{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# What are items are generating the most OOPE?\n",
    "\n",
    "* The most expensive single presentation is Cinacalcet HCl_Tab 30mg, costing £16,000 per month in OOPE\n",
    "  * Calcium and Magnesium supplements overall cost £25,000 per month in OOPE with a mean OOPE of £3.30\n",
    "* The most expensive single OOPE is for Tretinoin Cap 10mg with OOPE cost of £80.99\n",
    "* The most expensive commonly-prescribed presentation is Solgar_Mag Cit Tab 200mg with mean OOPE per item of £9.86\n",
    "* At a BNF paragraph level, the most expensive things are:\n",
    "   * Vitamin D (£67,000 per month)\n",
    "   * Foods for Special Diets (£52,000 per month)\n",
    "   * Enteral Nutrition (£43,000 per month)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A [recent case of a pharmacist being suspended](https://www.pharmacyregulation.org/sites/default/files/determinations/tasleem_fatma_2063889_determination.pdf) gives some background to how OOPE overcharging might happen. In this case (2013)it was carelessness and inexperience:\n",
    "\n",
    " * It affected Aciferol D3, Amlostin, Glutafin G/F bread, Mepradec capsules, ProD3 4, Simvador tablets\n",
    " * These were branded generics, prescribed as brand by GP on grounds they were cheaper (10%) that DT price\n",
    " * But there were not available from the usual wholesaler\n",
    " * They were therefore put through as \"specials\" to their alternative wholesaler (Pharmacare)\n",
    " * Each order to Pharmacare incurs a £50 plus VAT handling fee and a £15 plus VAT administration fee\n",
    "\n",
    "It was confirmed [on Twitter](https://twitter.com/CPPEOli/status/1082224683167244289) that the £36.98 amount mentioned below is the standard \"special obtains\" fee charged by Alliance; which they will always charge for items they don't usually hold.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "sql = \"\"\"\n",
    "SELECT\n",
    "  SUM(item_count) AS items,\n",
    "  SUM(item_pay_oope_amt) AS oope,\n",
    "  bnf_code,\n",
    "  bnf_name\n",
    "FROM\n",
    "  `ebmdatalab.dispensers.dispensing_with_metadata`\n",
    "WHERE year_month = '201703' AND item_pay_oope_amt > 0\n",
    "GROUP BY bnf_code, bnf_name\n",
    "ORDER BY SUM(item_pay_oope_amt) DESC\n",
    "LIMIT 1000\n",
    "\"\"\"\n",
    "\n",
    "items = pd.io.gbq.read_gbq(sql, 'ebmdatalab', dialect='standard')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>items</th>\n",
       "      <th>oope</th>\n",
       "      <th>bnf_code</th>\n",
       "      <th>bnf_name</th>\n",
       "      <th>oope_per_item</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>778</td>\n",
       "      <td>16130.23</td>\n",
       "      <td>0905012I0AAAAAA</td>\n",
       "      <td>Cinacalcet HCl_Tab 30mg</td>\n",
       "      <td>20.732943</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>572</td>\n",
       "      <td>12346.06</td>\n",
       "      <td>0606020Z0AAAAAA</td>\n",
       "      <td>Denosumab_Inj 60mg/1ml Pfs</td>\n",
       "      <td>21.584021</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>477</td>\n",
       "      <td>10347.45</td>\n",
       "      <td>130802000BBATA0</td>\n",
       "      <td>Dermacolor_Camouflage Creme</td>\n",
       "      <td>21.692767</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   items      oope         bnf_code                     bnf_name  \\\n",
       "0    778  16130.23  0905012I0AAAAAA      Cinacalcet HCl_Tab 30mg   \n",
       "1    572  12346.06  0606020Z0AAAAAA   Denosumab_Inj 60mg/1ml Pfs   \n",
       "2    477  10347.45  130802000BBATA0  Dermacolor_Camouflage Creme   \n",
       "\n",
       "   oope_per_item  \n",
       "0      20.732943  \n",
       "1      21.584021  \n",
       "2      21.692767  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "items['oope_per_item'] = items['oope'] / items['items']\n",
    "items.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>items</th>\n",
       "      <th>oope</th>\n",
       "      <th>bnf_code</th>\n",
       "      <th>bnf_name</th>\n",
       "      <th>oope_per_item</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>101</td>\n",
       "      <td>2584.71</td>\n",
       "      <td>090402000BBRMA0</td>\n",
       "      <td>Fresubin 2kcal_Drink (6 Flav)</td>\n",
       "      <td>25.591188</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>147</td>\n",
       "      <td>3581.08</td>\n",
       "      <td>090402000BBSIA0</td>\n",
       "      <td>Fortisip Compact_Liq (8 Flav)</td>\n",
       "      <td>24.361088</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>156</td>\n",
       "      <td>3795.75</td>\n",
       "      <td>0905012I0AAABAB</td>\n",
       "      <td>Cinacalcet HCl_Tab 60mg</td>\n",
       "      <td>24.331731</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    items     oope         bnf_code                       bnf_name  \\\n",
       "21    101  2584.71  090402000BBRMA0  Fresubin 2kcal_Drink (6 Flav)   \n",
       "15    147  3581.08  090402000BBSIA0  Fortisip Compact_Liq (8 Flav)   \n",
       "13    156  3795.75  0905012I0AAABAB        Cinacalcet HCl_Tab 60mg   \n",
       "\n",
       "    oope_per_item  \n",
       "21      25.591188  \n",
       "15      24.361088  \n",
       "13      24.331731  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "items[items['items'] > 100].sort_values('oope_per_item', ascending=False).head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">items</th>\n",
       "      <th colspan=\"2\" halign=\"left\">oope</th>\n",
       "      <th colspan=\"2\" halign=\"left\">oope_per_item</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>mean</th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "      <th>sum</th>\n",
       "      <th>mean</th>\n",
       "      <th>sum</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bnf_para</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>090604</th>\n",
       "      <td>59.426966</td>\n",
       "      <td>5289</td>\n",
       "      <td>750.558539</td>\n",
       "      <td>66799.71</td>\n",
       "      <td>20.154450</td>\n",
       "      <td>1793.746029</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>090401</th>\n",
       "      <td>9.676056</td>\n",
       "      <td>2061</td>\n",
       "      <td>246.200610</td>\n",
       "      <td>52440.73</td>\n",
       "      <td>29.515000</td>\n",
       "      <td>6286.695034</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>090402</th>\n",
       "      <td>26.000000</td>\n",
       "      <td>2392</td>\n",
       "      <td>464.477391</td>\n",
       "      <td>42731.92</td>\n",
       "      <td>37.859968</td>\n",
       "      <td>3483.117087</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>130802</th>\n",
       "      <td>93.250000</td>\n",
       "      <td>1119</td>\n",
       "      <td>2107.610000</td>\n",
       "      <td>25291.32</td>\n",
       "      <td>27.762150</td>\n",
       "      <td>333.145797</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>090501</th>\n",
       "      <td>77.000000</td>\n",
       "      <td>1155</td>\n",
       "      <td>1657.254667</td>\n",
       "      <td>24858.82</td>\n",
       "      <td>23.271590</td>\n",
       "      <td>349.073845</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>060602</th>\n",
       "      <td>165.750000</td>\n",
       "      <td>663</td>\n",
       "      <td>3767.602500</td>\n",
       "      <td>15070.41</td>\n",
       "      <td>28.796925</td>\n",
       "      <td>115.187701</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>040202</th>\n",
       "      <td>32.500000</td>\n",
       "      <td>390</td>\n",
       "      <td>877.387500</td>\n",
       "      <td>10528.65</td>\n",
       "      <td>30.282186</td>\n",
       "      <td>363.386238</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>091000</th>\n",
       "      <td>86.555556</td>\n",
       "      <td>779</td>\n",
       "      <td>1008.082222</td>\n",
       "      <td>9072.74</td>\n",
       "      <td>15.460353</td>\n",
       "      <td>139.143175</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>090102</th>\n",
       "      <td>176.500000</td>\n",
       "      <td>706</td>\n",
       "      <td>2134.987500</td>\n",
       "      <td>8539.95</td>\n",
       "      <td>17.571041</td>\n",
       "      <td>70.284164</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>190201</th>\n",
       "      <td>1133.000000</td>\n",
       "      <td>1133</td>\n",
       "      <td>7753.190000</td>\n",
       "      <td>7753.19</td>\n",
       "      <td>6.843063</td>\n",
       "      <td>6.843063</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                items               oope           oope_per_item             \n",
       "                 mean   sum         mean       sum          mean          sum\n",
       "bnf_para                                                                     \n",
       "090604      59.426966  5289   750.558539  66799.71     20.154450  1793.746029\n",
       "090401       9.676056  2061   246.200610  52440.73     29.515000  6286.695034\n",
       "090402      26.000000  2392   464.477391  42731.92     37.859968  3483.117087\n",
       "130802      93.250000  1119  2107.610000  25291.32     27.762150   333.145797\n",
       "090501      77.000000  1155  1657.254667  24858.82     23.271590   349.073845\n",
       "060602     165.750000   663  3767.602500  15070.41     28.796925   115.187701\n",
       "040202      32.500000   390   877.387500  10528.65     30.282186   363.386238\n",
       "091000      86.555556   779  1008.082222   9072.74     15.460353   139.143175\n",
       "090102     176.500000   706  2134.987500   8539.95     17.571041    70.284164\n",
       "190201    1133.000000  1133  7753.190000   7753.19      6.843063     6.843063"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Grouped at BNF paragraph level\n",
    "items['bnf_para'] = items['bnf_code'].str.slice(0,6)\n",
    "items.groupby('bnf_para').agg(['mean', 'sum']).sort_values(('oope', 'sum'), ascending=False).head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# How does it change over time?\n",
    "\n",
    "In the period for which we have data, OOPE has decreased overall.  This is because the number of items having OOPE added is decreasing; however, the amount of OOPE being added per item is increasing. See charts below."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "sql = \"\"\"SELECT \n",
    "  PARSE_DATE(\"%Y%m_%d\", CONCAT(year_month, \"_01\")) AS date,\n",
    "  SUM(item_pay_oope_amt) AS oope,\n",
    "  SUM(CASE WHEN item_pay_oope_amt > 0 THEN item_count ELSE 0 END) AS oope_items,\n",
    "  SUM(item_count) AS items\n",
    "FROM\n",
    "  `ebmdatalab.dispensers.dispensing_with_metadata`\n",
    "GROUP BY\n",
    "  year_month\n",
    "ORDER BY\n",
    "  year_month\n",
    "\"\"\"\n",
    "dftime = pd.io.gbq.read_gbq(sql, 'ebmdatalab', dialect='standard')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "lines_to_next_cell": 2
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7ff46b2b0c18>"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "dftime['oope_per_item'] = dftime['oope'] / dftime['oope_items']\n",
    "dftime.set_index('date')[['oope']].plot()\n",
    "dftime.set_index('date')[['oope_per_item']].plot()\n",
    "dftime.set_index('date')[['oope_items']].plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Look up by CCG\n",
    "# Add a column in bigquery that matches postcode to CCG\n",
    "\n",
    "\n",
    "sql = \"\"\"SELECT \n",
    "  PARSE_DATE(\"%Y%m_%d\", CONCAT(year_month, \"_01\")) AS date,\n",
    "  SUM(item_pay_oope_amt) AS oope,\n",
    "  SUM(CASE WHEN item_pay_oope_amt > 0 THEN item_count ELSE 0 END) AS oope_items,\n",
    "  SUM(item_count) AS items\n",
    "FROM\n",
    "  `ebmdatalab.dispensers.dispensing_with_metadata`\n",
    "  WHERE ccg = '08H'\n",
    "GROUP BY\n",
    "  year_month\n",
    "ORDER BY\n",
    "  year_month\n",
    "\"\"\"\n",
    "dftime = pd.io.gbq.read_gbq(sql, 'ebmdatalab', dialect='standard')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f6d94e81cc0>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "dftime['oope_per_item'] = dftime['oope'] / dftime['oope_items']\n",
    "dftime.set_index('date')[['oope']].plot()\n",
    "dftime.set_index('date')[['oope_per_item']].plot()\n",
    "dftime.set_index('date')[['oope_items']].plot()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Is there interesting variation for multiples?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# summarise OOPE spending per item, grouped by HQ\n",
    "sql = \"\"\"\n",
    "SELECT\n",
    "  hq_name,\n",
    "  COUNT(DISTINCT organisation_code) AS branches,\n",
    "  SUM(item_count) AS items,\n",
    "  SUM(actual_cost) AS actual_cost,\n",
    "  SUM(CASE WHEN item_pay_oope_amt > 0 THEN item_count ELSE 0 END) AS oope_items,\n",
    "  SUM(item_pay_oope_amt) AS oope,\n",
    "  MAX(item_pay_oope_amt/item_count) AS max_oope_per_item\n",
    "FROM\n",
    "  `ebmdatalab.dispensers.dispensing_with_metadata`\n",
    "WHERE year_month = '201703'\n",
    "GROUP BY\n",
    "  hq_name\n",
    "\"\"\"\n",
    "df = pd.io.gbq.read_gbq(sql, 'ebmdatalab', dialect='standard')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In a single month, there were 11816 branches, represented by 3762 HQs, prescribing 84970248 items\n",
      "27224 items had OOPE added\n",
      "Total OOPE per month £406966.0\n"
     ]
    }
   ],
   "source": [
    "print(\"In a single month, there were {} branches, represented by {} HQs, prescribing {} items\".format(\n",
    "    df['branches'].sum(),\n",
    "    df['branches'].count(),\n",
    "    df['items'].sum()))\n",
    "print(\"{} items had OOPE added\".format(df['oope_items'].sum()))\n",
    "print(\"Total OOPE per month £%s ({})\" % (round(df['oope']).sum()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## HQs with highest OOPE per item\n",
    "\n",
    "If we only count HQs with at least 100 OOPE items per month, per branch, then the top-spending (per OOPE item) HQ charges an average of £57 per OOPE item, and 6% of its items have OOPE added.  It charges £10,869 of OOPE in one month.\n",
    "\n",
    "Of the top 10 HQs, only one was a multiple (i.e. associated with more than contractor/location).\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['oope_per_item'] = df['oope'] / df['items']\n",
    "df['oope_per_oope_item'] = df['oope'] / df['oope_items']\n",
    "df['percent_oope_items'] = round(df['oope_items'] / df['items'] * 100)\n",
    "df['percent_oope_cost'] = round(df['oope'] / df['actual_cost'] * 100)\n",
    "df = df.sort_values('oope_per_oope_item', ascending=False)\n",
    "common_oopers = df[df['oope_items']/df['branches'] > 100]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>hq_name</th>\n",
       "      <th>branches</th>\n",
       "      <th>items</th>\n",
       "      <th>actual_cost</th>\n",
       "      <th>oope_items</th>\n",
       "      <th>oope</th>\n",
       "      <th>max_oope_per_item</th>\n",
       "      <th>oope_per_item</th>\n",
       "      <th>oope_per_oope_item</th>\n",
       "      <th>percent_oope_items</th>\n",
       "      <th>percent_oope_cost</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>924</th>\n",
       "      <td>ZAIMRA LTD</td>\n",
       "      <td>1</td>\n",
       "      <td>5196</td>\n",
       "      <td>56741.52</td>\n",
       "      <td>190</td>\n",
       "      <td>10869.00</td>\n",
       "      <td>79.00</td>\n",
       "      <td>2.091801</td>\n",
       "      <td>57.205263</td>\n",
       "      <td>4.0</td>\n",
       "      <td>19.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2817</th>\n",
       "      <td>C.K &amp; SONS PHARMACY</td>\n",
       "      <td>1</td>\n",
       "      <td>5221</td>\n",
       "      <td>35233.32</td>\n",
       "      <td>135</td>\n",
       "      <td>4079.00</td>\n",
       "      <td>79.00</td>\n",
       "      <td>0.781268</td>\n",
       "      <td>30.214815</td>\n",
       "      <td>3.0</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1570</th>\n",
       "      <td>MEDICX A G LIMITED</td>\n",
       "      <td>1</td>\n",
       "      <td>5459</td>\n",
       "      <td>54335.33</td>\n",
       "      <td>328</td>\n",
       "      <td>9167.59</td>\n",
       "      <td>73.68</td>\n",
       "      <td>1.679353</td>\n",
       "      <td>27.949970</td>\n",
       "      <td>6.0</td>\n",
       "      <td>17.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2724</th>\n",
       "      <td>MEDI-CALL HEALTHCARE LTD</td>\n",
       "      <td>1</td>\n",
       "      <td>2032</td>\n",
       "      <td>23935.24</td>\n",
       "      <td>125</td>\n",
       "      <td>2923.00</td>\n",
       "      <td>79.00</td>\n",
       "      <td>1.438484</td>\n",
       "      <td>23.384000</td>\n",
       "      <td>6.0</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1193</th>\n",
       "      <td>JM PATEL LTD</td>\n",
       "      <td>1</td>\n",
       "      <td>9196</td>\n",
       "      <td>101498.46</td>\n",
       "      <td>539</td>\n",
       "      <td>11686.98</td>\n",
       "      <td>79.00</td>\n",
       "      <td>1.270876</td>\n",
       "      <td>21.682709</td>\n",
       "      <td>6.0</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2495</th>\n",
       "      <td>SPRING PHARMACY</td>\n",
       "      <td>1</td>\n",
       "      <td>8877</td>\n",
       "      <td>80412.59</td>\n",
       "      <td>115</td>\n",
       "      <td>2480.00</td>\n",
       "      <td>50.00</td>\n",
       "      <td>0.279374</td>\n",
       "      <td>21.565217</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2964</th>\n",
       "      <td>S&amp;S CHEMISTS</td>\n",
       "      <td>1</td>\n",
       "      <td>15374</td>\n",
       "      <td>91065.44</td>\n",
       "      <td>154</td>\n",
       "      <td>3129.90</td>\n",
       "      <td>61.99</td>\n",
       "      <td>0.203584</td>\n",
       "      <td>20.324026</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2304</th>\n",
       "      <td>MEDIX-AG LTD</td>\n",
       "      <td>1</td>\n",
       "      <td>8242</td>\n",
       "      <td>46293.35</td>\n",
       "      <td>143</td>\n",
       "      <td>2135.59</td>\n",
       "      <td>90.39</td>\n",
       "      <td>0.259111</td>\n",
       "      <td>14.934196</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3020</th>\n",
       "      <td>BESPOKE PHARMACY LTD</td>\n",
       "      <td>3</td>\n",
       "      <td>55939</td>\n",
       "      <td>340462.72</td>\n",
       "      <td>1640</td>\n",
       "      <td>19031.46</td>\n",
       "      <td>86.99</td>\n",
       "      <td>0.340218</td>\n",
       "      <td>11.604549</td>\n",
       "      <td>3.0</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>592</th>\n",
       "      <td>HUSCHEM LTD</td>\n",
       "      <td>1</td>\n",
       "      <td>8631</td>\n",
       "      <td>57097.70</td>\n",
       "      <td>103</td>\n",
       "      <td>1146.90</td>\n",
       "      <td>36.98</td>\n",
       "      <td>0.132881</td>\n",
       "      <td>11.134951</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1825</th>\n",
       "      <td>CAMBRIDGE AVENUE MEDICAL SERVICES LTD</td>\n",
       "      <td>1</td>\n",
       "      <td>20735</td>\n",
       "      <td>125761.31</td>\n",
       "      <td>166</td>\n",
       "      <td>1675.30</td>\n",
       "      <td>42.50</td>\n",
       "      <td>0.080796</td>\n",
       "      <td>10.092169</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2416</th>\n",
       "      <td>BEACON PRIMARY HEALTHCARE LTD</td>\n",
       "      <td>1</td>\n",
       "      <td>18978</td>\n",
       "      <td>123550.69</td>\n",
       "      <td>492</td>\n",
       "      <td>3550.08</td>\n",
       "      <td>36.98</td>\n",
       "      <td>0.187063</td>\n",
       "      <td>7.215610</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2520</th>\n",
       "      <td>SIGCARE LIMITED</td>\n",
       "      <td>1</td>\n",
       "      <td>5128</td>\n",
       "      <td>32661.79</td>\n",
       "      <td>123</td>\n",
       "      <td>320.00</td>\n",
       "      <td>30.00</td>\n",
       "      <td>0.062402</td>\n",
       "      <td>2.601626</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                    hq_name  branches  items  actual_cost  \\\n",
       "924                              ZAIMRA LTD         1   5196     56741.52   \n",
       "2817                    C.K & SONS PHARMACY         1   5221     35233.32   \n",
       "1570                     MEDICX A G LIMITED         1   5459     54335.33   \n",
       "2724               MEDI-CALL HEALTHCARE LTD         1   2032     23935.24   \n",
       "1193                           JM PATEL LTD         1   9196    101498.46   \n",
       "2495                        SPRING PHARMACY         1   8877     80412.59   \n",
       "2964                           S&S CHEMISTS         1  15374     91065.44   \n",
       "2304                           MEDIX-AG LTD         1   8242     46293.35   \n",
       "3020                   BESPOKE PHARMACY LTD         3  55939    340462.72   \n",
       "592                             HUSCHEM LTD         1   8631     57097.70   \n",
       "1825  CAMBRIDGE AVENUE MEDICAL SERVICES LTD         1  20735    125761.31   \n",
       "2416          BEACON PRIMARY HEALTHCARE LTD         1  18978    123550.69   \n",
       "2520                        SIGCARE LIMITED         1   5128     32661.79   \n",
       "\n",
       "      oope_items      oope  max_oope_per_item  oope_per_item  \\\n",
       "924          190  10869.00              79.00       2.091801   \n",
       "2817         135   4079.00              79.00       0.781268   \n",
       "1570         328   9167.59              73.68       1.679353   \n",
       "2724         125   2923.00              79.00       1.438484   \n",
       "1193         539  11686.98              79.00       1.270876   \n",
       "2495         115   2480.00              50.00       0.279374   \n",
       "2964         154   3129.90              61.99       0.203584   \n",
       "2304         143   2135.59              90.39       0.259111   \n",
       "3020        1640  19031.46              86.99       0.340218   \n",
       "592          103   1146.90              36.98       0.132881   \n",
       "1825         166   1675.30              42.50       0.080796   \n",
       "2416         492   3550.08              36.98       0.187063   \n",
       "2520         123    320.00              30.00       0.062402   \n",
       "\n",
       "      oope_per_oope_item  percent_oope_items  percent_oope_cost  \n",
       "924            57.205263                 4.0               19.0  \n",
       "2817           30.214815                 3.0               12.0  \n",
       "1570           27.949970                 6.0               17.0  \n",
       "2724           23.384000                 6.0               12.0  \n",
       "1193           21.682709                 6.0               12.0  \n",
       "2495           21.565217                 1.0                3.0  \n",
       "2964           20.324026                 1.0                3.0  \n",
       "2304           14.934196                 2.0                5.0  \n",
       "3020           11.604549                 3.0                6.0  \n",
       "592            11.134951                 1.0                2.0  \n",
       "1825           10.092169                 1.0                1.0  \n",
       "2416            7.215610                 3.0                3.0  \n",
       "2520            2.601626                 2.0                1.0  "
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "common_oopers.head(13)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>hq_name</th>\n",
       "      <th>branches</th>\n",
       "      <th>items</th>\n",
       "      <th>actual_cost</th>\n",
       "      <th>oope_items</th>\n",
       "      <th>oope</th>\n",
       "      <th>max_oope_per_item</th>\n",
       "      <th>oope_per_item</th>\n",
       "      <th>oope_per_oope_item</th>\n",
       "      <th>percent_oope_items</th>\n",
       "      <th>percent_oope_cost</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>924</th>\n",
       "      <td>ZAIMRA LTD</td>\n",
       "      <td>1</td>\n",
       "      <td>5196</td>\n",
       "      <td>56741.52</td>\n",
       "      <td>190</td>\n",
       "      <td>10869.00</td>\n",
       "      <td>79.00</td>\n",
       "      <td>2.091801</td>\n",
       "      <td>57.205263</td>\n",
       "      <td>4.0</td>\n",
       "      <td>19.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1570</th>\n",
       "      <td>MEDICX A G LIMITED</td>\n",
       "      <td>1</td>\n",
       "      <td>5459</td>\n",
       "      <td>54335.33</td>\n",
       "      <td>328</td>\n",
       "      <td>9167.59</td>\n",
       "      <td>73.68</td>\n",
       "      <td>1.679353</td>\n",
       "      <td>27.949970</td>\n",
       "      <td>6.0</td>\n",
       "      <td>17.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2724</th>\n",
       "      <td>MEDI-CALL HEALTHCARE LTD</td>\n",
       "      <td>1</td>\n",
       "      <td>2032</td>\n",
       "      <td>23935.24</td>\n",
       "      <td>125</td>\n",
       "      <td>2923.00</td>\n",
       "      <td>79.00</td>\n",
       "      <td>1.438484</td>\n",
       "      <td>23.384000</td>\n",
       "      <td>6.0</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1193</th>\n",
       "      <td>JM PATEL LTD</td>\n",
       "      <td>1</td>\n",
       "      <td>9196</td>\n",
       "      <td>101498.46</td>\n",
       "      <td>539</td>\n",
       "      <td>11686.98</td>\n",
       "      <td>79.00</td>\n",
       "      <td>1.270876</td>\n",
       "      <td>21.682709</td>\n",
       "      <td>6.0</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2817</th>\n",
       "      <td>C.K &amp; SONS PHARMACY</td>\n",
       "      <td>1</td>\n",
       "      <td>5221</td>\n",
       "      <td>35233.32</td>\n",
       "      <td>135</td>\n",
       "      <td>4079.00</td>\n",
       "      <td>79.00</td>\n",
       "      <td>0.781268</td>\n",
       "      <td>30.214815</td>\n",
       "      <td>3.0</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       hq_name  branches  items  actual_cost  oope_items  \\\n",
       "924                 ZAIMRA LTD         1   5196     56741.52         190   \n",
       "1570        MEDICX A G LIMITED         1   5459     54335.33         328   \n",
       "2724  MEDI-CALL HEALTHCARE LTD         1   2032     23935.24         125   \n",
       "1193              JM PATEL LTD         1   9196    101498.46         539   \n",
       "2817       C.K & SONS PHARMACY         1   5221     35233.32         135   \n",
       "\n",
       "          oope  max_oope_per_item  oope_per_item  oope_per_oope_item  \\\n",
       "924   10869.00              79.00       2.091801           57.205263   \n",
       "1570   9167.59              73.68       1.679353           27.949970   \n",
       "2724   2923.00              79.00       1.438484           23.384000   \n",
       "1193  11686.98              79.00       1.270876           21.682709   \n",
       "2817   4079.00              79.00       0.781268           30.214815   \n",
       "\n",
       "      percent_oope_items  percent_oope_cost  \n",
       "924                  4.0               19.0  \n",
       "1570                 6.0               17.0  \n",
       "2724                 6.0               12.0  \n",
       "1193                 6.0               12.0  \n",
       "2817                 3.0               12.0  "
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "common_oopers.sort_values('percent_oope_cost', ascending=False).head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Distribution of OOPE per item by HQ\n",
    "The  majority of HQs have no OOPE at all. All but six have their mean OOPE across all items of under 50p; 87% had a mean OOPE of less than 1p.  In the first chart, note the log scale!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "lines_to_next_cell": 2
   },
   "outputs": [],
   "source": [
    "print(\"{}% of contracts never charge any OOPE\".format(round(df[df['oope'] == 0].count().iloc[0] / df.count().iloc[0] * 100)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>oope_per_oope_item</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>(-0.207, 1.02]</th>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>(1.02, 2.034]</th>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>(2.034, 3.048]</th>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>(3.048, 4.063]</th>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>(4.063, 5.077]</th>\n",
       "      <td>37</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    count\n",
       "oope_per_oope_item       \n",
       "(-0.207, 1.02]         12\n",
       "(1.02, 2.034]          17\n",
       "(2.034, 3.048]         30\n",
       "(3.048, 4.063]         21\n",
       "(4.063, 5.077]         37"
      ]
     },
     "execution_count": 207,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bins = pd.cut(df.oope_per_oope_item, 209)\n",
    "df.groupby(bins)['oope_per_oope_item'].agg(['count']).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'Mean OOPE per item (£)')"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "fig, ax = plt.subplots()\n",
    "df.hist(ax=ax, column=['oope_per_item'], bins=50)\n",
    "ax.set_yscale('log')\n",
    "ax.set_ylabel('HQ count (log scale)')\n",
    "ax.set_xlabel('Mean OOPE per item (£)')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'Mean OOPE per item that has OOPE (£)')"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "fig, ax = plt.subplots()\n",
    "df.hist(ax=ax, column=['oope_per_oope_item'], bins=50)\n",
    "ax.set_ylabel('HQ count')\n",
    "ax.set_xlabel('Mean OOPE per item that has OOPE (£)')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## How does high OOPE covary with group size?\n",
    "\n",
    "None of the large multiples has high mean OOPEs:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f7cc11012b0>"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.plot.scatter(x='branches', y='oope_per_oope_item')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Is there anything different about the prescriptions being dispensed?\n",
    "\n",
    "That is, do the HQs with the highest mean OOPE prescribe different amounts of kinds of things from those with the lowest?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The 10 HQs with highest mean OOPE account for 109981 items\n",
      "The 10 HQs with lowest mean OOPE account for 77398 items\n"
     ]
    }
   ],
   "source": [
    "print(\"The 10 HQs with highest mean OOPE account for %s items\" % df['items'].head(10).sum())\n",
    "print(\"The 10 HQs with lowest mean OOPE account for %s items\" % df['items'].tail(10).sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Investigate lowest and highest HQs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [],
   "source": [
    "sql = \"\"\"\n",
    "SELECT\n",
    "  item_pay_dr_nic,\n",
    "  item_count,\n",
    "  item_pay_oope_amt,\n",
    "  bnf_code,\n",
    "  bnf_name,\n",
    "  name,\n",
    "  postcode\n",
    "FROM\n",
    "  `ebmdatalab.dispensers.dispensing_with_metadata`\n",
    "WHERE year_month = '201703' AND \n",
    "(%s)\n",
    "\"\"\"\n",
    "hq_names = []\n",
    "for name in list(common_oopers.hq_name.head(10)):\n",
    "    hq_names.append(\"hq_name = '%s'\" % name)\n",
    "highest = pd.io.gbq.read_gbq(sql % \" OR \".join(hq_names), 'ebmdatalab', dialect='standard')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [],
   "source": [
    "sql = \"\"\"\n",
    "SELECT\n",
    "  item_pay_dr_nic,\n",
    "  item_count,\n",
    "  item_pay_oope_amt,\n",
    "  bnf_code,\n",
    "  bnf_name,\n",
    "  name,\n",
    "  postcode\n",
    "FROM\n",
    "  `ebmdatalab.dispensers.dispensing_with_metadata`\n",
    "WHERE year_month = '201703' AND \n",
    "(%s)\n",
    "\"\"\"\n",
    "hq_names = []\n",
    "for name in list(df.hq_name.tail(10)):\n",
    "    hq_names.append(\"hq_name = '%s'\" % name)\n",
    "lowest = pd.io.gbq.read_gbq(sql % \" OR \".join(hq_names), 'ebmdatalab', dialect='standard')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The 10 HQs with highest OOPE per items\n",
    "...are accounted for by 12 dispensers, with a mean item count of 10300, mean NIC of £73,000, and mean OOPE of £5550. 3.5% of prescription items come with OOPEs (compared with 0 for the best dispensers). In a single month they charged £69,000 in OOPE, equivalent to a 9% surcharge (the highest being 22%). 67% of dispensers have NO OOPE."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>item_pay_dr_nic</th>\n",
       "      <th>item_count</th>\n",
       "      <th>item_pay_oope_amt</th>\n",
       "      <th>uplift</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>STONE PHARMACY</th>\n",
       "      <td>96587.59</td>\n",
       "      <td>9196</td>\n",
       "      <td>11686.98</td>\n",
       "      <td>0.120999</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RYANS PHARMACY</th>\n",
       "      <td>49271.80</td>\n",
       "      <td>5196</td>\n",
       "      <td>10869.00</td>\n",
       "      <td>0.220593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BIRKWOOD PHARMACY</th>\n",
       "      <td>106179.76</td>\n",
       "      <td>19749</td>\n",
       "      <td>10341.15</td>\n",
       "      <td>0.097393</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WILBERFORCE PHARMACY ONLINE</th>\n",
       "      <td>48620.37</td>\n",
       "      <td>5459</td>\n",
       "      <td>9167.59</td>\n",
       "      <td>0.188555</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MARISCO PHARMACY</th>\n",
       "      <td>158056.87</td>\n",
       "      <td>20486</td>\n",
       "      <td>6286.61</td>\n",
       "      <td>0.039774</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C K &amp; SONS</th>\n",
       "      <td>33487.40</td>\n",
       "      <td>5221</td>\n",
       "      <td>4079.00</td>\n",
       "      <td>0.121807</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>S&amp;S CHEMISTS</th>\n",
       "      <td>93973.56</td>\n",
       "      <td>15374</td>\n",
       "      <td>3129.90</td>\n",
       "      <td>0.033306</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RYANS LOCAL</th>\n",
       "      <td>22613.05</td>\n",
       "      <td>2032</td>\n",
       "      <td>2923.00</td>\n",
       "      <td>0.129262</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SPRING PHARMACY LTD</th>\n",
       "      <td>83883.55</td>\n",
       "      <td>8877</td>\n",
       "      <td>2480.00</td>\n",
       "      <td>0.029565</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NEWMARKET PHARMACY</th>\n",
       "      <td>80359.16</td>\n",
       "      <td>15704</td>\n",
       "      <td>2403.70</td>\n",
       "      <td>0.029912</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WILBERFORCE PHARMACY</th>\n",
       "      <td>47292.67</td>\n",
       "      <td>8242</td>\n",
       "      <td>2135.59</td>\n",
       "      <td>0.045157</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MEDICARE PHARMACY</th>\n",
       "      <td>60112.96</td>\n",
       "      <td>8631</td>\n",
       "      <td>1146.90</td>\n",
       "      <td>0.019079</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                             item_pay_dr_nic  item_count  item_pay_oope_amt  \\\n",
       "name                                                                          \n",
       "STONE PHARMACY                      96587.59        9196           11686.98   \n",
       "RYANS PHARMACY                      49271.80        5196           10869.00   \n",
       "BIRKWOOD PHARMACY                  106179.76       19749           10341.15   \n",
       "WILBERFORCE PHARMACY ONLINE         48620.37        5459            9167.59   \n",
       "MARISCO PHARMACY                   158056.87       20486            6286.61   \n",
       "C K & SONS                          33487.40        5221            4079.00   \n",
       "S&S CHEMISTS                        93973.56       15374            3129.90   \n",
       "RYANS LOCAL                         22613.05        2032            2923.00   \n",
       "SPRING PHARMACY LTD                 83883.55        8877            2480.00   \n",
       "NEWMARKET PHARMACY                  80359.16       15704            2403.70   \n",
       "WILBERFORCE PHARMACY                47292.67        8242            2135.59   \n",
       "MEDICARE PHARMACY                   60112.96        8631            1146.90   \n",
       "\n",
       "                               uplift  \n",
       "name                                   \n",
       "STONE PHARMACY               0.120999  \n",
       "RYANS PHARMACY               0.220593  \n",
       "BIRKWOOD PHARMACY            0.097393  \n",
       "WILBERFORCE PHARMACY ONLINE  0.188555  \n",
       "MARISCO PHARMACY             0.039774  \n",
       "C K & SONS                   0.121807  \n",
       "S&S CHEMISTS                 0.033306  \n",
       "RYANS LOCAL                  0.129262  \n",
       "SPRING PHARMACY LTD          0.029565  \n",
       "NEWMARKET PHARMACY           0.029912  \n",
       "WILBERFORCE PHARMACY         0.045157  \n",
       "MEDICARE PHARMACY            0.019079  "
      ]
     },
     "execution_count": 127,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "totals = highest.groupby(\"name\").sum().sort_values('item_pay_oope_amt', ascending=False)\n",
    "totals['uplift'] = totals['item_pay_oope_amt']/totals['item_pay_dr_nic']\n",
    "totals"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "item_pay_dr_nic      73369.895000\n",
       "item_count           10347.250000\n",
       "item_pay_oope_amt     5554.118333\n",
       "uplift                   0.089617\n",
       "dtype: float64"
      ]
     },
     "execution_count": 130,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "totals.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The 10  HQs with lowest OOPE per item \n",
    "...are accounted for by 10 dispensers, with a mean item count of 7739, mean NIC of £58,000, and mean OOPE of £0 (indeed, a maximum OOPE of £0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>item_pay_dr_nic</th>\n",
       "      <th>item_count</th>\n",
       "      <th>item_pay_oope_amt</th>\n",
       "      <th>bnf_code</th>\n",
       "      <th>bnf_name</th>\n",
       "      <th>name</th>\n",
       "      <th>postcode</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5.24</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0304010G0BCACAB</td>\n",
       "      <td>Piriton_Syr 2mg/5ml</td>\n",
       "      <td>GOSPEL LANE PHARMACY</td>\n",
       "      <td>B27 7AJ</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   item_pay_dr_nic  item_count  item_pay_oope_amt         bnf_code  \\\n",
       "0             5.24           2                0.0  0304010G0BCACAB   \n",
       "\n",
       "              bnf_name                  name postcode  \n",
       "0  Piriton_Syr 2mg/5ml  GOSPEL LANE PHARMACY  B27 7AJ  "
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lowest.head(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>item_pay_dr_nic</th>\n",
       "      <th>item_count</th>\n",
       "      <th>item_pay_oope_amt</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>GIBSONS PHARMACY</th>\n",
       "      <td>86639.07</td>\n",
       "      <td>12992</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GOSPEL LANE PHARMACY</th>\n",
       "      <td>84491.16</td>\n",
       "      <td>9294</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HALTON PHARMACY</th>\n",
       "      <td>28602.85</td>\n",
       "      <td>4615</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HEATH PHARMACY</th>\n",
       "      <td>25253.77</td>\n",
       "      <td>4261</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MEDIBANK PHARMACY</th>\n",
       "      <td>24870.48</td>\n",
       "      <td>4609</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MILFORD PHARMACY</th>\n",
       "      <td>44619.98</td>\n",
       "      <td>5486</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MONKBAR PHARMACY</th>\n",
       "      <td>83188.79</td>\n",
       "      <td>7807</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PEAK PHARMACY</th>\n",
       "      <td>22131.30</td>\n",
       "      <td>3576</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>VERWOOD PHARMACY</th>\n",
       "      <td>80653.13</td>\n",
       "      <td>11322</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WHITTINGTON PHARMACY</th>\n",
       "      <td>25230.83</td>\n",
       "      <td>3604</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      item_pay_dr_nic  item_count  item_pay_oope_amt\n",
       "name                                                                \n",
       "GIBSONS PHARMACY             86639.07       12992                0.0\n",
       "GOSPEL LANE PHARMACY         84491.16        9294                0.0\n",
       "HALTON PHARMACY              28602.85        4615                0.0\n",
       "HEATH PHARMACY               25253.77        4261                0.0\n",
       "MEDIBANK PHARMACY            24870.48        4609                0.0\n",
       "MILFORD PHARMACY             44619.98        5486                0.0\n",
       "MONKBAR PHARMACY             83188.79        7807                0.0\n",
       "PEAK PHARMACY                22131.30        3576                0.0\n",
       "VERWOOD PHARMACY             80653.13       11322                0.0\n",
       "WHITTINGTON PHARMACY         25230.83        3604                0.0"
      ]
     },
     "execution_count": 132,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lowest.groupby(\"name\").sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "item_pay_dr_nic      50568.136\n",
       "item_count            6756.600\n",
       "item_pay_oope_amt        0.000\n",
       "dtype: float64"
      ]
     },
     "execution_count": 133,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lowest.groupby(\"name\").sum().mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Do the highest and lowest dispense very different things?\n",
    "\n",
    "Descriptive statistics for the \"highest\" and \"lowest\" groups are very similar.\n",
    "\n",
    "The mean OOPE per item (where OOPE was added at all) in the \"highest\" group was £36 (and 0 in the \"lowest\"). The \"highest\" group had about 120 presentations-with-OOPE not seen in the \"lowest\" group; for those presentations, the mean OOPE per item was £41; OOPE presentations also dispensed in the \"lowest\" group had a mean OOPE per item of £26.\n",
    "\n",
    "Therefore it looks like the \"highest\" group do routinely add more OOPE for everything; but a lot more for things only they see.\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>item_pay_dr_nic</th>\n",
       "      <th>item_count</th>\n",
       "      <th>item_pay_oope_amt</th>\n",
       "      <th>oope_count</th>\n",
       "      <th>oope_per_oope_item</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bnf_name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Coloplast_Sensura Mio Maxi DrnblBag Grey+Window C/Fit10-55mm</th>\n",
       "      <td>298.89</td>\n",
       "      <td>1</td>\n",
       "      <td>86.99</td>\n",
       "      <td>1</td>\n",
       "      <td>86.99</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Promixin_Pdr For Neb 1mega u</th>\n",
       "      <td>672.00</td>\n",
       "      <td>2</td>\n",
       "      <td>158.38</td>\n",
       "      <td>2</td>\n",
       "      <td>79.19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Nutrison_Energy M/Fibre Liq (Vanilla)</th>\n",
       "      <td>459.90</td>\n",
       "      <td>1</td>\n",
       "      <td>79.00</td>\n",
       "      <td>1</td>\n",
       "      <td>79.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Loprofin_L/P Fusilli</th>\n",
       "      <td>8.94</td>\n",
       "      <td>1</td>\n",
       "      <td>79.00</td>\n",
       "      <td>1</td>\n",
       "      <td>79.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Symbicort_Turbohaler 100mcg/6mcg (120 D)</th>\n",
       "      <td>396.00</td>\n",
       "      <td>11</td>\n",
       "      <td>79.00</td>\n",
       "      <td>1</td>\n",
       "      <td>79.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                    item_pay_dr_nic  \\\n",
       "bnf_name                                                              \n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window...           298.89   \n",
       "Promixin_Pdr For Neb 1mega u                                 672.00   \n",
       "Nutrison_Energy M/Fibre Liq (Vanilla)                        459.90   \n",
       "Loprofin_L/P Fusilli                                           8.94   \n",
       "Symbicort_Turbohaler 100mcg/6mcg (120 D)                     396.00   \n",
       "\n",
       "                                                    item_count  \\\n",
       "bnf_name                                                         \n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window...           1   \n",
       "Promixin_Pdr For Neb 1mega u                                 2   \n",
       "Nutrison_Energy M/Fibre Liq (Vanilla)                        1   \n",
       "Loprofin_L/P Fusilli                                         1   \n",
       "Symbicort_Turbohaler 100mcg/6mcg (120 D)                    11   \n",
       "\n",
       "                                                    item_pay_oope_amt  \\\n",
       "bnf_name                                                                \n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window...              86.99   \n",
       "Promixin_Pdr For Neb 1mega u                                   158.38   \n",
       "Nutrison_Energy M/Fibre Liq (Vanilla)                           79.00   \n",
       "Loprofin_L/P Fusilli                                            79.00   \n",
       "Symbicort_Turbohaler 100mcg/6mcg (120 D)                        79.00   \n",
       "\n",
       "                                                    oope_count  \\\n",
       "bnf_name                                                         \n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window...           1   \n",
       "Promixin_Pdr For Neb 1mega u                                 2   \n",
       "Nutrison_Energy M/Fibre Liq (Vanilla)                        1   \n",
       "Loprofin_L/P Fusilli                                         1   \n",
       "Symbicort_Turbohaler 100mcg/6mcg (120 D)                     1   \n",
       "\n",
       "                                                    oope_per_oope_item  \n",
       "bnf_name                                                                \n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window...               86.99  \n",
       "Promixin_Pdr For Neb 1mega u                                     79.19  \n",
       "Nutrison_Energy M/Fibre Liq (Vanilla)                            79.00  \n",
       "Loprofin_L/P Fusilli                                             79.00  \n",
       "Symbicort_Turbohaler 100mcg/6mcg (120 D)                         79.00  "
      ]
     },
     "execution_count": 186,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "highest['oope_count'] = np.where(highest['item_pay_oope_amt'] > 0, highest['item_count'], 0)\n",
    "df3 = highest.groupby('bnf_name').sum()\n",
    "df3['oope_per_oope_item'] = df3['item_pay_oope_amt'] / df3['oope_count']\n",
    "highest_presentations = df3.sort_values('oope_per_oope_item', ascending=False)\n",
    "highest_presentations[~np.isnan(highest_presentations['oope_per_oope_item'])].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The following shows that in the HQs with the highest OOPE per items, 3746 different presentations were dispensed, of which 286 had OOPE, with a mean OOPE per item of £36"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>item_pay_dr_nic</th>\n",
       "      <th>item_count</th>\n",
       "      <th>item_pay_oope_amt</th>\n",
       "      <th>oope_count</th>\n",
       "      <th>oope_per_oope_item</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>3885.000000</td>\n",
       "      <td>3885.000000</td>\n",
       "      <td>3885.000000</td>\n",
       "      <td>3885.000000</td>\n",
       "      <td>286.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>226.625158</td>\n",
       "      <td>31.960618</td>\n",
       "      <td>17.155578</td>\n",
       "      <td>0.893694</td>\n",
       "      <td>36.552648</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>696.446448</td>\n",
       "      <td>142.008236</td>\n",
       "      <td>139.804966</td>\n",
       "      <td>10.403610</td>\n",
       "      <td>19.573289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.140000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.152181</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>16.460000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>25.283333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>51.630000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>36.980000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>174.520000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>39.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>14929.820000</td>\n",
       "      <td>3385.000000</td>\n",
       "      <td>3476.120000</td>\n",
       "      <td>486.000000</td>\n",
       "      <td>86.990000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       item_pay_dr_nic   item_count  item_pay_oope_amt   oope_count  \\\n",
       "count      3885.000000  3885.000000        3885.000000  3885.000000   \n",
       "mean        226.625158    31.960618          17.155578     0.893694   \n",
       "std         696.446448   142.008236         139.804966    10.403610   \n",
       "min           0.140000     1.000000           0.000000     0.000000   \n",
       "25%          16.460000     1.000000           0.000000     0.000000   \n",
       "50%          51.630000     4.000000           0.000000     0.000000   \n",
       "75%         174.520000    13.000000           0.000000     0.000000   \n",
       "max       14929.820000  3385.000000        3476.120000   486.000000   \n",
       "\n",
       "       oope_per_oope_item  \n",
       "count          286.000000  \n",
       "mean            36.552648  \n",
       "std             19.573289  \n",
       "min              0.152181  \n",
       "25%             25.283333  \n",
       "50%             36.980000  \n",
       "75%             39.500000  \n",
       "max             86.990000  "
      ]
     },
     "execution_count": 144,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "highest_presentations.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And the next table shows in the HQs with the lowest OOPE per items, 3368 items were dispensed with no OOPE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>item_pay_dr_nic</th>\n",
       "      <th>item_count</th>\n",
       "      <th>item_pay_oope_amt</th>\n",
       "      <th>oope_count</th>\n",
       "      <th>oope_per_oope_item</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>3222.000000</td>\n",
       "      <td>3222.000000</td>\n",
       "      <td>3222.0</td>\n",
       "      <td>3222.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>156.946418</td>\n",
       "      <td>20.970205</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>396.685053</td>\n",
       "      <td>81.933010</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.190000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>14.980000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>44.930000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>138.450000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>7144.200000</td>\n",
       "      <td>1715.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       item_pay_dr_nic   item_count  item_pay_oope_amt  oope_count  \\\n",
       "count      3222.000000  3222.000000             3222.0      3222.0   \n",
       "mean        156.946418    20.970205                0.0         0.0   \n",
       "std         396.685053    81.933010                0.0         0.0   \n",
       "min           0.190000     1.000000                0.0         0.0   \n",
       "25%          14.980000     1.000000                0.0         0.0   \n",
       "50%          44.930000     3.000000                0.0         0.0   \n",
       "75%         138.450000    11.000000                0.0         0.0   \n",
       "max        7144.200000  1715.000000                0.0         0.0   \n",
       "\n",
       "       oope_per_oope_item  \n",
       "count                 0.0  \n",
       "mean                  NaN  \n",
       "std                   NaN  \n",
       "min                   NaN  \n",
       "25%                   NaN  \n",
       "50%                   NaN  \n",
       "75%                   NaN  \n",
       "max                   NaN  "
      ]
     },
     "execution_count": 155,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df3 = lowest.groupby('bnf_name').sum()\n",
    "lowest['oope_count'] = np.where(lowest['item_pay_oope_amt'] > 0, lowest['item_count'], 0)\n",
    "df3 = lowest.groupby('bnf_name').sum()\n",
    "df3['oope_per_oope_item'] = df3['item_pay_oope_amt'] / df3['oope_count']\n",
    "lowest_presentations = df3.sort_values('oope_per_oope_item', ascending=False)\n",
    "lowest_presentations.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can combine the two tables together, to find presentations common to both high and low OOPE dispensers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>item_pay_dr_nic_high</th>\n",
       "      <th>item_count_high</th>\n",
       "      <th>item_pay_oope_amt_high</th>\n",
       "      <th>oope_count_high</th>\n",
       "      <th>oope_per_oope_item_high</th>\n",
       "      <th>item_pay_dr_nic_low</th>\n",
       "      <th>item_count_low</th>\n",
       "      <th>item_pay_oope_amt_low</th>\n",
       "      <th>oope_count_low</th>\n",
       "      <th>oope_per_oope_item_low</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bnf_name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Coloplast_Sensura Mio Maxi DrnblBag Grey+Window C/Fit10-55mm</th>\n",
       "      <td>298.89</td>\n",
       "      <td>1.0</td>\n",
       "      <td>86.99</td>\n",
       "      <td>1.0</td>\n",
       "      <td>86.99</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Promixin_Pdr For Neb 1mega u</th>\n",
       "      <td>672.00</td>\n",
       "      <td>2.0</td>\n",
       "      <td>158.38</td>\n",
       "      <td>2.0</td>\n",
       "      <td>79.19</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Nutrison_Energy M/Fibre Liq (Vanilla)</th>\n",
       "      <td>459.90</td>\n",
       "      <td>1.0</td>\n",
       "      <td>79.00</td>\n",
       "      <td>1.0</td>\n",
       "      <td>79.00</td>\n",
       "      <td>61.32</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Loprofin_L/P Fusilli</th>\n",
       "      <td>8.94</td>\n",
       "      <td>1.0</td>\n",
       "      <td>79.00</td>\n",
       "      <td>1.0</td>\n",
       "      <td>79.00</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Symbicort_Turbohaler 100mcg/6mcg (120 D)</th>\n",
       "      <td>396.00</td>\n",
       "      <td>11.0</td>\n",
       "      <td>79.00</td>\n",
       "      <td>1.0</td>\n",
       "      <td>79.00</td>\n",
       "      <td>561.00</td>\n",
       "      <td>13.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                    item_pay_dr_nic_high  \\\n",
       "bnf_name                                                                   \n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window...                298.89   \n",
       "Promixin_Pdr For Neb 1mega u                                      672.00   \n",
       "Nutrison_Energy M/Fibre Liq (Vanilla)                             459.90   \n",
       "Loprofin_L/P Fusilli                                                8.94   \n",
       "Symbicort_Turbohaler 100mcg/6mcg (120 D)                          396.00   \n",
       "\n",
       "                                                    item_count_high  \\\n",
       "bnf_name                                                              \n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window...              1.0   \n",
       "Promixin_Pdr For Neb 1mega u                                    2.0   \n",
       "Nutrison_Energy M/Fibre Liq (Vanilla)                           1.0   \n",
       "Loprofin_L/P Fusilli                                            1.0   \n",
       "Symbicort_Turbohaler 100mcg/6mcg (120 D)                       11.0   \n",
       "\n",
       "                                                    item_pay_oope_amt_high  \\\n",
       "bnf_name                                                                     \n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window...                   86.99   \n",
       "Promixin_Pdr For Neb 1mega u                                        158.38   \n",
       "Nutrison_Energy M/Fibre Liq (Vanilla)                                79.00   \n",
       "Loprofin_L/P Fusilli                                                 79.00   \n",
       "Symbicort_Turbohaler 100mcg/6mcg (120 D)                             79.00   \n",
       "\n",
       "                                                    oope_count_high  \\\n",
       "bnf_name                                                              \n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window...              1.0   \n",
       "Promixin_Pdr For Neb 1mega u                                    2.0   \n",
       "Nutrison_Energy M/Fibre Liq (Vanilla)                           1.0   \n",
       "Loprofin_L/P Fusilli                                            1.0   \n",
       "Symbicort_Turbohaler 100mcg/6mcg (120 D)                        1.0   \n",
       "\n",
       "                                                    oope_per_oope_item_high  \\\n",
       "bnf_name                                                                      \n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window...                    86.99   \n",
       "Promixin_Pdr For Neb 1mega u                                          79.19   \n",
       "Nutrison_Energy M/Fibre Liq (Vanilla)                                 79.00   \n",
       "Loprofin_L/P Fusilli                                                  79.00   \n",
       "Symbicort_Turbohaler 100mcg/6mcg (120 D)                              79.00   \n",
       "\n",
       "                                                    item_pay_dr_nic_low  \\\n",
       "bnf_name                                                                  \n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window...                  NaN   \n",
       "Promixin_Pdr For Neb 1mega u                                        NaN   \n",
       "Nutrison_Energy M/Fibre Liq (Vanilla)                             61.32   \n",
       "Loprofin_L/P Fusilli                                                NaN   \n",
       "Symbicort_Turbohaler 100mcg/6mcg (120 D)                         561.00   \n",
       "\n",
       "                                                    item_count_low  \\\n",
       "bnf_name                                                             \n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window...             NaN   \n",
       "Promixin_Pdr For Neb 1mega u                                   NaN   \n",
       "Nutrison_Energy M/Fibre Liq (Vanilla)                          2.0   \n",
       "Loprofin_L/P Fusilli                                           NaN   \n",
       "Symbicort_Turbohaler 100mcg/6mcg (120 D)                      13.0   \n",
       "\n",
       "                                                    item_pay_oope_amt_low  \\\n",
       "bnf_name                                                                    \n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window...                    NaN   \n",
       "Promixin_Pdr For Neb 1mega u                                          NaN   \n",
       "Nutrison_Energy M/Fibre Liq (Vanilla)                                 0.0   \n",
       "Loprofin_L/P Fusilli                                                  NaN   \n",
       "Symbicort_Turbohaler 100mcg/6mcg (120 D)                              0.0   \n",
       "\n",
       "                                                    oope_count_low  \\\n",
       "bnf_name                                                             \n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window...             NaN   \n",
       "Promixin_Pdr For Neb 1mega u                                   NaN   \n",
       "Nutrison_Energy M/Fibre Liq (Vanilla)                          0.0   \n",
       "Loprofin_L/P Fusilli                                           NaN   \n",
       "Symbicort_Turbohaler 100mcg/6mcg (120 D)                       0.0   \n",
       "\n",
       "                                                    oope_per_oope_item_low  \n",
       "bnf_name                                                                    \n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window...                     NaN  \n",
       "Promixin_Pdr For Neb 1mega u                                           NaN  \n",
       "Nutrison_Energy M/Fibre Liq (Vanilla)                                  NaN  \n",
       "Loprofin_L/P Fusilli                                                   NaN  \n",
       "Symbicort_Turbohaler 100mcg/6mcg (120 D)                               NaN  "
      ]
     },
     "execution_count": 159,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Now compare these\n",
    "#baddy_presentations = df3[(df3['oope_per_item'] > 0) & (df3['item_count'] > 10)]\n",
    "compared = highest_presentations.merge(\n",
    "    lowest_presentations,\n",
    "    left_index=True,\n",
    "    right_index=True,\n",
    "    suffixes=[\"_high\", \"_low\"],\n",
    "    how=\"outer\"\n",
    ")\n",
    "compared.sort_values(\"oope_per_oope_item_high\", ascending=False).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "good_only = compared[np.isnan(compared['item_count_high'])]\n",
    "bad_only = compared[np.isnan(compared['item_count_low'])]\n",
    "both = compared[(~np.isnan(compared['item_count_low'])) & (~np.isnan(compared['item_count_high']))]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are 758 presentations only prescribed in the dispensaries with the lowest OOPE:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>item_pay_dr_nic_high</th>\n",
       "      <th>item_count_high</th>\n",
       "      <th>item_pay_oope_amt_high</th>\n",
       "      <th>oope_count_high</th>\n",
       "      <th>oope_per_oope_item_high</th>\n",
       "      <th>item_pay_dr_nic_low</th>\n",
       "      <th>item_count_low</th>\n",
       "      <th>item_pay_oope_amt_low</th>\n",
       "      <th>oope_count_low</th>\n",
       "      <th>oope_per_oope_item_low</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>758.000000</td>\n",
       "      <td>758.000000</td>\n",
       "      <td>758.0</td>\n",
       "      <td>758.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>68.327744</td>\n",
       "      <td>1.742744</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>121.116004</td>\n",
       "      <td>1.614349</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.220000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9.062500</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>25.860000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>70.355000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1256.400000</td>\n",
       "      <td>15.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       item_pay_dr_nic_high  item_count_high  item_pay_oope_amt_high  \\\n",
       "count                   0.0              0.0                     0.0   \n",
       "mean                    NaN              NaN                     NaN   \n",
       "std                     NaN              NaN                     NaN   \n",
       "min                     NaN              NaN                     NaN   \n",
       "25%                     NaN              NaN                     NaN   \n",
       "50%                     NaN              NaN                     NaN   \n",
       "75%                     NaN              NaN                     NaN   \n",
       "max                     NaN              NaN                     NaN   \n",
       "\n",
       "       oope_count_high  oope_per_oope_item_high  item_pay_dr_nic_low  \\\n",
       "count              0.0                      0.0           758.000000   \n",
       "mean               NaN                      NaN            68.327744   \n",
       "std                NaN                      NaN           121.116004   \n",
       "min                NaN                      NaN             0.220000   \n",
       "25%                NaN                      NaN             9.062500   \n",
       "50%                NaN                      NaN            25.860000   \n",
       "75%                NaN                      NaN            70.355000   \n",
       "max                NaN                      NaN          1256.400000   \n",
       "\n",
       "       item_count_low  item_pay_oope_amt_low  oope_count_low  \\\n",
       "count      758.000000                  758.0           758.0   \n",
       "mean         1.742744                    0.0             0.0   \n",
       "std          1.614349                    0.0             0.0   \n",
       "min          1.000000                    0.0             0.0   \n",
       "25%          1.000000                    0.0             0.0   \n",
       "50%          1.000000                    0.0             0.0   \n",
       "75%          2.000000                    0.0             0.0   \n",
       "max         15.000000                    0.0             0.0   \n",
       "\n",
       "       oope_per_oope_item_low  \n",
       "count                     0.0  \n",
       "mean                      NaN  \n",
       "std                       NaN  \n",
       "min                       NaN  \n",
       "25%                       NaN  \n",
       "50%                       NaN  \n",
       "75%                       NaN  \n",
       "max                       NaN  "
      ]
     },
     "execution_count": 161,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "good_only.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "...and 117 presentations-that-have-OOPE which only prescribed in those with the highest OOPE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>item_pay_dr_nic_high</th>\n",
       "      <th>item_count_high</th>\n",
       "      <th>item_pay_oope_amt_high</th>\n",
       "      <th>oope_count_high</th>\n",
       "      <th>oope_per_oope_item_high</th>\n",
       "      <th>item_pay_dr_nic_low</th>\n",
       "      <th>item_count_low</th>\n",
       "      <th>item_pay_oope_amt_low</th>\n",
       "      <th>oope_count_low</th>\n",
       "      <th>oope_per_oope_item_low</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>117.000000</td>\n",
       "      <td>117.000000</td>\n",
       "      <td>117.000000</td>\n",
       "      <td>117.000000</td>\n",
       "      <td>117.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>192.259573</td>\n",
       "      <td>3.666667</td>\n",
       "      <td>98.354444</td>\n",
       "      <td>3.128205</td>\n",
       "      <td>41.470097</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>237.674093</td>\n",
       "      <td>5.333693</td>\n",
       "      <td>119.587144</td>\n",
       "      <td>5.066010</td>\n",
       "      <td>16.411073</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>2.420000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>36.980000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>9.523810</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>41.940000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>36.980000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>36.980000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>98.260000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>73.960000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>36.980000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>268.320000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>110.940000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>40.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1078.580000</td>\n",
       "      <td>30.000000</td>\n",
       "      <td>732.860000</td>\n",
       "      <td>29.000000</td>\n",
       "      <td>86.990000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       item_pay_dr_nic_high  item_count_high  item_pay_oope_amt_high  \\\n",
       "count            117.000000       117.000000              117.000000   \n",
       "mean             192.259573         3.666667               98.354444   \n",
       "std              237.674093         5.333693              119.587144   \n",
       "min                2.420000         1.000000               36.980000   \n",
       "25%               41.940000         1.000000               36.980000   \n",
       "50%               98.260000         2.000000               73.960000   \n",
       "75%              268.320000         3.000000              110.940000   \n",
       "max             1078.580000        30.000000              732.860000   \n",
       "\n",
       "       oope_count_high  oope_per_oope_item_high  item_pay_dr_nic_low  \\\n",
       "count       117.000000               117.000000                  0.0   \n",
       "mean          3.128205                41.470097                  NaN   \n",
       "std           5.066010                16.411073                  NaN   \n",
       "min           1.000000                 9.523810                  NaN   \n",
       "25%           1.000000                36.980000                  NaN   \n",
       "50%           1.000000                36.980000                  NaN   \n",
       "75%           2.000000                40.000000                  NaN   \n",
       "max          29.000000                86.990000                  NaN   \n",
       "\n",
       "       item_count_low  item_pay_oope_amt_low  oope_count_low  \\\n",
       "count             0.0                    0.0             0.0   \n",
       "mean              NaN                    NaN             NaN   \n",
       "std               NaN                    NaN             NaN   \n",
       "min               NaN                    NaN             NaN   \n",
       "25%               NaN                    NaN             NaN   \n",
       "50%               NaN                    NaN             NaN   \n",
       "75%               NaN                    NaN             NaN   \n",
       "max               NaN                    NaN             NaN   \n",
       "\n",
       "       oope_per_oope_item_low  \n",
       "count                     0.0  \n",
       "mean                      NaN  \n",
       "std                       NaN  \n",
       "min                       NaN  \n",
       "25%                       NaN  \n",
       "50%                       NaN  \n",
       "75%                       NaN  \n",
       "max                       NaN  "
      ]
     },
     "execution_count": 162,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bad_only[bad_only['item_pay_oope_amt_high'] > 0].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "...and 2498 in both, of which 192 have any OOPE:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>item_pay_dr_nic_high</th>\n",
       "      <th>item_count_high</th>\n",
       "      <th>item_pay_oope_amt_high</th>\n",
       "      <th>oope_count_high</th>\n",
       "      <th>oope_per_oope_item_high</th>\n",
       "      <th>item_pay_dr_nic_low</th>\n",
       "      <th>item_count_low</th>\n",
       "      <th>item_pay_oope_amt_low</th>\n",
       "      <th>oope_count_low</th>\n",
       "      <th>oope_per_oope_item_low</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>169.000000</td>\n",
       "      <td>169.000000</td>\n",
       "      <td>169.000000</td>\n",
       "      <td>169.000000</td>\n",
       "      <td>169.000000</td>\n",
       "      <td>169.000000</td>\n",
       "      <td>169.000000</td>\n",
       "      <td>169.0</td>\n",
       "      <td>169.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>1133.642308</td>\n",
       "      <td>56.189349</td>\n",
       "      <td>326.283728</td>\n",
       "      <td>18.378698</td>\n",
       "      <td>33.148260</td>\n",
       "      <td>539.685976</td>\n",
       "      <td>26.408284</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>2179.961654</td>\n",
       "      <td>138.535494</td>\n",
       "      <td>578.767144</td>\n",
       "      <td>46.437475</td>\n",
       "      <td>20.867779</td>\n",
       "      <td>1087.537552</td>\n",
       "      <td>65.862990</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>8.080000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>9.920000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.152181</td>\n",
       "      <td>3.640000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>141.600000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>61.990000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>18.246667</td>\n",
       "      <td>48.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>417.600000</td>\n",
       "      <td>15.000000</td>\n",
       "      <td>79.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>35.909091</td>\n",
       "      <td>128.800000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1097.600000</td>\n",
       "      <td>52.000000</td>\n",
       "      <td>237.000000</td>\n",
       "      <td>14.000000</td>\n",
       "      <td>39.500000</td>\n",
       "      <td>446.910000</td>\n",
       "      <td>21.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>14929.820000</td>\n",
       "      <td>1451.000000</td>\n",
       "      <td>3476.120000</td>\n",
       "      <td>486.000000</td>\n",
       "      <td>79.000000</td>\n",
       "      <td>7144.200000</td>\n",
       "      <td>688.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       item_pay_dr_nic_high  item_count_high  item_pay_oope_amt_high  \\\n",
       "count            169.000000       169.000000              169.000000   \n",
       "mean            1133.642308        56.189349              326.283728   \n",
       "std             2179.961654       138.535494              578.767144   \n",
       "min                8.080000         1.000000                9.920000   \n",
       "25%              141.600000         5.000000               61.990000   \n",
       "50%              417.600000        15.000000               79.000000   \n",
       "75%             1097.600000        52.000000              237.000000   \n",
       "max            14929.820000      1451.000000             3476.120000   \n",
       "\n",
       "       oope_count_high  oope_per_oope_item_high  item_pay_dr_nic_low  \\\n",
       "count       169.000000               169.000000           169.000000   \n",
       "mean         18.378698                33.148260           539.685976   \n",
       "std          46.437475                20.867779          1087.537552   \n",
       "min           1.000000                 0.152181             3.640000   \n",
       "25%           2.000000                18.246667            48.000000   \n",
       "50%           5.000000                35.909091           128.800000   \n",
       "75%          14.000000                39.500000           446.910000   \n",
       "max         486.000000                79.000000          7144.200000   \n",
       "\n",
       "       item_count_low  item_pay_oope_amt_low  oope_count_low  \\\n",
       "count      169.000000                  169.0           169.0   \n",
       "mean        26.408284                    0.0             0.0   \n",
       "std         65.862990                    0.0             0.0   \n",
       "min          1.000000                    0.0             0.0   \n",
       "25%          2.000000                    0.0             0.0   \n",
       "50%          5.000000                    0.0             0.0   \n",
       "75%         21.000000                    0.0             0.0   \n",
       "max        688.000000                    0.0             0.0   \n",
       "\n",
       "       oope_per_oope_item_low  \n",
       "count                     0.0  \n",
       "mean                      NaN  \n",
       "std                       NaN  \n",
       "min                       NaN  \n",
       "25%                       NaN  \n",
       "50%                       NaN  \n",
       "75%                       NaN  \n",
       "max                       NaN  "
      ]
     },
     "execution_count": 163,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "both[both['item_pay_oope_amt_high'] > 0].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "bnf_name\n",
       "Ensure Plus Advance_Liq (5 Flav)     79.0\n",
       "Complan_Shake Pdr Sach 57g (Choc)    79.0\n",
       "Aveeno_Crm                           79.0\n",
       "Janumet_Tab 50mg/1g                  79.0\n",
       "Nifedipine_Cap 10mg M/R              79.0\n",
       "Name: oope_per_oope_item_high, dtype: float64"
      ]
     },
     "execution_count": 174,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Total OOPE spent per item, for things prescribed in both places - top 5 (of 192 with any OOPE)\n",
    "both['oope_per_oope_item_high'].sort_values(ascending=False).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "bnf_name\n",
       "Coloplast_Sensura Mio Maxi DrnblBag Grey+Window C/Fit10-55mm    86.99\n",
       "Promixin_Pdr For Neb 1mega u                                    79.19\n",
       "Calshake_Pdr Sach 87g (Vanilla)                                 79.00\n",
       "Fresubin Energy Fibre_Liq (6 Flav)                              79.00\n",
       "Fortisip Compact Fibre_Liq S/Pack                               79.00\n",
       "Name: oope_per_oope_item_high, dtype: float64"
      ]
     },
     "execution_count": 175,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Total OOPE spent per item, for things prescribed in only high-oope places - top 5 (of 103 with any OOPE)\n",
    "bad_only['oope_per_oope_item_high'].sort_values(ascending=False).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Finally, there is a weird £36.98 thing going on\n",
    "I've noticed a lot of the things with high OOPE have identical OOPE-per item - for example, £20.99, £36.98, and £79, all examined here.\n",
    "\n",
    "These appear to be fixed OOPE prices, regardless of the item - for example, the relatively common `Bio-Vitamin D3_Cap 800u` costs 44p per pack, but £36.98 fixed OOPE.\n",
    "\n",
    "At the other end, an unusual special like `Levocarnitine_Oral Soln Paed 1.5g/5ml` costs £1571 per pack, and still £36.98 fixed OOPE.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "lines_to_next_cell": 2
   },
   "outputs": [],
   "source": [
    "sql = \"\"\"\n",
    "SELECT\n",
    "  *\n",
    "FROM\n",
    "  dispensers.dispensing_with_metadata\n",
    "WHERE\n",
    "  (item_pay_oope_amt / item_count = 36.98\n",
    "    OR (item_pay_oope_amt / item_count) = 20.99 \n",
    "    OR (item_pay_oope_amt / item_count) = 79.0)\n",
    "  AND year_month = '201703'\n",
    "\"\"\"\n",
    "df = pd.io.gbq.read_gbq(sql, 'ebmdatalab', dialect='standard')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['oope_per_nic'] = (df['item_pay_oope_amt'] / df['item_count']) / df['item_pay_dr_nic']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "A total of 50145.39 spent in OOPE for items with OOPE at these three levels\n"
     ]
    }
   ],
   "source": [
    "print(\"A total of {} spent in OOPE for items with OOPE at these three levels\".format(df['item_pay_oope_amt'].sum()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>bnf_name</th>\n",
       "      <th>bnf_code</th>\n",
       "      <th>item_pay_dr_nic</th>\n",
       "      <th>item_pay_oope_amt</th>\n",
       "      <th>oope_per_nic</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>16301</th>\n",
       "      <td>Aloe Pura_Aloe Vera Max Strgh Juice</td>\n",
       "      <td>091104000BBJUA0</td>\n",
       "      <td>0.01</td>\n",
       "      <td>20.99</td>\n",
       "      <td>2099.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>597</th>\n",
       "      <td>Lamb_Pure Fish Oil Cap 1100mg</td>\n",
       "      <td>090601100BBGSA0</td>\n",
       "      <td>0.10</td>\n",
       "      <td>36.98</td>\n",
       "      <td>369.800000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21117</th>\n",
       "      <td>Fluoxetine HCl_Tab Disper 20mg S/F</td>\n",
       "      <td>0403030E0AAAKAK</td>\n",
       "      <td>0.12</td>\n",
       "      <td>20.99</td>\n",
       "      <td>174.916667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15597</th>\n",
       "      <td>Lamb_Nicotinamide Tab 250mg (8054)</td>\n",
       "      <td>090602200BBBHA0</td>\n",
       "      <td>0.31</td>\n",
       "      <td>36.98</td>\n",
       "      <td>119.290323</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5164</th>\n",
       "      <td>Ergocalciferol_Tab 12.5mcg</td>\n",
       "      <td>0906040N0AAEWEW</td>\n",
       "      <td>0.37</td>\n",
       "      <td>36.98</td>\n",
       "      <td>99.945946</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                  bnf_name         bnf_code  item_pay_dr_nic  \\\n",
       "16301  Aloe Pura_Aloe Vera Max Strgh Juice  091104000BBJUA0             0.01   \n",
       "597          Lamb_Pure Fish Oil Cap 1100mg  090601100BBGSA0             0.10   \n",
       "21117   Fluoxetine HCl_Tab Disper 20mg S/F  0403030E0AAAKAK             0.12   \n",
       "15597   Lamb_Nicotinamide Tab 250mg (8054)  090602200BBBHA0             0.31   \n",
       "5164            Ergocalciferol_Tab 12.5mcg  0906040N0AAEWEW             0.37   \n",
       "\n",
       "       item_pay_oope_amt  oope_per_nic  \n",
       "16301              20.99   2099.000000  \n",
       "597                36.98    369.800000  \n",
       "21117              20.99    174.916667  \n",
       "15597              36.98    119.290323  \n",
       "5164               36.98     99.945946  "
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[['bnf_name', 'bnf_code', 'item_pay_dr_nic', 'item_pay_oope_amt', 'oope_per_nic']].sort_values('oope_per_nic', ascending=False).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The most common of these:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "bnf_name\n",
       "Dermacolor_Camouflage Creme        1169\n",
       "Dermacolor_Fixing Pdr               758\n",
       "Juvela_G/F Fresh Loaf Wte (Cut)     536\n",
       "Cyanocobalamin_Tab 1mg M/R          521\n",
       "Veil_Cover Crm                      477\n",
       "Veil_Finishing Pdr                  414\n",
       "Unspec Drug Code_                   383\n",
       "Vega_ZM3 Multivit/Min V/Cap         346\n",
       "Solgar_Vit D3 Tab 1,000u            303\n",
       "Solgar_Mag Cit Tab 200mg            275\n",
       "Name: item_count, dtype: int64"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby('bnf_name').agg('sum')['item_count'].sort_values(ascending=False).head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## What DT categories are the OOPE items in?\n",
    "Only Category C items can be claimed, but:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>vmpp</th>\n",
       "      <th>tariff_category</th>\n",
       "      <th>name</th>\n",
       "      <th>hq_name</th>\n",
       "      <th>bnf_code</th>\n",
       "      <th>item_count</th>\n",
       "      <th>item_pay_oope_amt</th>\n",
       "      <th>actual_cost</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Simvastatin 20mg tablets 28 tablet</td>\n",
       "      <td>Part VIIIA Category M</td>\n",
       "      <td>BOOTS UK LIMITED</td>\n",
       "      <td>BOOTS GROUP PLC</td>\n",
       "      <td>0212000Y0AAABAB</td>\n",
       "      <td>197</td>\n",
       "      <td>24.95</td>\n",
       "      <td>205.53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Paracetamol 250mg/5ml oral suspension sugar fr...</td>\n",
       "      <td>Part VIIIA Category A</td>\n",
       "      <td>KAMSONS PHARMACY</td>\n",
       "      <td>WAREMOSS LTD</td>\n",
       "      <td>0407010H0AABGBG</td>\n",
       "      <td>17</td>\n",
       "      <td>6.50</td>\n",
       "      <td>78.90</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Sulpiride 200mg/5ml oral solution sugar free 1...</td>\n",
       "      <td>Part VIIIA Category A</td>\n",
       "      <td>LLOYDSPHARMACY</td>\n",
       "      <td>LLOYDS PHARMACY LTD</td>\n",
       "      <td>0402010U0AAAEAE</td>\n",
       "      <td>3</td>\n",
       "      <td>43.00</td>\n",
       "      <td>256.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Sildenafil 25mg tablets 4 tablet</td>\n",
       "      <td>Part VIIIA Category M</td>\n",
       "      <td>YOUR LOCAL BOOTS PHARMACY</td>\n",
       "      <td>BOOTS GROUP PLC</td>\n",
       "      <td>0704050Z0AAABAB</td>\n",
       "      <td>1</td>\n",
       "      <td>20.00</td>\n",
       "      <td>23.60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Haloperidol 5mg/5ml oral solution sugar free 1...</td>\n",
       "      <td>Part VIIIA Category A</td>\n",
       "      <td>MAGSON PHARMACY</td>\n",
       "      <td>ADAM MYERS LTD</td>\n",
       "      <td>0402010J0AABCBC</td>\n",
       "      <td>2</td>\n",
       "      <td>11.99</td>\n",
       "      <td>71.91</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                vmpp        tariff_category  \\\n",
       "0                 Simvastatin 20mg tablets 28 tablet  Part VIIIA Category M   \n",
       "1  Paracetamol 250mg/5ml oral suspension sugar fr...  Part VIIIA Category A   \n",
       "2  Sulpiride 200mg/5ml oral solution sugar free 1...  Part VIIIA Category A   \n",
       "3                   Sildenafil 25mg tablets 4 tablet  Part VIIIA Category M   \n",
       "4  Haloperidol 5mg/5ml oral solution sugar free 1...  Part VIIIA Category A   \n",
       "\n",
       "                        name              hq_name         bnf_code  \\\n",
       "0           BOOTS UK LIMITED      BOOTS GROUP PLC  0212000Y0AAABAB   \n",
       "1           KAMSONS PHARMACY         WAREMOSS LTD  0407010H0AABGBG   \n",
       "2             LLOYDSPHARMACY  LLOYDS PHARMACY LTD  0402010U0AAAEAE   \n",
       "3  YOUR LOCAL BOOTS PHARMACY      BOOTS GROUP PLC  0704050Z0AAABAB   \n",
       "4            MAGSON PHARMACY       ADAM MYERS LTD  0402010J0AABCBC   \n",
       "\n",
       "   item_count  item_pay_oope_amt  actual_cost  \n",
       "0         197              24.95       205.53  \n",
       "1          17               6.50        78.90  \n",
       "2           3              43.00       256.75  \n",
       "3           1              20.00        23.60  \n",
       "4           2              11.99        71.91  "
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sql = \"\"\"\n",
    "SELECT\n",
    "  vmpp,\n",
    "  tariff_category,\n",
    "  d.name,\n",
    "  d.hq_name,\n",
    "  d.bnf_code,\n",
    "  item_count,\n",
    "  item_pay_oope_amt,\n",
    "  actual_cost\n",
    "FROM\n",
    "  dispensers.dispensing_with_metadata d\n",
    "LEFT JOIN\n",
    "  `dmd.dt_viewer` v\n",
    "ON\n",
    "  d.bnf_code = v.bnf_code\n",
    "WHERE\n",
    "  year_month = '201703'\n",
    "  AND v.date = '2017-03-01'\n",
    "  AND item_pay_oope_amt > 0\n",
    "  AND tariff_category != \"Part VIIIA Category C\"\n",
    "\"\"\"\n",
    "df = pd.io.gbq.read_gbq(sql, 'ebmdatalab', dialect='standard')\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "A total of 1165.0700000000002 OOPE was paid on things that shouldn't\n"
     ]
    }
   ],
   "source": [
    "print(\"A total of {} OOPE was paid on things that shouldn't\".format(df.item_pay_oope_amt.sum()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>vmpp</th>\n",
       "      <th>name</th>\n",
       "      <th>hq_name</th>\n",
       "      <th>bnf_code</th>\n",
       "      <th>item_count</th>\n",
       "      <th>item_pay_oope_amt</th>\n",
       "      <th>actual_cost</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tariff_category</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Part VIIIA Category A</th>\n",
       "      <td>21</td>\n",
       "      <td>21</td>\n",
       "      <td>21</td>\n",
       "      <td>21</td>\n",
       "      <td>21</td>\n",
       "      <td>21</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Part VIIIA Category M</th>\n",
       "      <td>29</td>\n",
       "      <td>29</td>\n",
       "      <td>29</td>\n",
       "      <td>29</td>\n",
       "      <td>29</td>\n",
       "      <td>29</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       vmpp  name  hq_name  bnf_code  item_count  \\\n",
       "tariff_category                                                    \n",
       "Part VIIIA Category A    21    21       21        21          21   \n",
       "Part VIIIA Category M    29    29       29        29          29   \n",
       "\n",
       "                       item_pay_oope_amt  actual_cost  \n",
       "tariff_category                                        \n",
       "Part VIIIA Category A                 21           21  \n",
       "Part VIIIA Category M                 29           29  "
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby('tariff_category').count()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python (jupyter virtualenv)",
   "language": "python",
   "name": "jupyter"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}