update data processing method

analysis/codelists.py

@@ -69,8 +69,8 @@
   column = "code",
 )
 copd_medications = codelist_from_csv(
-  "codelists/bristol-copd-medications-bnf.csv",
-  column = "code",
+  "codelists/user-emprestige-copd-medications-new-dmd.csv",
+  column = "dmd_id",
 )
 
 # pulmonary fibrosis
@@ -131,6 +131,12 @@
 
 # rsv primary - sensitive 
 
+# rsv primary - sensitive (prescriptions)
+rsv_prescriptions_codelist = codelist_from_csv(
+  "codelists/user-emprestige-rsv-identification-prescriptions-maximal-sensitivity-dmd.csv",
+  column = "dmd_id",
+)
+
 # rsv secondary - sensitive
 
 # covid primary - sensitive
@@ -139,6 +145,12 @@
   column = "CTV3ID",
 )
 
+# covid primary - sensitive (prescriptions)
+covid_prescriptions_codelist = codelist_from_csv(
+  "codelists/user-emprestige-covid-19-identification-prescriptions-dmd.csv",
+  column = "dmd_id",
+)
+
 # covid secondary - sensitive
 covid_secondary_codelist = codelist_from_csv(
   "codelists/opensafely-covid-identification.csv",
@@ -147,12 +159,18 @@
 
 # flu primary - sensitive
 
+# flu primary - sensitive (prescriptions)
+flu_prescriptions_codelist = codelist_from_csv(
+  "codelists/user-emprestige-influenza-identification-prescriptions-maximal-sensitivity-dmd.csv",
+  column = "dmd_id",
+)
+
 # flu secondary - sensitive
 
 ##exclusion criteria
 
 # care home 
-care_home_codelist = codelist_from_csv(
+carehome_codelist = codelist_from_csv(
   "codelists/nhsd-primary-care-domain-refsets-carehome_cod.csv",
   column = "code",
 )

analysis/data_processing.R

@@ -0,0 +1,81 @@
+library("tidyverse")
+library("here")
+library("arrow")
+library("ggplot2")
+library("data.table")
+library("lubridate")
+
+## create output directories ----
+fs::dir_create(here("analysis"))
+
+#define study start date and study end date
+source(here("analysis", "design", "design.R"))
+args <- commandArgs(trailingOnly = TRUE)
+study_start_date <- as.Date(as.numeric(study_dates[args[2]]), format = "%Y-%m-%d", origin = "1970-01-01")
+study_end_date <- as.Date(as.numeric(study_dates[args[3]]), format = "%Y-%m-%d", origin = "1970-01-01")
+
+df_input <- read_feather(
+  here::here("output", paste0("input_", args[[1]], "_", year(study_start_date),
+                              "_", year(study_end_date), ".arrow")))
+
+#assign ethnicity group
+df_input <- df_input %>%
+  mutate(
+    latest_ethnicity_group = ifelse(df_input$latest_ethnicity_code == "1", "White",
+                             ifelse(df_input$latest_ethnicity_code == "2", "Mixed",
+                             ifelse(df_input$latest_ethnicity_code == "3", "Asian or Asian British",
+                             ifelse(df_input$latest_ethnicity_code == "4", "Black or Black British",
+                             ifelse(df_input$latest_ethnicity_code == "5", "Other Ethnic Groups", "Unknown"))))
+    ))
+
+#calculate age bands
+if(args[[1]] == "older_adults") {
+  df_input <- df_input %>%
+    mutate(age_band = case_when(
+      df_input$age >= 65 & df_input$age <= 74 ~ "65-74y",
+      df_input$age >= 75 & df_input$age <= 89 ~ "75-89y",
+      df_input$age >= 90 ~ "90+y",
+      TRUE ~ NA_character_
+    ))
+} else if(args[[1]] == "adults") {
+df_input <- df_input %>%
+  mutate(age_band = case_when(
+    df_input$age >= 18 & df_input$age <= 39 ~ "18-29y",
+    df_input$age >= 40 & df_input$age <= 64 ~ "40-64y",
+    TRUE ~ NA_character_
+  ))
+} else if(args[[1]] == "children_adults") {
+  df_input <- df_input %>%
+    mutate(age_band = case_when(
+      df_input$age >= 2 & df_input$age <= 5 ~ "2-5y",
+      df_input$age >= 6 & df_input$age <= 9 ~ "6-9y",
+      df_input$age >= 10 & df_input$age <= 13 ~ "10-13y",
+      df_input$age >= 14 & df_input$age <= 17 ~ "14-17y",
+      TRUE ~ NA_character_
+    ))
+} else {
+  df_input <- df_input %>%
+    mutate(age_band = case_when(
+      df_input$age >= 0 & df_input$age <= 2 ~ "0-2m",
+      df_input$age >= 3 & df_input$age <= 5 ~ "3-5m",
+      df_input$age >= 6 & df_input$age <= 11 ~ "6-11m",
+      df_input$age >= 12 & df_input$age <= 23 ~ "12-23m",
+      TRUE ~ NA_character_
+    ))
+}
+
+#calculate IMD quintile
+df_input <- df_input %>%
+  mutate(imd_quintile = case_when(
+    df_input$imd_rounded >= 0 & df_input$imd_rounded < as.integer(32800 * 1 / 5) ~ "1",
+    df_input$imd_rounded < as.integer(32800 * 2 / 5) ~ "2",
+    df_input$imd_rounded < as.integer(32800 * 3 / 5) ~ "3",
+    df_input$imd_rounded < as.integer(32800 * 4 / 5) ~ "4",
+    df_input$imd_rounded < as.integer(32800 * 5 / 5) ~ "5 (least deprived)",
+    TRUE ~ NA_character_
+  ))
+
+#write the new input file
+write_feather(df_input, here::here("output", 
+  paste0("input_processed_", args[[1]], "_", year(study_start_date),
+         "_", year(study_end_date), ".arrow")))

analysis/dataset_definition.py

@@ -69,6 +69,7 @@
 dataset.registered = registered_patients
 dataset.sex = patients.sex
 dataset.age = case(
+  when(args[1] == "older_adults").then(age_at_start),
   when(args[1] == "adults").then(age_at_start),
   when(args[1] == "children_adolescents").then(age_at_start),
   when(args[1] == "infants").then(age_at_start_months),
@@ -109,14 +110,15 @@
 #medication date
 medication_date = index_date - years(1)
 
-#has asthma if there is a recent asthma diagnosis and a medication prescribed 
-dataset.has_asthma = (
-  clinical_events.where(clinical_events.snomedct_code.is_in(codelists.asthma_codelist))
-  .exists_for_patient() & medications.where(medications.dmd_code
-  .is_in(codelists.asthma_medications))
-  .where(medications.date.is_on_or_between(medication_date, index_date))
-  .exists_for_patient()
-)
+#has asthma if there is an asthma diagnosis and a recent medication prescribed 
+if args[1] != "infants" and args[1] != "infants_subgroup" :
+  dataset.has_asthma = (
+    clinical_events.where(clinical_events.snomedct_code.is_in(codelists.asthma_codelist))
+    .exists_for_patient() & medications.where(medications.dmd_code
+    .is_in(codelists.asthma_medications))
+    .where(medications.date.is_on_or_between(medication_date, index_date))
+    .exists_for_patient()
+  )
 
 #reactive airway disease diagnosis 
 if args[1] == "children_adolescents" : 
@@ -307,5 +309,6 @@ def has_prior_event(codelist, where=True):
 
 #care home resident
 if args[1] == "older_adults" :
-  dataset.care_home_tpp = (addresses.care_home_is_potential_match.if_null_then(False))
+  dataset.care_home_tpp = (addresses.for_patient_on(index_date)
+  .care_home_is_potential_match.when_null_then(False))
   dataset.care_home_code = (has_prior_event(codelists.carehome_codelist))

analysis/dummydata/dummydata_infants.R

@@ -191,7 +191,7 @@ sim_list = lst(
   ##exclusion criteria 
 
   #severe combined immunodeficiency syndrome 
-  severe_immmunodeficiency = bn_node(
+  severe_immunodeficiency = bn_node(
     ~ rbernoulli(n = ..n, p = 0.02)
   ),
 

codelists/codelists.json

@@ -155,6 +155,30 @@
       "url": "https://www.opencodelists.org/codelist/nhsd-primary-care-domain-refsets/tempcarhome_cod/20200812/",
       "downloaded_at": "2024-02-09 10:54:25.774557Z",
       "sha": "931872c7b2c96884a1b4bdb5062356985875cea4"
+    },
+    "user-emprestige-rsv-identification-prescriptions-maximal-sensitivity-dmd.csv": {
+      "id": "user/emprestige/rsv-identification-prescriptions-maximal-sensitivity-dmd/65062298",
+      "url": "https://www.opencodelists.org/codelist/user/emprestige/rsv-identification-prescriptions-maximal-sensitivity-dmd/65062298/",
+      "downloaded_at": "2024-02-12 11:06:07.464507Z",
+      "sha": "3df89edeab0e2e010cda91cf8f00c03cd91c739d"
+    },
+    "user-emprestige-influenza-identification-prescriptions-maximal-sensitivity-dmd.csv": {
+      "id": "user/emprestige/influenza-identification-prescriptions-maximal-sensitivity-dmd/4bf6af63",
+      "url": "https://www.opencodelists.org/codelist/user/emprestige/influenza-identification-prescriptions-maximal-sensitivity-dmd/4bf6af63/",
+      "downloaded_at": "2024-02-12 11:06:07.639146Z",
+      "sha": "000e8053fa45aa912ac09bd9e7c51758fe69e04e"
+    },
+    "user-emprestige-covid-19-identification-prescriptions-dmd.csv": {
+      "id": "user/emprestige/covid-19-identification-prescriptions-dmd/32e73c2e",
+      "url": "https://www.opencodelists.org/codelist/user/emprestige/covid-19-identification-prescriptions-dmd/32e73c2e/",
+      "downloaded_at": "2024-02-12 11:13:31.540363Z",
+      "sha": "7c17a141fc9c02fa85d885fa6c21bc398e9a957e"
+    },
+    "user-emprestige-copd-medications-new-dmd.csv": {
+      "id": "user/emprestige/copd-medications-new-dmd/1f3d48e5",
+      "url": "https://www.opencodelists.org/codelist/user/emprestige/copd-medications-new-dmd/1f3d48e5/",
+      "downloaded_at": "2024-02-13 11:49:39.628031Z",
+      "sha": "8ca158de8df88cb9d1a961792c89ed650f5bd445"
     }
   }
 }

codelists/codelists.txt

@@ -23,4 +23,8 @@ opensafely/covid-identification/2020-06-03
 nhsd-primary-care-domain-refsets/carehome_cod/20211221
 opensafely/chronic-cardiac-disease/2020-04-08
 nhsd-primary-care-domain-refsets/pulmohyp_cod/20210127
-nhsd-primary-care-domain-refsets/tempcarhome_cod/20200812
+nhsd-primary-care-domain-refsets/tempcarhome_cod/20200812
+user/emprestige/rsv-identification-prescriptions-maximal-sensitivity-dmd/65062298
+user/emprestige/influenza-identification-prescriptions-maximal-sensitivity-dmd/4bf6af63
+user/emprestige/covid-19-identification-prescriptions-dmd/32e73c2e
+user/emprestige/copd-medications-new-dmd/1f3d48e5/