Merge pull request #17 from opensafely/update-apr-19

Update apr 19
opensafely · Apr 19, 2021 · 216deb4 · 216deb4
2 parents 473d352 + 3fa3faf
commit 216deb4
Show file tree

Hide file tree

Showing 73 changed files with 56,179 additions and 30,705 deletions.
diff --git a/analysis/codelists.py b/analysis/codelists.py
@@ -166,3 +166,7 @@
 wider_ld_codes = codelist_from_csv(
     "codelists/primis-covid19-vacc-uptake-learndis.csv", system="snomed", column="code"
 )
+
+covid_vacc_declined = codelist_from_csv(
+    "codelists/primis-covid19-vacc-uptake-cov1decl.csv", system="snomed", column="code"
+)
diff --git a/analysis/study_definition_delivery.py b/analysis/study_definition_delivery.py
@@ -166,30 +166,30 @@
         returning="binary_flag",
         return_expectations={"incidence": 0.01,},
     ),
-    cystic_fibrosis=patients.with_these_clinical_events(
-        cystic_fibrosis_codes,
-        on_or_before=index_date,
-        returning="binary_flag",
-        return_expectations={"incidence": 0.01,},
-    ),
-    sickle_cell_disease=patients.with_these_clinical_events(
-        sickle_cell_disease_codes,
-        on_or_before=index_date,
-        returning="binary_flag",
-        return_expectations={"incidence": 0.01,},
-    ),
-    permanant_immunosuppression=patients.with_these_clinical_events(
-        permanent_immunosuppression_codes,
-        on_or_before=index_date,
-        returning="binary_flag",
-        return_expectations={"incidence": 0.01,},
-    ),
-    temporary_immunosuppression=patients.with_these_clinical_events(
-        temporary_immunosuppression_codes,
-        on_or_before=index_date,
-        returning="binary_flag",
-        return_expectations={"incidence": 0.01,},
-    ),
+    # cystic_fibrosis=patients.with_these_clinical_events(
+    #     cystic_fibrosis_codes,
+    #     on_or_before=index_date,
+    #     returning="binary_flag",
+    #     return_expectations={"incidence": 0.01,},
+    # ),
+    # sickle_cell_disease=patients.with_these_clinical_events(
+    #     sickle_cell_disease_codes,
+    #     on_or_before=index_date,
+    #     returning="binary_flag",
+    #     return_expectations={"incidence": 0.01,},
+    # ),
+    # permanant_immunosuppression=patients.with_these_clinical_events(
+    #     permanent_immunosuppression_codes,
+    #     on_or_before=index_date,
+    #     returning="binary_flag",
+    #     return_expectations={"incidence": 0.01,},
+    # ),
+    # temporary_immunosuppression=patients.with_these_clinical_events(
+    #     temporary_immunosuppression_codes,
+    #     on_or_before=index_date,
+    #     returning="binary_flag",
+    #     return_expectations={"incidence": 0.01,},
+    # ),
     #
     psychosis_schiz_bipolar=patients.with_these_clinical_events(
         psychosis_schizophrenia_bipolar_affective_disease_codes,
@@ -199,12 +199,12 @@
     ),
 
     # https://github.com/opensafely/codelist-development/issues/4
-    asplenia=patients.with_these_clinical_events(
-        asplenia_codes,
-        on_or_before=index_date,
-        returning="binary_flag",
-        return_expectations={"incidence": 0.01,},
-    ),
+    # asplenia=patients.with_these_clinical_events(
+    #     asplenia_codes,
+    #     on_or_before=index_date,
+    #     returning="binary_flag",
+    #     return_expectations={"incidence": 0.01,},
+    # ),
 
     ###############################################################################
     # COVID VACCINATION

diff --git a/analysis/study_definition_delivery_common.py b/analysis/study_definition_delivery_common.py
@@ -30,27 +30,24 @@
             OR
             shielded
             OR
-            (care_home)
-            OR
             (LD)
         )
-        """
+        """,
+        registered=patients.registered_as_of(
+            index_date, 
+        ),
+        has_died=patients.died_from_any_cause(
+            on_or_before=index_date,
+            returning="binary_flag",
+        ),
     ),
 
     has_follow_up=patients.registered_with_one_practice_between(
         start_date="index_date - 12 months",
         end_date=index_date,
         return_expectations={"incidence": 0.90},
     ),
-    registered=patients.registered_as_of(
-        index_date, 
-        return_expectations={"incidence": 0.98},
-    ),
-    has_died=patients.died_from_any_cause(
-        on_or_before=index_date,
-        returning="binary_flag",
-        return_expectations={"incidence": 0.05},
-    ),
+
 
 
     ### PRIMIS care home flag 
@@ -94,37 +91,6 @@
             },
         },
     ),
-    # age bands for patients not in care homes (ie living in the community)
-    # this is used to define eligible groups not defined by clinical criteria
-    ageband_community=patients.categorised_as(
-        {
-            "care home" : "DEFAULT",
-            "16-29": """ age >= 16 AND age < 30 AND NOT care_home""",
-            "30-39": """ age >= 30 AND age < 40 AND NOT care_home""",
-            "40-49": """ age >= 40 AND age < 50 AND NOT care_home""",
-            "50-59": """ age >= 50 AND age < 60 AND NOT care_home""",
-            "60-64": """ age >= 60 AND age < 65 AND NOT care_home""",
-            "65-69": """ age >= 65 AND age < 70 AND NOT care_home""",
-            "70-79": """ age >= 70 AND age < 80 AND NOT care_home""",
-            "80+": """ age >=  80 AND age < 120 AND NOT care_home""",
-        },
-        return_expectations={
-            "rate": "universal",
-            "category": {
-                "ratios": {
-                    "care home":0.125,
-                    "16-29": 0.125,
-                    "30-39": 0.125,
-                    "40-49": 0.125,
-                    "50-59": 0.125,
-                    "60-64": 0.0625,
-                    "65-69": 0.0625,
-                    "70-79": 0.125,
-                    "80+": 0.125,
-                }
-            },
-        },
-    ),
 
         # 5 year age bands
     ageband_5yr=patients.categorised_as(
@@ -300,17 +266,7 @@
         returning="binary_flag", 
         return_expectations={"incidence": 0.01,},
     ),
-    adrenaline_pen=patients.with_these_medications(
-        adrenaline_pen,
-        on_or_after="index_date - 24 months",  # look for last two years
-        returning="binary_flag",
-        return_last_date_in_period=False,
-        include_month=False,
-        return_expectations={
-            "date": {"earliest": "2018-12-01", "latest": index_date},
-            "incidence": 0.001,
-        },
-    ),
+
 
 
     ### PRIMIS overall flag for shielded group
@@ -387,5 +343,20 @@
             wider_ld_codes, 
             return_expectations={"incidence": 0.02,},
     ),
+
+    # declined vaccination / vaccination course / first dose (not including declined second dose)
+    covid_vacc_declined_date = patients.with_these_clinical_events(
+        covid_vacc_declined,
+        returning="date",
+        find_first_match_in_period=True,
+        date_format = "YYYY-MM-DD",
+        return_expectations={
+            "date": {
+                "earliest": "2020-12-08",  # first vaccine administered on the 8/12
+                "latest": index_date,
+            },
+                "incidence":0.04
+        },
+    ),
 
 )
diff --git a/codelists/codelists.json b/codelists/codelists.json
@@ -174,17 +174,17 @@
       "downloaded_at": "2021-02-10 14:28:12.124536Z",
       "sha": "99f6c0787e17e16430549a4cfb5084612faba2ac"
     },
-    "user-peter-inglesby-eth2001.csv": {
-      "id": "user/peter-inglesby/eth2001/0f56fa3b",
-      "url": "https://codelists.opensafely.org/codelist/user/peter-inglesby/eth2001/0f56fa3b/",
-      "downloaded_at": "2021-02-10 14:28:12.230527Z",
-      "sha": "708774541b925cf190016e2e160eec31b02d2512"
-    },
     "primis-covid19-vacc-uptake-learndis.csv": {
       "id": "primis-covid19-vacc-uptake/learndis/v1",
       "url": "https://codelists.opensafely.org/codelist/primis-covid19-vacc-uptake/learndis/v1/",
       "downloaded_at": "2021-03-09 11:58:49.613669Z",
       "sha": "41f81fcb9fadb082a77cadd6cf60614d7c8d860f"
+    },
+    "primis-covid19-vacc-uptake-cov1decl.csv": {
+      "id": "primis-covid19-vacc-uptake/cov1decl/v1.1",
+      "url": "https://codelists.opensafely.org/codelist/primis-covid19-vacc-uptake/cov1decl/v1.1/",
+      "downloaded_at": "2021-04-19 13:00:27.222055Z",
+      "sha": "c8f6ef075bb2f267ee8524453967ba927703fe9d"
     }
   }
 }
diff --git a/codelists/codelists.txt b/codelists/codelists.txt
@@ -27,5 +27,5 @@ opensafely/sickle-cell-disease/2020-04-14
 opensafely/solid-organ-transplantation/2020-04-10
 opensafely/permanent-immunosuppression/2020-06-02
 opensafely/temporary-immunosuppression/2020-04-24
-user/peter-inglesby/eth2001/0f56fa3b
 primis-covid19-vacc-uptake/learndis/v1
+primis-covid19-vacc-uptake/cov1decl/v1.1
diff --git a/codelists/primis-covid19-vacc-uptake-cov1decl.csv b/codelists/primis-covid19-vacc-uptake-cov1decl.csv
@@ -0,0 +1,4 @@
+code,term
+1324721000000108,Severe acute respiratory syndrome coronavirus 2 vaccination dose declined
+1324741000000101,Severe acute respiratory syndrome coronavirus 2 vaccination first dose declined
+1324811000000107,Severe acute respiratory syndrome coronavirus 2 immunisation course declined
diff --git a/codelists/user-peter-inglesby-eth2001.csv b/codelists/user-peter-inglesby-eth2001.csv
diff --git a/lib/data_processing.py b/lib/data_processing.py
@@ -10,7 +10,7 @@
 
 
 
-def load_data(input_file='input_delivery.csv', input_path="output"):
+def load_data(input_file='input_delivery.csv.gz', input_path="output"):
     """
     This reads in a csv that must be in output/ directory and cleans the
     data ready for use in the graphs and tables
@@ -33,7 +33,7 @@ def load_data(input_file='input_delivery.csv', input_path="output"):
     """
 
     # import data and fill nulls with 0
-    df = pd.read_csv(os.path.join("..",input_path, input_file)).fillna(0)
+    df = pd.read_csv(os.path.join("..",input_path, input_file), compression='gzip').fillna(0)
 
     # fill unknown ethnicity from GP records with ethnicity from SUS (secondary care)
     df.loc[df["ethnicity"]==0, "ethnicity"] = df["ethnicity_6_sus"]
@@ -62,6 +62,9 @@ def load_data(input_file='input_delivery.csv', input_path="output"):
         covid_vacc_2nd = np.where(df["covid_vacc_second_dose_date"]!=0, 1, 0),
         covid_vacc_bin = np.where(df["covid_vacc_date"]!=0, 1, 0))
 
+    # declined - suppress if vaccine has been received
+    df["covid_vacc_declined_date"] = np.where(df["covid_vacc_date"]==0, df["covid_vacc_declined_date"], 0)
+
     # create an additional field for 2nd dose to use as a flag for each eligible group
     df["2nd_dose"] = df["covid_vacc_2nd"]
 
@@ -84,7 +87,7 @@ def load_data(input_file='input_delivery.csv', input_path="output"):
 
     # drop unnecssary columns or columns created for processing 
     df = df.drop(["imd","ethnicity_16", "ethnicity", 'ethnicity_6_sus',
-       'ethnicity_16_sus', "adrenaline_pen", "has_died", "has_follow_up"], 1)
+       'ethnicity_16_sus', "has_follow_up"], 1)
 
     # categorise into priority groups (similar to the national groups but not exactly the same)
     conditions = [
@@ -111,9 +114,7 @@ def load_data(input_file='input_delivery.csv', input_path="output"):
     for c in ["2nd_dose", "LD", "newly_shielded_since_feb_15", "dementia", 
           "chronic_cardiac_disease", "current_copd", "dialysis", "dmards","psychosis_schiz_bipolar",
          "solid_organ_transplantation", "chemo_or_radio", "intel_dis_incl_downs_syndrome","ssri",
-          "lung_cancer", "cancer_excl_lung_and_haem", "haematological_cancer", "bone_marrow_transplant",
-          "cystic_fibrosis", "sickle_cell_disease", "permanant_immunosuppression",
-          "temporary_immunosuppression", "asplenia"]:
+          "lung_cancer", "cancer_excl_lung_and_haem", "haematological_cancer"]:
           df[c] = np.where(df[c]==1, "yes", "no")
 
 
@@ -122,7 +123,7 @@ def load_data(input_file='input_delivery.csv', input_path="output"):
     df = df.merge(stps, left_on="stp", right_on="stp_id", how="left").rename(columns={"name":"stp_name"})
 
     # drop additional columns
-    df = df.drop(['registered', 'care_home', 'age',"stp_id", "ageband_community"], 1)  
+    df = df.drop(['care_home', 'age',"stp_id"], 1)  
 
     return df