📊 antibiotics: adding aggregated antimicrobial use data by class

etl/steps/data/garden/antibiotics/2024-11-12/antimicrobial_usage.meta.yml

@@ -4,8 +4,7 @@ definitions:
     presentation:
       topic_tags:
         - Global Health
-  aware_description:
-    <% if aware == "A" %>
+  aware_description: <% if aware == "A" %>
     Access antibiotics have activity against a wide range of common pathogens and show lower resistance potential than antibiotics in the other groups.
     <% elif aware == "W" %>
     Watch antibiotic have higher resistance potential and include most of the highest priority agents among the Critically Important Antimicrobials for Human Medicine and/or antibiotics that are at relatively high risk of bacterial resistance.
@@ -14,8 +13,7 @@ definitions:
     <% elif aware == "O" %>
     The use of the Not classified/Not recommended antibiotics is not evidence-based, nor recommended in high-quality international guidelines.  WHO does not recommend the use of these antibiotics in clinical practice.
     <% endif %>
-  routeofadministration:
-    <% if routeofadministration == "O" %>
+  routeofadministration: <% if routeofadministration == "O" %>
     orally administered
     <% elif routeofadministration == "P" %>
     parentearally administered
@@ -24,36 +22,84 @@ definitions:
     <% elif routeofadministration == "I" %>
     inhaled
     <% endif %>
-
+  antimicrobialclass:
+    <% if antimicrobialclass == "Antibacterials (ATC J01, A07AA, P01AB, ATC J04A)" %>
+    antibiotics including antituberculosis drugs
+    <% elif antimicrobialclass == "Antimalarials (ATC P01B)" %>
+    antimalarials
+    <% elif antimicrobialclass == "Antimycotics and antifungals for systemic use (J02, D01B)" %>
+    antifungals
+    <% elif antimicrobialclass == "Antivirals for systemic use (ATC J05)" %>
+    antivirals
+    <% elif antimicrobialclass == "Drugs for the treatment of tuberculosis (ATC J04A)" %>
+    antituberculosis drugs
+    <% endif %>
 
 # Learn more about the available fields:
 # http://docs.owid.io/projects/etl/architecture/metadata/reference/
 dataset:
-  update_period_days: 308
-
+  update_period_days: 365
 
 tables:
   class:
     variables:
       ddd:
-        title: Defined daily doses of {definitions.routeofadministration}  << antimicrobialclass>> - << atc4name.lower() >>  used
-        description_short: Volume of antimicrobials used in a given year.
-        #description_processing: <<notes>>
+        title: Defined daily doses of {definitions.routeofadministration}  << antimicrobialclass.lower()>> - << atc4name.lower() >>  used
+        description_short: Total [defined daily doses](#dod:defined-daily-doses) of antimicrobials used in a given year.
         unit: defined daily doses
       did:
         title: Defined daily doses per 1000 inhabitants per day of {definitions.routeofadministration} << antimicrobialclass>> - << atc4name.lower() >> used
-        description_short: Volume of antimicrobials used per 1000 inhabitants per day.
-        #description_processing: <<notes>>
+        description_short: Total [defined daily doses](#dod:defined-daily-doses) of antimicrobials used per 1000 inhabitants per day.
         unit: defined daily doses per 1000 inhabitants per day
+  class_aggregated:
+    variables:
+      ddd_anti_malarials:
+        title: Defined daily doses of antimalarials used
+        description_short: Total [defined daily doses](#dod:defined-daily-doses) of antimalarials used in a given year.
+        unit: defined daily doses
+      ddd_antibacterials_and_antituberculosis:
+        title: Defined daily doses of antibiotics and antituberculosis drugs used
+        description_short: Total [defined daily doses](#dod:defined-daily-doses) of antibiotics and antituberculosis drugs used in a given year.
+        unit: defined daily doses
+      ddd_antifungals:
+        title: Defined daily doses of antifungals used
+        description_short: Total [defined daily doses](#dod:defined-daily-doses) of antifungals used in a given year.
+        unit: defined daily doses
+      ddd_antituberculosis:
+        title: Defined daily doses of antituberculosis drugs used
+        description_short: Total [defined daily doses](#dod:defined-daily-doses) of antituberculosis drugs used in a given year.
+        unit: defined daily doses
+      ddd_antivirals:
+        title: Defined daily doses of antivirals used
+        description_short: Total [defined daily doses](#dod:defined-daily-doses) of antivirals used in a given year.
+        unit: defined daily doses
+      did_anti_malarials:
+        title: Defined daily doses of antimalarials used per 1,000 inhabitants per day
+        description_short: Total [defined daily doses](#dod:defined-daily-doses) of antimalarials used in a given year per 1,000 inhabitants per day.
+        unit: defined daily doses per 1,000 inhabitants per day
+      did_antibacterials_and_antituberculosis:
+        title: Defined daily doses of antibiotics and antituberculosis drugs used per 1,000 inhabitants per day
+        description_short: Total [defined daily doses](#dod:defined-daily-doses) of antibiotics and antituberculosis drugs used in a given year per 1,000 inhabitants per day.
+        unit: defined daily doses per 1,000 inhabitants per day
+      did_antifungals:
+        title: Defined daily doses of antifungals used per 1,000 inhabitants per day
+        description_short: Total [defined daily doses](#dod:defined-daily-doses) of antifungals used in a given year per 1,000 inhabitants per day.
+        unit: defined daily doses per 1,000 inhabitants per day
+      did_antituberculosis:
+        title: Defined daily doses of antituberculosis drugs used per 1,000 inhabitants per day
+        description_short: Total [defined daily doses](#dod:defined-daily-doses) of antituberculosis drugs used in a given year per 1,000 inhabitants per day.
+        unit: defined daily doses per 1,000 inhabitants per day
+      did_antivirals:
+        title: Defined daily doses of antivirals used per 1,000 inhabitants per day
+        description_short: Total [defined daily doses](#dod:defined-daily-doses) of antivirals used in a given year per 1,000 inhabitants per day.
+        unit: defined daily doses per 1,000 inhabitants per day
   aware:
     variables:
       ddd:
         title: Defined daily doses of << awarelabel >> antibiotics used
-        description_short: "Volume of AWaRe category: << awarelabel >> antibiotics used in a given year. {definitions.aware_description}"
-        #description_processing: <<notes>>
+        description_short: "Total [defined daily doses](#dod:defined-daily-doses) of AWaRe category: << awarelabel >> antibiotics used in a given year. {definitions.aware_description}"
         unit: defined daily doses
       did:
         title: Defined daily doses per 1000 inhabitants per day of << awarelabel>> antibiotics used
-        description_short: "Volume of AWaRe category: <<awarelabel>> used per 1000 inhabitants per day. {definitions.aware_description}"
-        #description_processing: <<notes>>
+        description_short: "Total [defined daily doses](#dod:defined-daily-doses) of AWaRe category: <<awarelabel>> used per 1000 inhabitants per day. {definitions.aware_description}"
         unit: defined daily doses per 1000 inhabitants per day

etl/steps/data/garden/antibiotics/2024-11-12/antimicrobial_usage.py

@@ -1,5 +1,9 @@
 """Load a meadow dataset and create a garden dataset."""
 
+import pandas as pd
+from owid.catalog import Table
+from owid.catalog import processing as pr
+
 from etl.data_helpers import geo
 from etl.helpers import PathFinder, create_dataset
 
@@ -23,6 +27,9 @@ def run(dest_dir: str) -> None:
     tb_class = geo.harmonize_countries(df=tb_class, countries_file=paths.country_mapping_path)
     tb_aware = geo.harmonize_countries(df=tb_aware, countries_file=paths.country_mapping_path)
 
+    # Aggregate by antimicrobial class
+    tb_class_agg, tb_notes = aggregate_antimicrobial_classes(tb_class)
+    # Save the origins of the aggregated table to insert back in later
     # Drop columns that are not needed in the garden dataset.
     tb_class = tb_class.drop(
         columns=["whoregioncode", "whoregionname", "countryiso3", "incomeworldbankjune", "atc4", "notes"]
@@ -31,14 +38,128 @@ def run(dest_dir: str) -> None:
 
     tb_class = tb_class.format(["country", "year", "antimicrobialclass", "atc4name", "routeofadministration"])
     tb_aware = tb_aware.format(["country", "year", "awarelabel"])
+    tb_class_agg = pivot_aggregated_table(tb_class_agg, tb_notes)
+    tb_class_agg = tb_class_agg.format(["country", "year"], short_name="class_aggregated")
 
     #
     # Save outputs.
     #
     # Create a new garden dataset with the same metadata as the meadow dataset.
     ds_garden = create_dataset(
-        dest_dir, tables=[tb_class, tb_aware], check_variables_metadata=True, default_metadata=ds_meadow.metadata
+        dest_dir,
+        tables=[tb_class, tb_aware, tb_class_agg],
+        check_variables_metadata=True,
+        default_metadata=ds_meadow.metadata,
     )
 
     # Save changes in the new garden dataset.
     ds_garden.save()
+
+
+def pivot_aggregated_table(tb_class_agg: Table, tb_notes: Table) -> Table:
+    """
+    Pivot the aggregated table to have a column for each antimicrobial class, then add the description_processing metadata
+    """
+
+    tb_notes_dict = {
+        "Antibacterials (ATC J01, A07AA, P01AB)": "antibacterials",
+        "Antimalarials (ATC P01B)": "anti_malarials",
+        "Antimycotics and antifungals for systemic use (J02, D01B)": "antifungals",
+        "Antivirals for systemic use (ATC J05)": "antivirals",
+        "Drugs for the treatment of tuberculosis (ATC J04A)": "antituberculosis",
+        "Antibacterials (ATC J01, A07AA, P01AB, ATC J04A)": "antibacterials_and_antituberculosis",
+    }
+    tb_notes["category"] = tb_notes["antimicrobialclass"].map(tb_notes_dict)
+    tb_class_agg = tb_class_agg.copy(deep=True)
+    tb_class_agg["antimicrobialclass"] = tb_class_agg["antimicrobialclass"].replace(tb_notes_dict)
+    tb_class_agg = tb_class_agg.pivot(index=["country", "year"], columns="antimicrobialclass", values=["ddd", "did"])
+    tb_class_agg.columns = tb_class_agg.columns.to_flat_index()
+    tb_class_agg.columns = [f"{col[0]}_{col[1]}" for col in tb_class_agg.columns]
+    tb_class_agg = tb_class_agg.reset_index()
+
+    for key in tb_notes_dict.values():
+        if f"ddd_{key}" in tb_class_agg.columns:
+            tb_class_agg[f"ddd_{key}"].metadata.description_processing = tb_notes["description_processing"][
+                tb_notes["category"] == key
+            ]
+        if f"did_{key}" in tb_class_agg.columns:
+            tb_class_agg[f"did_{key}"].metadata.description_processing = tb_notes["description_processing"][
+                tb_notes["category"] == key
+            ].astype(str)
+
+    return tb_class_agg
+
+
+def aggregate_antimicrobial_classes(tb: Table) -> Table:
+    """
+    Aggregating by antimicrobial class, we want to combine antibacterials and antituberculosis, but also keep antituberculosis separately
+    """
+    tb = tb.copy(deep=True)
+    # Convert the column to strings (if not already done)
+    tb["antimicrobialclass"] = tb["antimicrobialclass"].astype(str)
+
+    # Create a completely independent copy of antituberculosis rows and reset its index
+    msk = tb["antimicrobialclass"] == "Drugs for the treatment of tuberculosis (ATC J04A)"
+    tb_anti_tb = tb[msk].reset_index(drop=True)
+    assert len(tb_anti_tb["antimicrobialclass"].unique()) == 1
+
+    # Modify antimicrobialclass in tb
+    tb["antimicrobialclass"] = tb["antimicrobialclass"].replace(
+        {
+            "Drugs for the treatment of tuberculosis (ATC J04A)": "Antibacterials (ATC J01, A07AA, P01AB, ATC J04A)",
+            "Antibacterials (ATC J01, A07AA, P01AB)": "Antibacterials (ATC J01, A07AA, P01AB, ATC J04A)",
+        },
+    )
+    assert len(tb["antimicrobialclass"].unique()) == 4
+    # Format the notes tables before it's removed
+    tb_notes = tb[["country", "year", "antimicrobialclass", "notes"]].dropna(subset=["notes"])
+    tb_notes = format_notes(tb_notes)
+
+    # Aggregate the data
+    tb = tb.groupby(["country", "year", "antimicrobialclass"], dropna=False)[["ddd", "did"]].sum().reset_index()
+    assert len(tb["antimicrobialclass"].unique()) == 4
+    # Add the antituberculosis data back to tb
+    tb_anti_tb = (
+        tb_anti_tb.groupby(["country", "year", "antimicrobialclass"], dropna=False)[["ddd", "did"]].sum().reset_index()
+    )
+    tb_combined = pr.concat([tb, tb_anti_tb])
+
+    tb_combined.set_index(["country", "year", "antimicrobialclass"], verify_integrity=True)
+
+    return tb_combined, tb_notes
+
+
+def format_notes(tb_notes: Table) -> Table:
+    """
+    Format notes column
+    """
+    for note in tb_notes["notes"].unique():
+        if pd.notna(note):
+            msk = tb_notes["notes"] == note
+            tb_note = tb_notes[msk]
+            countries = tb_note["country"].unique()
+            countries_formatted = combine_countries(countries)
+            description_processing_string = f"- In {countries_formatted}: {note}\n"
+            tb_notes.loc[msk, "description_processing"] = description_processing_string
+    # Creating onedescription processing for each antimicrobial class, the variable unit
+    tb_desc = (
+        tb_notes.dropna(subset=["description_processing"])  # Remove NaNs
+        .groupby(["antimicrobialclass"])["description_processing"]
+        .apply(lambda x: "; ".join(set(x)))  # Combine unique values
+        .reset_index()
+    )
+    # tb = pr.merge(tb, tb_desc, on=["antimicrobialclass"])
+
+    return tb_desc
+
+
+def combine_countries(countries):
+    # Combine countries into a string
+    if not countries:
+        return ""
+    elif len(countries) == 1:
+        return countries[0]
+    elif len(countries) == 2:
+        return " and ".join(countries)
+    else:
+        return ", ".join(countries[:-1]) + " and " + countries[-1]

etl/steps/data/grapher/antibiotics/2024-11-12/antimicrobial_usage.py

@@ -16,6 +16,7 @@ def run(dest_dir: str) -> None:
     # Read table from garden dataset.
     tb_class = ds_garden["class"]
     tb_aware = ds_garden["aware"]
+    tb_class_agg = ds_garden["class_aggregated"]
 
     #
     # Process data.
@@ -26,7 +27,10 @@ def run(dest_dir: str) -> None:
     #
     # Create a new grapher dataset with the same metadata as the garden dataset.
     ds_grapher = create_dataset(
-        dest_dir, tables=[tb_class, tb_aware], check_variables_metadata=True, default_metadata=ds_garden.metadata
+        dest_dir,
+        tables=[tb_class, tb_aware, tb_class_agg],
+        check_variables_metadata=True,
+        default_metadata=ds_garden.metadata,
     )
 
     # Save changes in the new grapher dataset.