVfVL4H1Jvj5hjEur6sX86Z/DpIfc9ciiQZIkzVebjS1GF825BiDJLgsow6xDJh7vz1kHuBuwNnAa8JSq2i3JMyYa/9fA/cYXDGdFWDQsI4sGSZI0Xz11zvUBCzjDrMMZ9lVsCpwJvHfqAFV1EnAl8D7gNeOBd1TVF6bKkOQ5VXUfxj+HJD+Yauye2XJVkiSpA1V1MPBb4FvAI4C1kjxz4gyvSfKGKce8kQwvZFgedRLD4W5HJnl7y0w9cKZBkiSpD/dKss14/cmqOqFBhkcDTYsGhoJh6yTXVdUqwAmARcMysmiQJEnqw+pVdZskV1bVrYGVG2RYrapO4U83IU+9t2NRkuvGsa+tqmsnHr9LFg2SJEl92B84tap+zLCe/7UNMuzTYMwlHVdVRzMs03o4cHzjPF1wT4MkSVInqmpNYEPg3CS/bjT+44BVGDpKrd/ooLsnAJsApyfx3IblwKJBkiRJy0VVfQM4HdgcuAq4MskT26bS8uCJ0JIkSVpeFiXZk2FPw2OBNRvn0XLingZJkqROVNXawEbATxudTn1dVa0OrMFwqJrPmp3wf0hJkqQOVNVewEuBHwObVtUbk3xo4hjvHjN8ETgfOG7i8RkPdnsvcEfgQ8CPk3xm6hy9sWiQJEnqw/OAzZNcVVW3Ab7B8NA8mSQfm72uqqOSXDbl+KP9gecwnEp9MPB5wKJhGbmnQZIkqQ8XA9eN138AWnRP2qOqTqmq04ATx/fJJTkLmEnyK+D3LTL0xpkGSZKkPqwE/GA8Cfr+wCpVdQRMesDai4HHA7+ZaLwbc2lV7QGsUVVPB37bMEs3LBokSZL68KY514c3yvBD4Pwk1zcaH+C5wCuBS4AHjp+1jCwaJEmS+nAWcHvgeuD/Ae9K8oOJM3wVOKeqzmY43G0myaMmznDAhDMrC4ZFgyRJUh+OAF4H/CNwNPBOYNuJM+wBPJW2S4JWq6r7AmcCiwGSXNMwTxcsGiRJkvqwGPgm8KokH6mq5zXI8HPgu0kWNxh71sbAp+Z8ngE2bJSlGxYNkiRJfVgFeBvwzaraFli1QYbVgFOr6scMD+tTbsJmHG/zKcdbKCwaJEmS+vAc4LHA+4EnAc9qkOHNDcb8E1W1I8MSrVUY9lWsleS+bVPNf57TIEmS1IdzgGuAfYFfApMdrFZVO4yX9wZqidfU9mPY23E+cCjwowYZumPRIEmS1IcDgbsxzDbcDjhswrHXGt/XBdab81p3wgyzLkzybYAkHwTu0iBDd1yeJEmS1IeNkuxWVQ9PckxVvWKqgZMcOl7WCtDu9Oqq2obhcLvHAWs3ztMFiwZJkqQ+3Kqq1gaoqtsxthud2KorQLvT5zMsi9oPeOP4rmXk8iRJkqQ+vAo4nuEU5BOB1zfIUAztTk8HApzRIMPqwDpJTgN+yvBnoWVk0SBJktSHuyYpYCPgPkm+PHWAsd3pPYGHMCyXanE+wmHAueP1Z4GDG2TojkWDJElSH3YHSPKrJDMtAlTVk4GzgM8BZ1bVY1vkSHLi+P4tfN5dLtzTIEmS1IfVquoUhmVBs/sJpt6U/BrgwUl+WVXrAMcAX5o4w2+ranfg28CDgN9PPH6XLBokSZL6sE/rAMCvk/wSIMnFVTXZWRFzPIvhrIonA6cBuzbI0J1FMzNNZq8kSZK0HFTVC4H3JrnuRn52K2CvJP8xUZZPALcBvgFswXBWw9cBkrxyigy6ZTjTIEmSNL+dAhxbVT8BfghcDNwB2ArYlGm7KH1yzvUvJhxXtzBnGiRJkjowbjp+JMNhZr9k+Ib/q602RasvFg2SJEnqRlV9D/gQcFiSS1vn6YUtqCRJktSTxwDXAMdU1Ueq6jGtA/XAmQZJkiQtF1V1V2BnhlOZAUjyhkZZNgFezVBEnAu8JcknWmTpgRuhJUmSOlBVawB3BK5lOOjtsCTnTRzjKODLwPkTj/tHVbUX8EzgMuB9DC1YVwFOBCwalpJFgyRJUh+OBg4AdmI4n+Ag4HETZ/h9kn0nHnNJmwE7Jzl3zr1rq2qPVoF6YNEgSZLUh9sAnwZenOSZjdby/7iqns7QBnYGIMmZE2f4qyUKBsYc3544R1csGiRJkvqwKvBi4OSq2hRYo0GG+42vWTPAoybOcGlVvRgIsBggyRcnztAdiwZJkqQ+vAx4EvAm4BkMBcSkkmw79Zg34tf8afEyA1g0LCO7J0mSJHViXJK0IcOm3zOTXDXRuEcneUpVXci4LGlWkvWnyPDnVNV6SS5smaEHFg2SJEkdqKp/ATYANgH+E9g+yc5tU02vqt4APJ9hudZtGIqnzdqmmv883E2SJKkPD0/yTODyJIcC92gdqJEdGYqnwxkKqF+0jdMHiwZJkqQ+3KqqVgdmqmpl4PrWgRq5MMnVwO2SnMUw46Bl5EZoSZKkPrwTOBm4E3DS+Hkh+nlV7QpcUVVvBu7QOE8X3NMgSZLUiaq6I3BP4Nwkl0w47rn86QboaxlOYb46ySZT5RizrATcFbgUeDbw5SSnT5mhR840SJIkzWNV9QGW6Fg03ifJrhPFuDewCHg3cGCS71TV/YG9Jhp/rr8EngisPn5+ImDRsIzc0yBJkjS/fQT4KLAmcAZwMPBDbnhovsUluXps77pRku+M904BaqoMc3yK4c/i6jkvLSNnGiRJkuaxJF8AqKqXJXnbePv4qvpSgzi/rao3At8BHgq0OB/h/CSvazBu1ywaJEmS+nDbqnoU8F2GB/bJZhrm+HtgT2AH4CfA6xpkOKaq3gKcNnsjyWENcnTF5UmSJEl92BV4KUPRsAfwrKkGrqoHjpcPA34EfIxhqdQjpsowx9OB2zOc0bAJw34LLSNnGiRJkjqQ5Iyq+mdgU4ZTkM+ZcPhHA98DljyBegb44oQ5YOjY9PyJx+yeLVclSZI6UFUvYnhoP4lhedKRSd4+cYaXAIcm+c2U4y6R4SDgXOD7jF2lkkxduHTHmQZJkqQ+7AxsneS6qloFOAGYtGgAVga+XFVnAO9L8vWJx4fhfIiNxxe0me3ojnsaJEmS+rAoyXUASa5lOGBtUknekWQL4N+BvarqzKnGrqrZL8P3WOK151QZeuZMgyRJUh+Oq6qjgW8BWwPHTx2gqm4N7MSwCXsR8NoJhz8M2AUINxx2t2i83nDCHF1yT4MkSVInquoJDB2DTkvyuQbj/xQ4Gjg4yVlTjz9meEaSD7UYu2cuT5IkSepAVT0ReNC4+flFVbVdgxibAB8C7l9V92swPsDzGo3bNZcnSZIk9eH1wLbj9dOAzzP9BuDnMywROgn4p6qavIMTsFpVncINy5RmkuwycYbuWDRIkiT14dokvwNI8ruqur5Bhl1o38Fpn4nHWxAsGiRJkvrwnao6Avg2sCVwSoMMf9LBqaom7+AEnA68iqHl6k+ANzXI0B2LBkmSpA4keWFVPQko4Ogkn24Qo3kHJ+Cj4+sQ4GHAfwE7NMjRFbsnSZIkabmZ08Hp9CSfbTD+15JsO+fz15M8cuocvXGmQZIkScusqnYHDkny2aq6DNisUZQzqurvga8BWwC/rqqNAZJMdthcb2y5KkmSpGVSVa8DtgNWHW+dD2xXVa9pEOfewG7A4cDewJrAgcABDbJ0w+VJkiRJHaiqxzCsIlkJeBfw6iRHTDT2ScBWSWbm3FsFOCHJllNk0C3LmQZJkqQ+vAn4KfAihg3Ae0449uVzCwYYuicBv58wg25BFg2SJEl9uBK4GLguyUUMB5tN5Q9VteHcG+Nnl7R0wo3QkiRJffg9cCxwUFX9I/DLCcfeB/hkVX0FOAe4G/A44FkTZvijqrozsPrs5yT/3SJHT9zTIEmS1IGqWg3YKMlpVXUf4Mwk10w4/u2BvwHWB84DPpNk8uVJVfUe4PHABcAiYCbJQ6fO0RtnGiRJkuaxqtotyfuB146f5/74lVPlSPI74LCpxrsJDwI2TLK4dZCeWDRIkiTNb+eP72maYsVxFsPSpCtbB+mJRYMkSdI8luQL4+XRwB2B64DnsWJ869/C3YDzquqs8bPLk5YDiwZJkqQ+HA28F3gKcBpwEMNm5IVm59YBemTLVUmSpD7cBjgG2CDJW4CVG+eZVFXtNl7uCeyxxEvLyJkGSZKkPqwKvBg4uao2BdZonGdqs3s7zmiaolO2XJUkSepAVT0UeBLDydD/AHwnyXeahlI3LBokSZI6UVVPADZlOKPhU63zqB8WDZIkSR2oqjcD9wKOA7YBzk3ysrap1Av3NEiSJPVhmyQPA6iq/YETG+eZVFW95s/9LMkbpszSI7snSZIk9WGVqpp9tlsELLTlJBePr4cA6wJnA2sC92uYqRvONEiSJPXho8DxVXUi8ODx84KR5ECAqtopyV7j7cOr6ksNY3XDmQZJkqQOJHkHw0nQxwO7J3ln40itrFlVGwFUVQG3b5ynC26EliRJ6kBV3ZXhNOTVZ+8txLX8VfVw4D3AOsDPgT2TfLdtqvnP5UmSJEl9OAr4MjcccrYgJTkOuG/rHL2xaJAkSerD75Ps2zpEK1V1dJKnVNWFLLEJPMn6jWJ1w+VJkiRJHaiqdwInAacwPjQnObNpqAaq6k5JftU6R2+caZAkSerD/fjT9qIzwKOaJGnr2Ko6G3hfEjsnLSfONEiSJKkrVfVA4DnAg4BPJnlT40jznjMNkiRJHaiqXwB3Bn4FrA1cxXDY2V4L8Bv3nwDfBu4JbN04Sxc8p0GSJKkP3wTuM2763QT4JPDXwBtbhppaVR0CnAxsDOyRZPvGkbpg0SBJktSHDZIEIMnZwN2SnAVc1zbW5D4BbJbkNUl+1jpML1yeJEmS1IcLq+otwAnAQ4GLquqxwDVtY03uPOA7VbUBcBHw3CTfb5xp3nOmQZIkqQ/PBC5gWJL038CzgcsZToleSPYHdkuyHsNm6Hc1ztMFZxokSZL6cB3wXeAHwCLgyUk+3DRRGyslORUgyQ+qaqEtz7pFWDRIkiT14RPAKsBdgJUZZh0WYtFwXVXtAHwL2Aa4unGeLrg8SZIkqQ9rj52CTgK2AFZvnKeVXYFnAccD/wA8r22cPlg0SJIk9eHK8X2NJH9gOBF6IbocOCjJfRja0P6ucZ4uWDRIkiT14eNV9Rrg1Ko6kYW7LOcjwGrj9aXAhxpm6caimZmFWoRKkiT1qao2B36a5KrWWaZWVSckeeicz19N8qiWmXrgRmhJkqR5rKr2TbJfVX2Y/7kkaZcWmRq7Zjyf4kTgQcDixnm6YNEgSZI0vx0zvh/QNMWKYzfg7cB/AKcBe7SN0weXJ0mSJM1zVXVf4CnA2sDPgaOS/LRtKvXEjdCSJEnzWFX9HXAIcD5wLPB74Oiq+pumwdQVlydJkiTNby8GHpHkitkbVXUo8KnxtaBU1aIkLqVZzpxpkCRJmt+um1swACS5DLi+UZ7WvtA6QI+caZAkSZrf/lx3oIX65fBvxqVZYfyzSXJm20jzn0WDJEnS/LZZVR2xxL1FwKYtwqwA7gy8ZM7nGcBzGpaR3ZMkSZLmsap6xJ/7WZJvTJllRVFVtwfuDpyd5PLGcbpg0SBJkqRuVNVOwL4MK2qOBGaS7Nc21fy3UNe6SZIkqU97A1sBlwD7AU9uG6cPFg2SJEnqyfVJrmaYYZgBrvjf/gX97ywaJEmS1JPjqurDwAZVdQDw3daBeuCeBkmSJHWlqrYHNgdOT/KZ1nl6YNEgSZKkblTV3YGnALeZvZfkDc0CdcLlSZIkSerJh4E1gIvnvLSMPNxNkiRJPbkyyetbh+iNy5MkSZI071XVxuPl64BjgO8znAZNkjMbxeqGMw2SJEnqwYFzrnefcz0DPGriLN2xaJAkSdK8l2RbgKraYW7HpKp6artU/XB5kiRJkua9qtoBeCiwC3DEeHsl4G+SbNIsWCecaZAkSVIPTgXWAv4AZLy3GPhIs0QdsWiQJElSDy5McmhVHQlc3zpMbywaJEmS1IPDGJYmncbYNQlYNF5v2CpUL9zTIEmSpO5U1VrApUl82F0OLBokSZLUjaraBngPsDJwFHBekoPbppr/VmodQJIkSVqO9gO2AS4C/gXYq22cPlg0SJIkqSeLk1wKzCS5Cvh960A9sGiQJElST86qqjcDa1XVK4DzWgfqgUWDJEmSerInQ6FwHHAF8Ly2cfrgRmhJkiRJN8mZBkmSJEk3yaJBkiRJ0k3yRGhJkiTNe1X1mj/3syRvmDJLj5xpkCRJUg8uHl8PAdYFzgbWBO7XMFM33AgtSZKkblTVF5NsN+fzl5I8tmWmHjjTIEmSpJ6sWVUbAVRVAbdvnKcL7mmQJElST14CfKKq1gF+znBug5aRy5MkSZIk3SRnGiRJktSNqnom8Apg9dl7STZsl6gPFg2SJEnqyT7AjsD5rYP0xKJBkiRJPTknyVmtQ/TGokGSJEk9ubKqPg/8AJgBSPLKpok6YNEgSZKknnyudYAeWTRIkiSpJ4cDWwKrAIuA9dvG6YNFgyRJknryCYaC4S7AysAFwIebJuqAJ0JLkiSpJ2sn2R44CdiCOa1XtfQsGiRJktSTK8f3NZL8gXEztJaNJ0JLkiSpG1X1j8CawLXAk4DLkzymaagOWDRIkiSpK1W1KMlMVW0OnDXOOGgZWDRIkiRJuknuaZAkSZJ0kywaJEmS1I2q2mGJz09tlaUnLk+SJEnSvDcWCw8DdgaOGG+vDOyYZJNmwTrh4W6SJEnqwanAWsAfgIz3FuPBbsuFMw2SJEnqRlXNLr9fCXgIcFKSaxpG6oIzDZIkSerJvwGnA38JPAC4GHhW00QdcCO0JEmSerJlkgOBhyTZHtigdaAeWDRIkiSpJytX1RbAz6pqVeB2rQP1wOVJkiRJ6slhwHuAXYG3AQe2jdMHN0JLkiSpK1V1e4Y9DWcnuaJ1nh5YNEiSJKkbVbUTsC/DipojgZkk+7VNNf+5p0GSJEk92RvYCrgE2A94cts4fbBokCRJUk+uT3I1wwzDDODypOXAokGSJEk9Oa6qjgA2qKoDgO+2DtQD9zRIkiSpK1W1PbA5cEaSY1rn6YFFgyRJkua9qloZWBn4CPA0YNH4+bNJHtUyWw88p0GSJEk92BV4JbAukPHeYuC4Zok64kyDJEmSulFVuyY5pHWO3lg0SJIkSbpJdk+SJEmSdJMsGiRJktSdqlqzdYaeuDxJkiRJ3aiqRwDvZuicdBRwXpKD26aa/5xpkCRJUk/eCGwDXAT8C7BX2zh9sGiQJElSTxYnuRSYSXIV8PvWgXpg0SBJkqSenFVVbwbWqqpXAOe1DtQDiwZJkiT1ZE+GQuE44ArgeW3j9METoSVJktSTo4D3AQcmsePPcuJMgyRJknqyH/B44JSqel1V3a11oB7YclWSJEndqao7Au8FnpxktdZ55juXJ0mSJKkbVbU18GxgS4alSi9vGqgTFg2SJEnqyUuAg4Dd3NOw/LinQZIkSfNeVT1wvDwImAEeW1XbVdV2DWN1w5kGSZIk9eDRwPeApy9xfwb44vRx+mLRIEmSpHkvyVvHy+OTvH/2flW9qFGkrtg9SZIkSfNeVe0M7AhsC3x1vL0SsHmSzZoF64QzDZIkSerBscCFwFrAgeO9xcDZzRJ1xKJBkiRJPVgjyder6oIl7t+2SZrOWDRIkiSpB3uPrwOXuD8DPGr6OH1xT4MkSZK6VFV3TXJ+6xw9sGiQJElSN6rqn4DfAncAngMcm2Tvlpl64OFukiRJ6slOwKHAXyfZFLh/4zxdsGiQJElST64H1gUuHj/fumGWbrgRWpIkST35+vh6RlW9E/hs0zSdcE+DJEmSulNVdwJ+m+Ta1ll6YNEgSZKkblTVI4FDgMsYNkM/L8mXWmbqgXsaJEmS1JP9gIcnuR/wsPGzlpFFgyRJknpyfZILAJL8AriqcZ4uuBFakiRJPbmsql4IfBPYBri0cZ4uONMgSZKknjwDuBvDsqS7Aru2jdMHiwZJkiR1I8nvgBOA44FvJvlN40hdsHuSJEmSulFV7wduC3ybYSP0L5K8tG2q+c89DZIkSerJ5kkePF7vX1UnNk3TCZcnSZIkqSdnVdU9AKrqzsB/N87TBZcnSZIkqRtVdTZwF4ZiYQPgaoa2qzNJ1m+ZbT6zaJAkSZJ0k1yeJEmSJOkmWTRIkiRJukkWDZIkSepGVe2wxOentsrSE/c0SJIkad4bi4WHATsDR4y3VwZ2TLJJs2Cd8JwGSZIk9eBUYC3gD0DGe4uBDzdL1BFnGiRJktSNqlopyeLWOXrjTIMkSZJ6sk9V7QNcCSzC8xmWC4sGSZIk9eTpwPpJrmwdpCd2T5IkSVJPzmXY16DlyJkGSZIk9WRV4EdV9SNgBiDJLm0jzX8WDZIkSerJW1sH6JHLkyRJktST7wOPBZ7F0IL1F23j9MGiQZIkST05BDgHuBdwEXBw2zh9sGiQJElST9ZKcghwbZIT8Hl3ufAPUZIkSV2pqnuP7xsA1zWO0wU3QkuSJKknLwI+AGwCHA3s1TZOH5xpkCRJUk9+CuyV5A7AW4AftY3TB4sGSZIk9eRw4H7j9cbAoe2i9MOiQZIkST25S5IPACR5G7Be4zxdsGiQJElST2aqamOAqtoIWLlxni64EVqSJEk9eSnw0apaB7gA2LNxni4smpmZaZ1BkiRJWm6q6vbA3YGzk1zeOE4XLBokSZLUjaraCdiXYUXNkcBMkv3appr/3NMgSZKknuwNbAVcAuwHPLltnD5YNEiSJKkn1ye5mmGGYQa4onWgHlg0SJIkqSfHVdURwAZVdQDw3daBeuCeBkmSJHWlqrYHNgfOSHJM6zw9sGiQJEnSvFdVtwJ2BH6T5GvjvXWB/ZM8rWm4DnhOgyRJknpwOHAdsF5VbQacCxwM7N80VScsGiRJktSDjZI8sKpWBU4Grga2TXJ641xdcCO0JEmSenAZQJJrGJ5xt7NgWH4sGiRJktSbi5Nc2jpET9wILUmSpHmvqi4GvgIsAh41XgOQZJdWuXrhngZJkiT14Klzrg9olqJTzjRIkiRJuknuaZAkSZJ0kywaJEmSJN0kiwZJkiRJN8miQZIkSdJN+v94RWeUcYNibAAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 504x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "produce_matrix(df_list = [tp_ae_hosp_discharge, tp_ae_cov, tp_ae_resp, tp_ae_suspected, tp_ae_probable, tp_ae_test, tp_ae_test_2_weeks, tp_ae_hos_resp_cov_plus_test_2_weeks, tp_ae_hosp_prim_care_before_plus_2_week_test], names=[\"Hospital discharge\", \"Covid diagnosis code\", \"Respiratory diagnosis code\", \"Probable covid code\", \"Suspected covid code\", \"Recent positive test (2 weeks)\", \"Recent positive test (4 weeks)\", \"Diagnosis codes (resp or cov) with recent positive test\", \"Covid in primary care with recent positive test\", \"Recent positive test and either covid in primary care or diagnosis code (resp or cov)\"], filename=\"cm_no_hosp_discharge_primary\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Restrict to A&E attendance with hospital discharge"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of A&E attendances with hospital discharge: 400\n"
     ]
    }
   ],
   "source": [
    "ae_with_hosp_discharge = round_to_5(len(df[\n",
    "    (df[\"ae_attendance_hosp_discharge_date\"].notnull())\n",
    "]))\n",
    "print(f'Number of A&E attendances with hospital discharge: {ae_with_hosp_discharge}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 1 - Discharge destination\n",
    "Restrict to those that attended A&E and had a hospital discharge destination (hospital ward, emergency short stay ward, icu)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with hospital discharge before admission: 500 (166.67%)\n"
     ]
    },
    {
     "ename": "ZeroDivisionError",
     "evalue": "division by zero",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/1777310896.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     39\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     40\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 41\u001b[0;31m \u001b[0mdischarge_destination\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_hosp_discharge\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_hosp_discharge\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_hosp_discharge\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_hosp_discharge\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     42\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf'False positives (FDR): {round_to_5(len(fp_ae_hosp_discharge))} ({discharge_destination[6]}%)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/3350781174.py\u001b[0m in \u001b[0;36mperformance_metrics\u001b[0;34m(tp, fp, tn, fn)\u001b[0m\n\u001b[1;32m     38\u001b[0m     \u001b[0msensitivity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     39\u001b[0m     \u001b[0mspecificity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 40\u001b[0;31m     \u001b[0mfdr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     41\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msensitivity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mspecificity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfdr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     42\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
     ]
    }
   ],
   "source": [
    "ae_hosp_before = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_date\"].notnull()\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with hospital discharge before admission: {round_to_5(len(ae_hosp_before))} ({round((round_to_5(len(ae_hosp_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_hosp_discharge = df[\n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_hosp_discharge = df[\n",
    "    ~(\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_hosp_discharge = df[\n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_hosp_discharge = df[\n",
    "    ~(df[\"discharge_destination\"].isin(hosp_discharge_list)) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "discharge_destination = performance_metrics(tp_ae_hosp_discharge, fp_ae_hosp_discharge, tn_ae_hosp_discharge, fn_ae_hosp_discharge)\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_ae_hosp_discharge))} ({discharge_destination[6]}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 2 - Diagnosis code\n",
    "Restrict to those with a COVID-19 diagnosis code or respiratory-related diagnosis code from the [emergency care diagnosis refset](https://termbrowser.nhs.uk/?perspective=full&conceptId1=991411000000109&edition=uk-edition&release=v20220216&server=https://termbrowser.nhs.uk/sct-browser-api/snomed&langRefset=900000000000508004)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance and covid diagnosis code before admission: 500 (166.67%)\n",
      "Number with A&E attendance and respiratory code before admission: 500 (166.67%)\n"
     ]
    },
    {
     "ename": "ZeroDivisionError",
     "evalue": "division by zero",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/2422805883.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     89\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     90\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 91\u001b[0;31m \u001b[0mdiagnosis_code_cov\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_cov\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_cov\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_cov\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_cov\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     92\u001b[0m \u001b[0mdiagnosis_code_resp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_resp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_resp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_resp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_resp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     93\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/3350781174.py\u001b[0m in \u001b[0;36mperformance_metrics\u001b[0;34m(tp, fp, tn, fn)\u001b[0m\n\u001b[1;32m     38\u001b[0m     \u001b[0msensitivity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     39\u001b[0m     \u001b[0mspecificity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 40\u001b[0;31m     \u001b[0mfdr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     41\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msensitivity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mspecificity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfdr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     42\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
     ]
    }
   ],
   "source": [
    "ae_cov_before = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_covid\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "ae_resp_before = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_resp\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance and covid diagnosis code before admission: {round_to_5(len(ae_cov_before))} ({round((round_to_5(len(ae_cov_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "print(f'Number with A&E attendance and respiratory code before admission: {round_to_5(len(ae_resp_before))} ({round((round_to_5(len(ae_resp_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_cov = df[\n",
    "    (df[\"ae_cov\"]==1) &\n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list)) &\n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_cov = df[\n",
    "    ~(\n",
    "        (df[\"ae_cov\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_cov = df[\n",
    "    (df[\"ae_cov\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_cov = df[\n",
    "    ~(\n",
    "        (df[\"ae_cov\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))\n",
    "    ) &\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "tp_ae_resp = df[\n",
    "    (df[\"ae_resp\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_resp = df[\n",
    "    ~(\n",
    "        (df[\"ae_resp\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_resp = df[\n",
    "    (df[\"ae_resp\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_resp = df[\n",
    "    ~(\n",
    "        (df[\"ae_resp\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "diagnosis_code_cov = performance_metrics(tp_ae_cov, fp_ae_cov, tn_ae_cov, fn_ae_cov)\n",
    "diagnosis_code_resp = performance_metrics(tp_ae_resp, fp_ae_resp, tn_ae_resp, fn_ae_resp)\n",
    "\n",
    "print(f'False positives (FDR) - A&E attendance and covid diagnosis code: {round_to_5(len(fp_ae_cov))} ({diagnosis_code_cov[6]}%)')\n",
    "print(f'False positives (FDR) - A&E attendance and respiratory diagnosis code: {round_to_5(len(fp_ae_resp))} ({diagnosis_code_resp[6]}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 3 - Identification in primary care\n",
    "\n",
    "Restrict to those with a probable or suspected COVID-19 code recorded in primary care in the 2 weeks prior to A&E attendance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance and recent suspected covid code in primary care before admission: 500 (166.67%)\n",
      "Number with A&E attendance and recent probable covid code in primary care before admission: 500 (166.67%)\n"
     ]
    },
    {
     "ename": "ZeroDivisionError",
     "evalue": "division by zero",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/3989611963.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     89\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     90\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 91\u001b[0;31m \u001b[0mprim_care_suspected\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_suspected\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_suspected\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_suspected\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_suspected\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     92\u001b[0m \u001b[0mprim_care_probable\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_probable\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_probable\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_probable\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_probable\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     93\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/3350781174.py\u001b[0m in \u001b[0;36mperformance_metrics\u001b[0;34m(tp, fp, tn, fn)\u001b[0m\n\u001b[1;32m     38\u001b[0m     \u001b[0msensitivity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     39\u001b[0m     \u001b[0mspecificity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 40\u001b[0;31m     \u001b[0mfdr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     41\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msensitivity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mspecificity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfdr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     42\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
     ]
    }
   ],
   "source": [
    "ae_prim_care_probable_before = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_prim_care_probable\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "ae_prim_care_suspected_before = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_prim_care_suspected\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance and recent suspected covid code in primary care before admission: {round_to_5(len(ae_prim_care_suspected_before))} ({round((round_to_5(len(ae_prim_care_suspected_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "print(f'Number with A&E attendance and recent probable covid code in primary care before admission: {round_to_5(len(ae_prim_care_probable_before))} ({round((round_to_5(len(ae_prim_care_probable_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_suspected = df[\n",
    "    (df[\"suspected_covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_suspected = df[\n",
    "    ~(\n",
    "        (df[\"suspected_covid_primary_care_before_ae\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_suspected = df[\n",
    "    (df[\"suspected_covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_suspected = df[\n",
    "    ~(\n",
    "        (df[\"suspected_covid_primary_care_before_ae\"]==1)& \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))\n",
    "        \n",
    "    )& \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "tp_ae_probable = df[\n",
    "    (df[\"covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_probable = df[\n",
    "    ~(\n",
    "        (df[\"covid_primary_care_before_ae\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_probable = df[\n",
    "    (df[\"covid_primary_care_before_ae\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "tn_ae_probable = df[\n",
    "    ~(\n",
    "        (df[\"covid_primary_care_before_ae\"]==1) &\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "prim_care_suspected = performance_metrics(tp_ae_suspected, fp_ae_suspected, tn_ae_suspected, fn_ae_suspected)\n",
    "prim_care_probable = performance_metrics(tp_ae_probable, fp_ae_probable, tn_ae_probable, fn_ae_probable)\n",
    "\n",
    "print(f'False positives (FDR) - A&E attendance and probable covid code in primary care: {round_to_5(len(fp_ae_probable))} ({prim_care_probable[6]}%)')\n",
    "print(f'False positives (FDR) - A&E attendance and suspected covid code in primary care: {round_to_5(len(fp_ae_suspected))} ({prim_care_suspected[6]}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Method 4 - Positive test result\n",
    "\n",
    "Restrict to those with a positive COVID-19 test result recorded up to 4 weeks prior to A&E attendance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance and recent positive test (2 weeks) before admission: 500 (166.67%)\n",
      "Number with A&E attendance and recent positive test (4 weeks) before admissio: 500 (166.67%)\n"
     ]
    },
    {
     "ename": "ZeroDivisionError",
     "evalue": "division by zero",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/4249659114.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     92\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     93\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 94\u001b[0;31m \u001b[0mpositive_test_2_weeks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_test_2_weeks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_test_2_weeks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_test_2_weeks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_test_2_weeks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     95\u001b[0m \u001b[0mpositive_test_4_weeks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     96\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_35459/3350781174.py\u001b[0m in \u001b[0;36mperformance_metrics\u001b[0;34m(tp, fp, tn, fn)\u001b[0m\n\u001b[1;32m     38\u001b[0m     \u001b[0msensitivity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     39\u001b[0m     \u001b[0mspecificity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 40\u001b[0;31m     \u001b[0mfdr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     41\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msensitivity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mspecificity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfdr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     42\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
     ]
    }
   ],
   "source": [
    "ae_pos_test_before_2_weeks = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_positive_covid_test_2_weeks\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "ae_pos_test_before = (df.loc[\n",
    "    df[\"ae_before_with_hospital_discharge_positive_covid_test\"]==1\n",
    ", \"patient_id\"])\n",
    "\n",
    "\n",
    "print(f'Number with A&E attendance and recent positive test (2 weeks) before admission: {round_to_5(len(ae_pos_test_before_2_weeks))} ({round((round_to_5(len(ae_pos_test_before_2_weeks))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "print(f'Number with A&E attendance and recent positive test (4 weeks) before admissio: {round_to_5(len(ae_pos_test_before))} ({round((round_to_5(len(ae_pos_test_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_test = df[\n",
    "    (df[\"pos_test\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_test = df[\n",
    "    ~(\n",
    "        (df[\"pos_test\"]==1) &\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_test = df[\n",
    "    (df[\"pos_test\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_test = df[\n",
    "    ~(\n",
    "        (df[\"pos_test\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tp_ae_test_2_weeks = df[\n",
    "    (df[\"pos_test_2_weeks\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_test_2_weeks = df[\n",
    "    ~(\n",
    "        (df[\"pos_test_2_weeks\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_test_2_weeks = df[\n",
    "    (df[\"pos_test_2_weeks\"]==1) & \n",
    "    (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_test_2_weeks = df[\n",
    "    ~(\n",
    "        (df[\"pos_test_2_weeks\"]==1) & \n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "\n",
    "positive_test_2_weeks = performance_metrics(tp_ae_test_2_weeks, fp_ae_test_2_weeks, tn_ae_test_2_weeks, fn_ae_test_2_weeks)\n",
    "positive_test_4_weeks = performance_metrics(tp_ae_test, fp_ae_test, tn_ae_test, fn_ae_test)\n",
    "\n",
    "print(f'False positives (FDR) - A&E attendance and recent positive test (2 weeks): {round_to_5(len(fp_ae_test_2_weeks))} ({positive_test_2_weeks[6]}%)')\n",
    "print(f'False positives (FDR) - A&E attendance and recent positive test (4 weeks): {round_to_5(len(fp_ae_test))} ({positive_test_4_weeks[6]}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "####  Combining approaches"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Diagnosis codes (resp or cov) with recent positive test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with either covid or respiratory diagnosis code and recent positive test (2 weeks) before admission: 375 (46.88%)\n"
     ]
    },
    {
     "ename": "ZeroDivisionError",
     "evalue": "division by zero",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_34360/3552651922.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     70\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     71\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 72\u001b[0;31m \u001b[0mcode_plus_test\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_hos_resp_cov_plus_test_2_weeks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_hos_resp_cov_plus_test_2_weeks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_hos_resp_cov_plus_test_2_weeks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_hos_resp_cov_plus_test_2_weeks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     73\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf'False positives (FDR): {fp_ae_hos_resp_cov_plus_test_2_weeks} ({code_plus_test[6]}%)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_34360/2412537391.py\u001b[0m in \u001b[0;36mperformance_metrics\u001b[0;34m(tp, fp, tn, fn)\u001b[0m\n\u001b[1;32m     37\u001b[0m     \u001b[0msensitivity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     38\u001b[0m     \u001b[0mspecificity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 39\u001b[0;31m     \u001b[0mfdr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     40\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msensitivity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mspecificity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfdr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     41\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
     ]
    }
   ],
   "source": [
    "ae_hos_resp_cov_plus_test_2_weeks = (df.loc[(\n",
    "        (df[\"ae_before_with_hospital_discharge_covid\"]==1) | \n",
    "        (df[\"ae_before_with_hospital_discharge_resp\"]==1) \n",
    "    )&\n",
    "    (df[\"ae_before_with_hospital_discharge_positive_covid_test_2_weeks\"]==1)\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with either covid or respiratory diagnosis code and recent positive test (2 weeks) before admission: {round_to_5(len(ae_hos_resp_cov_plus_test_2_weeks))} ({round((round_to_5(len(ae_hos_resp_cov_plus_test_2_weeks))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)    \n",
    "        )&\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    ~(\n",
    "        (\n",
    "            (\n",
    "                (df[\"ae_cov\"]==1) |\n",
    "                (df[\"ae_resp\"]==1)    \n",
    "            )&\n",
    "            (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "            (df[\"pos_test_2_weeks\"]==1)\n",
    "        ) & \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)    \n",
    "        )&\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_hos_resp_cov_plus_test_2_weeks = df[\n",
    "    ~((\n",
    "        (\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)    \n",
    "        )&\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "code_plus_test = performance_metrics(tp_ae_hos_resp_cov_plus_test_2_weeks, fp_ae_hos_resp_cov_plus_test_2_weeks, tn_ae_hos_resp_cov_plus_test_2_weeks, fn_ae_hos_resp_cov_plus_test_2_weeks)\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_ae_hos_resp_cov_plus_test_2_weeks))} ({code_plus_test[6]}%)')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Covid in primary care with recent positive test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with suspected or probable covid in primary care and recent positive test (2 weeks) before admission: 375 (46.88%)\n"
     ]
    },
    {
     "ename": "ZeroDivisionError",
     "evalue": "division by zero",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_34360/807346584.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     73\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     74\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 75\u001b[0;31m \u001b[0mprimary_care_plus_test\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_ae_hosp_prim_care_before_plus_2_week_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_ae_hosp_prim_care_before_plus_2_week_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_ae_hosp_prim_care_before_plus_2_week_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_ae_hosp_prim_care_before_plus_2_week_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     76\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf'False positives (FDR): {fp_ae_hosp_prim_care_before_plus_2_week_test} ({primary_care_plus_test[6]}%)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_34360/2412537391.py\u001b[0m in \u001b[0;36mperformance_metrics\u001b[0;34m(tp, fp, tn, fn)\u001b[0m\n\u001b[1;32m     37\u001b[0m     \u001b[0msensitivity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     38\u001b[0m     \u001b[0mspecificity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 39\u001b[0;31m     \u001b[0mfdr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     40\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msensitivity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mspecificity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfdr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     41\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
     ]
    }
   ],
   "source": [
    "ae_hosp_prim_care_before_plus_2_week_test = (df.loc[(\n",
    "        (df[\"ae_before_with_hospital_discharge_prim_care_suspected\"]==1) | \n",
    "        (df[\"ae_before_with_hospital_discharge_prim_care_probable\"]==1) \n",
    "    )\n",
    "&\n",
    "    (df[\"ae_before_with_hospital_discharge_positive_covid_test_2_weeks\"]==1)\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with suspected or probable covid in primary care and recent positive test (2 weeks) before admission: {round_to_5(len(ae_hosp_prim_care_before_plus_2_week_test))} ({round((round_to_5(len(ae_hosp_prim_care_before_plus_2_week_test))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "        )&\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    ~(\n",
    "        (\n",
    "            (\n",
    "                (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "                (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "            )&\n",
    "            (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "            (df[\"pos_test_2_weeks\"]==1)\n",
    "        )& \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    ((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "        )&\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )\n",
    "    ) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_ae_hosp_prim_care_before_plus_2_week_test = df[\n",
    "    ~((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1)    \n",
    "        )&\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "\n",
    "primary_care_plus_test = performance_metrics(tp_ae_hosp_prim_care_before_plus_2_week_test, fp_ae_hosp_prim_care_before_plus_2_week_test, tn_ae_hosp_prim_care_before_plus_2_week_test, fn_ae_hosp_prim_care_before_plus_2_week_test)\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_ae_hosp_prim_care_before_plus_2_week_test))} ({primary_care_plus_test[6]}%)')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Recent positive test and either covid in primary care or diagnosis code (resp or cov)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number with A&E attendance with recent positive test and either suspected or probable covid in primary care or a covid or respiratory diagnosis code on attendance: 475 (59.38%)\n"
     ]
    },
    {
     "ename": "ZeroDivisionError",
     "evalue": "division by zero",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_34360/539280281.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     90\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     91\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 92\u001b[0;31m \u001b[0mpos_test_prim_care_or_diagnosis_code\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mperformance_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp_combined_before\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp_combined_before\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn_combined_before\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn_combined_before\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     93\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf'False positives (FDR): {fp_combined_before} ({pos_test_prim_care_or_diagnosis_code[6]}%)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_34360/2412537391.py\u001b[0m in \u001b[0;36mperformance_metrics\u001b[0;34m(tp, fp, tn, fn)\u001b[0m\n\u001b[1;32m     37\u001b[0m     \u001b[0msensitivity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     38\u001b[0m     \u001b[0mspecificity\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtn\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 39\u001b[0;31m     \u001b[0mfdr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfp\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     40\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msensitivity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mspecificity\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfdr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     41\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
     ]
    }
   ],
   "source": [
    "combined_before = (df.loc[\n",
    "        (df[\"ae_before_with_hospital_discharge_positive_covid_test_2_weeks\"]==1) &\n",
    "        (\n",
    "            (df[\"ae_before_with_hospital_discharge_prim_care_suspected\"]==1) | \n",
    "            (df[\"ae_before_with_hospital_discharge_prim_care_probable\"]==1) |\n",
    "            (df[\"ae_before_with_hospital_discharge_covid\"]==1) | \n",
    "            (df[\"ae_before_with_hospital_discharge_resp\"]==1)   \n",
    "        )\n",
    ", \"patient_id\"])\n",
    "\n",
    "print(f'Number with A&E attendance with recent positive test and either suspected or probable covid in primary care or a covid or respiratory diagnosis code on attendance: {round_to_5(len(combined_before))} ({round((round_to_5(len(combined_before))/num_emergency_cov_admissions)*100, 2)}%)')\n",
    "\n",
    "tp_combined_before = df[\n",
    "    (\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    ) & \n",
    "    (\n",
    "        (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "        (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "    \n",
    "]\n",
    "\n",
    "fn_combined_before = df[\n",
    "    ~(\n",
    "        (\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "        )& \n",
    "        (\n",
    "            (df[\"ae_attendance_date\"] <= df[\"emergency_covid_hospital_admission_date\"]) &\n",
    "            (df[\"ae_attendance_date\"] > (df[\"emergency_covid_hospital_admission_date\"] - timedelta(days = 14)))\n",
    "        )\n",
    "    )&\n",
    "    (df[\"emergency_covid_hospital_admission_date\"].notnull())\n",
    "]\n",
    "\n",
    "\n",
    "fp_combined_before = df[\n",
    "    ((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )\n",
    "    ) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "tn_combined_before = df[\n",
    "    ~((\n",
    "        (\n",
    "            (df[\"covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"suspected_covid_primary_care_before_ae\"]==1) |\n",
    "            (df[\"ae_cov\"]==1) |\n",
    "            (df[\"ae_resp\"]==1)\n",
    "        \n",
    "        )\n",
    "        &\n",
    "        (df[\"discharge_destination\"].isin(hosp_discharge_list))&\n",
    "        (df[\"pos_test_2_weeks\"]==1)\n",
    "    )) & \n",
    "    (df[\"emergency_covid_hospital_admission_date\"].isnull())\n",
    "]\n",
    "\n",
    "\n",
    "pos_test_prim_care_or_diagnosis_code = performance_metrics(tp_combined_before, fp_combined_before, tn_combined_before, fn_combined_before)\n",
    "print(f'False positives (FDR): {round_to_5(len(fp_combined_before))} ({pos_test_prim_care_or_diagnosis_code[6]}%)')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'discharge_destination' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m/var/folders/6d/gn8xz2l52x32r5r2m_yxl_vr0000gp/T/ipykernel_34360/3823603461.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0midentification_table_index\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m\"tp\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"fp\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"tn\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"fn\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"sensitivity\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"specificity\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"fdr\"\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m identification_table = pd.DataFrame({\n\u001b[0;32m----> 3\u001b[0;31m     \u001b[0;34m\"discharge_destination\"\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mSeries\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdischarge_destination\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0midentification_table_index\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      4\u001b[0m     \u001b[0;34m\"covid_diagnosis_code\"\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mSeries\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdiagnosis_code_cov\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0midentification_table_index\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m     \u001b[0;34m\"resp_diagnosis_code\"\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mSeries\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdiagnosis_code_resp\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0midentification_table_index\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'discharge_destination' is not defined"
     ]
    }
   ],
   "source": [
    "identification_table_index = [\"tp\", \"fp\", \"tn\", \"fn\",\"sensitivity\", \"specificity\", \"fdr\"]\n",
    "identification_table = pd.DataFrame({\n",
    "    \"discharge_destination\": pd.Series(discharge_destination,index=identification_table_index),\n",
    "    \"covid_diagnosis_code\": pd.Series(diagnosis_code_cov,index=identification_table_index),\n",
    "    \"resp_diagnosis_code\": pd.Series(diagnosis_code_resp,index=identification_table_index),\n",
    "    \"primary_care_suspected\": pd.Series(prim_care_suspected,index=identification_table_index),\n",
    "    \"primary_care_probable\": pd.Series(prim_care_probable,index=identification_table_index),\n",
    "    \"positive_test_2_weeks\": pd.Series(positive_test_2_weeks,index=identification_table_index),\n",
    "    \"positive_test_4_weeks\": pd.Series(positive_test_4_weeks,index=identification_table_index),\n",
    "    \"diagnosis_code_with_positive_test\": pd.Series(code_plus_test,index=identification_table_index),\n",
    "    \"primary_care_with_positive_test\": pd.Series(primary_care_plus_test,index=identification_table_index),\n",
    "    \"pos_test_prim_care_or_diagnosis\": pd.Series(pos_test_prim_care_or_diagnosis_code,index=identification_table_index)\n",
    "    \n",
    "})\n",
    "identification_table.to_csv(\"output/jun/identification_table_hosp_discharge_primary.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 504x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "produce_matrix(df_list = [tp_ae_hosp_discharge, tp_ae_cov, tp_ae_resp, tp_ae_suspected, tp_ae_probable, tp_ae_test, tp_ae_test_2_weeks, tp_ae_hos_resp_cov_plus_test_2_weeks, tp_ae_hosp_prim_care_before_plus_2_week_test], names=[\"Hospital discharge\", \"Covid diagnosis code\", \"Respiratory diagnosis code\", \"Probable covid code\", \"Suspected covid code\", \"Recent positive test (2 weeks)\", \"Recent positive test (4 weeks)\", \"Diagnosis codes (resp or cov) with recent positive test\", \"Covid in primary care with recent positive test\", \"Recent positive test and either covid in primary care or diagnosis code (resp or cov)\"], filename=\"cm_hosp_discharge_primary\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.9.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}