Add config file and add helper functions

analysis/config.py

@@ -0,0 +1,15 @@
+# measures_definition_pf_breakdown.py
+start_date_measure1 = "2023-11-01"
+monthly_intervals_measure1 = 12
+
+# measures_definition_pf_condition_provider.py
+start_date_measure2 = "2023-01-01"
+monthly_intervals_measure2 = 22
+
+# measures_definition_pf_descriptive_stats.py
+start_date_measure3 = "2024-02-01"
+monthly_intervals_measure3 = 9
+
+# measures_definition_pf_medications.py
+start_date_measure4 = "2023-11-01"
+monthly_intervals_measure4 = 9

analysis/measures_definition_pf_breakdown.py

@@ -13,12 +13,14 @@
 
 from pf_dataset import get_latest_ethnicity
 from codelists import pharmacy_first_event_codes, pharmacy_first_consultation_codelist
+from config import start_date_measure1, monthly_intervals_measure1
+from pf_variables_library import get_events_between, get_events
 
 measures = create_measures()
 measures.configure_dummy_data(population_size=1000)
 
-start_date = "2023-11-01"
-monthly_intervals = 12
+start_date = start_date_measure1
+monthly_intervals = monthly_intervals_measure1
 
 registration = practice_registrations.for_patient_on(INTERVAL.end_date)
 ethnicity_combined = get_latest_ethnicity(
@@ -57,16 +59,22 @@
     otherwise="Missing",
 )
 
-pharmacy_first_ids = clinical_events.where(
-    clinical_events.snomedct_code.is_in(
-        pharmacy_first_consultation_codelist
-    )
-).consultation_id
+# pharmacy_first_ids = clinical_events.where(
+#     clinical_events.snomedct_code.is_in(
+#         pharmacy_first_consultation_codelist
+#     )
+# ).consultation_id
+
+pharmacy_first_ids = get_events(clinical_events, pharmacy_first_consultation_codelist).consultation_id
 
-# Select clinical events in interval date range
-selected_events = clinical_events.where(
-    clinical_events.date.is_on_or_between(INTERVAL.start_date, INTERVAL.end_date)
-).where(clinical_events.consultation_id.is_in(pharmacy_first_ids))
+# # Select clinical events in interval date range
+# selected_events = clinical_events.where(
+#     clinical_events.date.is_on_or_between(INTERVAL.start_date, INTERVAL.end_date)
+# ).where(clinical_events.consultation_id.is_in(pharmacy_first_ids))
+
+selected_events = get_events_between(clinical_events, INTERVAL.start_date, INTERVAL.end_date).where(
+    clinical_events.consultation_id.is_in(pharmacy_first_ids)
+)
 
 # Breakdown metrics to be produced as graphs
 breakdown_metrics = {

analysis/measures_definition_pf_condition_provider.py

@@ -9,25 +9,31 @@
     imd_quintile,
 )
 from codelists import pharmacy_first_consultation_codelist
+from config import start_date_measure2, monthly_intervals_measure2
+from pf_variables_library import get_events_between, get_events
 
 measures = create_measures()
 measures.configure_dummy_data(population_size=1000)
 
-start_date = "2023-01-01"
-monthly_intervals = 22
+start_date = start_date_measure2
+monthly_intervals = monthly_intervals_measure2
 
 registration = practice_registrations.for_patient_on(INTERVAL.end_date)
 
-selected_events = clinical_events.where(
-    clinical_events.date.is_on_or_between(INTERVAL.start_date, INTERVAL. end_date)
-)
+# selected_events = clinical_events.where(
+#     clinical_events.date.is_on_or_between(INTERVAL.start_date, INTERVAL. end_date)
+# )
+
+selected_events = get_events_between(clinical_events, INTERVAL.start_date, INTERVAL.end_date)
 
 # Create variable which contains boolean values of whether pharmacy first event exists for patient
-has_pharmacy_first = selected_events.where(
-    selected_events.snomedct_code.is_in(
-        pharmacy_first_consultation_codelist
-    )
-).exists_for_patient()
+# has_pharmacy_first = selected_events.where(
+#     selected_events.snomedct_code.is_in(
+#         pharmacy_first_consultation_codelist
+#     )
+# ).exists_for_patient()
+
+has_pharmacy_first = get_events(selected_events, pharmacy_first_consultation_codelist).exists_for_patient()
 
 for condition_name, condition_code in pharmacy_first_conditions_codes.items():
     condition_events = selected_events.where(

analysis/measures_definition_pf_descriptive_stats.py

@@ -2,26 +2,27 @@
 from ehrql.tables.raw.tpp import medications
 from ehrql.tables.tpp import practice_registrations, patients, clinical_events
 
-from pf_variables_library import get_consultation_ids, get_consultationid_events
+from pf_variables_library import get_events, get_consultationid_events
 from codelists import (
     pharmacy_first_med_codelist,
     pharmacy_first_consultation_codelist,
     pharmacy_first_conditions_codelist,
 )
+from config import start_date_measure3, monthly_intervals_measure3
 
 measures = create_measures()
 measures.configure_dummy_data(population_size=1000)
 measures.configure_disclosure_control(enabled=True)
 
-start_date = "2024-02-01"
-monthly_intervals = 9
+start_date = start_date_measure3
+monthly_intervals = monthly_intervals_measure3
 
 registration = practice_registrations.for_patient_on(INTERVAL.end_date)
 
 # Function to retrieve consultation ids from clinical events that are PF consultations
-pharmacy_first_ids = get_consultation_ids(
+pharmacy_first_ids = get_events(
     clinical_events, pharmacy_first_consultation_codelist
-)
+).consultation_id
 
 # Function to retrieve selected events using pharmacy first ids
 selected_clinical_events = get_consultationid_events(
@@ -33,14 +34,18 @@
 )
 
 # Create variable which contains boolean values of whether pharmacy first event exists for patient
-has_pf_consultation = selected_clinical_events.where(
-    selected_clinical_events.snomedct_code.is_in(pharmacy_first_consultation_codelist)
-).exists_for_patient()
+# has_pf_consultation = selected_clinical_events.where(
+#     selected_clinical_events.snomedct_code.is_in(pharmacy_first_consultation_codelist)
+# ).exists_for_patient()
+
+has_pf_consultation = get_events(selected_clinical_events, pharmacy_first_consultation_codelist).exists_for_patient()
 
 # PF consultations with PF clinical condition
-has_pf_condition = selected_clinical_events.where(
-    selected_clinical_events.snomedct_code.is_in(pharmacy_first_conditions_codelist)
-).exists_for_patient()
+# has_pf_condition = selected_clinical_events.where(
+#     selected_clinical_events.snomedct_code.is_in(pharmacy_first_conditions_codelist)
+# ).exists_for_patient()
+
+has_pf_condition = get_events(selected_clinical_events, pharmacy_first_conditions_codelist).exists_for_patient()
 
 # PF consultations with prescribed PF medication
 has_pf_medication = selected_med_events.where(

analysis/measures_definition_pf_medications.py

@@ -5,34 +5,44 @@
     practice_registrations,
 )
 from ehrql.tables.raw.tpp import medications
-from codelists import *
+from codelists import pharmacy_first_consultation_codelist, pharmacy_first_med_codelist
+from config import start_date_measure4, monthly_intervals_measure4
+from pf_variables_library import get_events_between
 
 measures = create_measures()
 measures.configure_dummy_data(population_size=1000)
 # Turn off during code development, but turn on before running against on the server
 measures.configure_disclosure_control(enabled=True)
 
-start_date = "2023-11-01"
-monthly_intervals = 9
+start_date = start_date_measure4
+monthly_intervals = monthly_intervals_measure4
 
 registration = practice_registrations.for_patient_on(INTERVAL.end_date)
 
 # Select Pharmacy First consultations during interval date range
-pharmacy_first_events = clinical_events.where(
-    clinical_events.date.is_on_or_between(INTERVAL.start_date, INTERVAL.end_date)
-).where(
-    clinical_events.snomedct_code.is_in(
-        pharmacy_first_consultation_codelist
-    )
+# pharmacy_first_events = clinical_events.where(
+#     clinical_events.date.is_on_or_between(INTERVAL.start_date, INTERVAL.end_date)
+# ).where(
+#     clinical_events.snomedct_code.is_in(
+#         pharmacy_first_consultation_codelist
+#     )
+# )
+
+pharmacy_first_events = get_events_between(clinical_events, INTERVAL.start_date, INTERVAL.end_date).where(
+    clinical_events.snomedct_code.is_in(pharmacy_first_consultation_codelist)
 )
 
 pharmacy_first_ids = pharmacy_first_events.consultation_id
 has_pharmacy_first_consultation = pharmacy_first_events.exists_for_patient()
 
 # Select medications prescribed with PF consultation ID
-selected_medications = medications.where(
-    medications.date.is_on_or_between(INTERVAL.start_date, INTERVAL.end_date)
-).where(medications.consultation_id.is_in(pharmacy_first_ids))
+# selected_medications = medications.where(
+#     medications.date.is_on_or_between(INTERVAL.start_date, INTERVAL.end_date)
+# ).where(medications.consultation_id.is_in(pharmacy_first_ids))
+
+selected_medications = get_events_between(medications, INTERVAL.start_date, INTERVAL.end_date).where(
+    medications.consultation_id.is_in(pharmacy_first_ids)
+)
 
 # Select first medication for group_by argument in measures
 first_selected_medication = (

analysis/pf_dataset.py

@@ -2,6 +2,9 @@
 
 from pf_variables_library import check_pregnancy_status, count_past_events
 
+# This file contains functions for the denominators of the patient population for each clinical condition.
+# It will be used to calculate rates, and is separate from pf_variables_library
+
 # Create denominator variables for each clinical condition
 # These are based on NHS England rules using sex, age, pregnancy status and repeated diagnoses
 # NOTE: The following exclusions have not been added:

analysis/pf_variables_library.py

@@ -1,7 +1,6 @@
 # Function to check status of a condition within a specified time window
 from ehrql import months
 
-
 def check_pregnancy_status(index_date, selected_events, codelist):
     return (
         selected_events.where(selected_events.snomedct_code.is_in(codelist))
@@ -11,7 +10,6 @@ def check_pregnancy_status(index_date, selected_events, codelist):
         .exists_for_patient()
     )
 
-
 # Function to count number of coded events within a specified time window
 def count_past_events(index_date, selected_events, codelist, num_months):
     return (
@@ -24,19 +22,24 @@ def count_past_events(index_date, selected_events, codelist, num_months):
         .count_for_patient()
     )
 
+# Function to get events linked to a specified codelist
+def get_events(event_frame, codelist):
+    selected_events = event_frame.where(
+        event_frame.snomedct_code.is_in(codelist)
+    )
 
-# Function to get consultation IDs linked to a specified codelist
-def get_consultation_ids(clinical_events, codelist):
-    consultation_ids = clinical_events.where(
-        clinical_events.snomedct_code.is_in(codelist)
-    ).consultation_id
-
-    return consultation_ids
-
+    return selected_events
 
 # Function to get events with specific consultation IDs
 def get_consultationid_events(event_frame, consultation_ids):
     selected_events = event_frame.where(
         event_frame.consultation_id.is_in(consultation_ids)
     )
     return selected_events
+
+# Function to get events within a time frame
+def get_events_between(event_frame, start_date, end_date):
+    selected_events = event_frame.where(
+        event_frame.date.is_on_or_between(start_date, end_date)
+    )
+    return selected_events