Reference pop data

analysis/000_baseline_data_midpoint.do

@@ -0,0 +1,192 @@
+version 16
+
+/*==============================================================================
+DO FILE NAME:			Incidence graphs
+PROJECT:				OpenSAFELY Disease Incidence project
+DATE: 					23/08/2024
+AUTHOR:					J Galloway / M Russell									
+DESCRIPTION OF FILE:	Baseline data for reference population
+DATASETS USED:			Dataset definition
+OTHER OUTPUT: 			logfiles, printed to folder $Logdir
+USER-INSTALLED ADO: 	 
+  (place .ado file(s) in analysis folder)						
+==============================================================================*/
+
+*Set filepaths
+*global projectdir "C:\Users\Mark\OneDrive\PhD Project\OpenSAFELY Incidence\disease_incidence"
+*global projectdir "C:\Users\k1754142\OneDrive\PhD Project\OpenSAFELY Incidence\disease_incidence"
+global projectdir `c(pwd)'
+di "$projectdir"
+
+capture mkdir "$projectdir/output/data"
+capture mkdir "$projectdir/output/tables"
+capture mkdir "$projectdir/output/figures"
+
+global logdir "$projectdir/logs"
+di "$logdir"
+
+*Open a log file
+cap log close
+log using "$logdir/baseline_data_midpoint.log", replace
+
+*Set Ado file path
+adopath + "$projectdir/analysis/extra_ados"
+
+*Import dataset
+import delimited "$projectdir/output/dataset_definition_midpoint.csv", clear
+
+set scheme plotplainblind
+
+*Create and label variables ===========================================================*/
+
+**Age
+lab var age "Age"
+codebook age
+keep if age !=.
+
+**Sex
+gen gender = 1 if sex == "female"
+replace gender = 2 if sex == "male"
+lab var gender "Gender"
+lab define gender 1 "Female" 2 "Male", modify
+lab val gender gender
+tab gender, missing
+keep if gender !=.
+drop sex
+
+**Ethnicity
+gen ethnicity_n = 1 if ethnicity == "White"
+replace ethnicity_n = 2 if ethnicity == "Asian or Asian British"
+replace ethnicity_n = 3 if ethnicity == "Black or Black British"
+replace ethnicity_n = 4 if ethnicity == "Mixed"
+replace ethnicity_n = 5 if ethnicity == "Chinese or Other Ethnic Groups"
+replace ethnicity_n = 6 if ethnicity == "Unknown"
+
+
+label define ethnicity_n	1 "White"  						///
+							2 "Asian or Asian British"		///
+							3 "Black or Black British"  	///
+							4 "Mixed"						///
+							5 "Chinese or Other Ethnic Groups" ///
+							6 "Unknown", modify
+
+label values ethnicity_n ethnicity_n
+lab var ethnicity_n "Ethnicity"
+tab ethnicity_n, missing
+drop ethnicity
+rename ethnicity_n ethnicity
+
+**IMD
+gen imd = 1 if imd_quintile == "1 (most deprived)"
+replace imd = 2 if imd_quintile == "2"
+replace imd = 3 if imd_quintile == "3"
+replace imd = 4 if imd_quintile == "4"
+replace imd = 5 if imd_quintile == "5 (least deprived)"
+replace imd = 6 if imd_quintile == "Unknown"
+
+label define imd 1 "1 (most deprived)" 2 "2" 3 "3" 4 "4" 5 "5 (least deprived)" 6 "Unknown", modify
+label values imd imd 
+lab var imd "Index of multiple deprivation"
+tab imd, missing
+drop imd_quintile
+
+save "$projectdir/output/data/reference_data_processed.dta", replace
+
+/*Tables================================================================*/
+
+use "$projectdir/output/data/reference_data_processed.dta", clear
+
+**Baseline table for reference population
+preserve
+table1_mc, total(before) onecol nospacelowpercent missing iqrmiddle(",")  ///
+	vars(age contn %5.1f \ ///
+		 gender cat %5.1f \ ///
+		 ethnicity cat %5.1f \ ///
+		 imd cat %5.1f \ ///
+		 )
+restore
+
+**Rounded and redacted baseline table for full population
+clear *
+save "$projectdir/output/data/reference_table_rounded.dta", replace emptyok
+use "$projectdir/output/data/reference_data_processed.dta", clear
+
+set type double
+
+foreach var of varlist imd ethnicity gender  {
+	preserve
+	contract `var'
+	local v : variable label `var' 
+	gen variable = `"`v'"'
+    decode `var', gen(categories)
+	gen count = round(_freq, 5)
+	egen total = total(count)
+	gen percent = round((count/total)*100, 0.1)
+	order total, after(percent)
+	gen countstr = string(count)
+	replace countstr = "<8" if count<=7
+	order countstr, after(count)
+	drop count
+	rename countstr count
+	tostring percent, gen(percentstr) force format(%9.1f)
+	replace percentstr = "-" if count =="<8"
+	order percentstr, after(percent)
+	drop percent
+	rename percentstr percent
+	gen totalstr = string(total)
+	replace totalstr = "-" if count =="<8"
+	order totalstr, after(count)
+	drop total
+	rename totalstr total
+	gen cohort = "All"
+	order cohort, first
+	list cohort variable categories count percent total
+	keep cohort variable categories count percent total
+	append using "$projectdir/output/data/reference_table_rounded.dta"
+	save "$projectdir/output/data/reference_table_rounded.dta", replace
+	restore
+}
+use "$projectdir/output/data/reference_table_rounded.dta", clear
+export excel "$projectdir/output/tables/reference_table_rounded.xls", replace sheet("Overall") keepcellfmt firstrow(variables)
+
+**Table of mean age
+clear *
+save "$projectdir/output/data/reference_mean_age_rounded.dta", replace emptyok
+use "$projectdir/output/data/reference_data_processed.dta", clear
+
+	preserve
+	collapse (count) count=age (mean) mean_age=age (sd) stdev_age=age
+	rename *count freq
+	gen count = round(freq, 5)
+	gen countstr = string(count)
+	replace countstr = "<8" if count<=7
+	order countstr, after(count)
+	drop count
+	rename countstr count
+	tostring mean_age, gen(meanstr) force format(%9.1f)
+	replace meanstr = "-" if count =="<8"
+	drop mean_age
+	rename meanstr mean_age
+	tostring stdev_age, gen(stdevstr) force format(%9.1f)
+	replace stdevstr = "-" if count =="<8"
+	order stdevstr, after(stdev_age)
+	drop stdev_age
+	rename stdevstr stdev_age
+	order count, first
+	list count mean_age stdev_age
+	keep count mean_age stdev_age
+	append using "$projectdir/output/data/reference_mean_age_rounded.dta"
+	save "$projectdir/output/data/reference_mean_age_rounded.dta", replace	
+	restore
+
+use "$projectdir/output/data/reference_mean_age_rounded.dta", clear
+export excel "$projectdir/output/tables/reference_mean_age_rounded.xls", replace keepcellfmt firstrow(variables)
+
+***Output tables as CSVs		 
+import excel "$projectdir/output/tables/reference_table_rounded.xls", clear
+export delimited using "$projectdir/output/tables/reference_table_rounded.csv", novarnames  replace
+
+import excel "$projectdir/output/tables/reference_mean_age_rounded.xls", clear
+export delimited using "$projectdir/output/tables/reference_mean_age_rounded.csv", novarnames  replace	
+
+log close	

analysis/001_baseline_data.do

@@ -46,18 +46,23 @@ set scheme plotplainblind
 
 *Create and label variables ===========================================================*/
 
+**Age
+codebook age_reg
+keep if age_reg !=.
+
 **Sex
 gen gender = 1 if sex == "female"
 replace gender = 2 if sex == "male"
 lab var gender "Gender"
 lab define gender 1 "Female" 2 "Male", modify
 lab val gender gender
 tab gender, missing
+keep if gender !=.
 drop sex
 
 **Ethnicity
 gen ethnicity_n = 1 if ethnicity == "White"
-replace ethnicity_n = 2 if ethnicity == "Asian or Asian British" /* mixed to 6 */
+replace ethnicity_n = 2 if ethnicity == "Asian or Asian British"
 replace ethnicity_n = 3 if ethnicity == "Black or Black British"
 replace ethnicity_n = 4 if ethnicity == "Mixed"
 replace ethnicity_n = 5 if ethnicity == "Chinese or Other Ethnic Groups"
@@ -150,7 +155,7 @@ table1_mc, total(before) onecol nospacelowpercent missing iqrmiddle(",")  ///
 		 )
 restore
 
-**Baseline table for reference population
+**Baseline table by disease
 foreach disease in $diseases {
 	preserve
 	keep if `disease'==1

analysis/dataset_definition_midpoint.py

@@ -0,0 +1,52 @@
+from ehrql import create_dataset, days, months, years, case, when, create_measures, INTERVAL, minimum_of, maximum_of
+from ehrql.tables.tpp import patients, medications, practice_registrations, clinical_events, apcs, addresses, ons_deaths, appointments
+from ehrql.codes import ICD10Code
+from datetime import date, datetime
+import codelists_ehrQL as codelists
+
+dataset = create_dataset()
+dataset.configure_dummy_data(population_size=1000)
+
+index_date = "2020-08-01"
+
+# Demographics
+dataset.age = patients.age_on(index_date)
+dataset.sex = patients.sex
+
+# Currently registered at mid-point
+any_registration = practice_registrations.for_patient_on(index_date).exists_for_patient()
+
+# Define patient ethnicity
+latest_ethnicity_code = (
+    clinical_events.where(clinical_events.snomedct_code.is_in(codelists.ethnicity_codes))
+    .where(clinical_events.date.is_on_or_before(index_date))
+    .sort_by(clinical_events.date)
+    .last_for_patient().snomedct_code.to_category(codelists.ethnicity_codes)
+)
+
+dataset.ethnicity = case(
+    when(latest_ethnicity_code == "1").then("White"),
+    when(latest_ethnicity_code == "2").then("Mixed"),
+    when(latest_ethnicity_code == "3").then("Asian or Asian British"),
+    when(latest_ethnicity_code == "4").then("Black or Black British"),
+    when(latest_ethnicity_code == "5").then("Chinese or Other Ethnic Groups"),
+    otherwise="Unknown",
+)
+
+# Define patient IMD
+imd = addresses.for_patient_on(index_date).imd_rounded
+
+dataset.imd_quintile = case(
+    when((imd >= 0) & (imd < int(32844 * 1 / 5))).then("1 (most deprived)"),
+    when(imd < int(32844 * 2 / 5)).then("2"),
+    when(imd < int(32844 * 3 / 5)).then("3"),
+    when(imd < int(32844 * 4 / 5)).then("4"),
+    when(imd < int(32844 * 5 / 5)).then("5 (least deprived)"),
+    otherwise="Unknown",
+)
+
+# Define population as any registered patient after index date - then apply further restrictions later
+dataset.define_population(
+    any_registration
+    & dataset.sex.is_in(["male", "female"]) 
+)  

generate_yaml.py

@@ -18,6 +18,13 @@
       highly_sensitive:
         cohort: output/dataset_definition.csv
 
+  generate_dataset_midpoint:
+    run: ehrql:v1 generate-dataset analysis/dataset_definition_midpoint.py
+      --output output/dataset_definition_midpoint.csv
+    outputs:
+      highly_sensitive:
+        cohort: output/dataset_definition_midpoint.csv        
+
   # generate_dataset_data_avail:
   #   run: ehrql:v1 generate-dataset analysis/dataset_definition_data_avail.py
   #     --output output/dataset_definition_data_avail.csv
@@ -35,6 +42,15 @@
   #       log1: logs/data_avail_tables.log   
   #       data1: output/tables/data_check_*.csv
 
+  run_baseline_data_midpoint:
+    run: stata-mp:latest analysis/000_baseline_data_midpoint.do
+    needs: [generate_dataset_midpoint]
+    outputs:
+      moderately_sensitive:
+        log1: logs/baseline_data_midpoint.log   
+        table1: output/tables/reference_table_rounded.csv
+        table2: output/tables/reference_mean_age_rounded.csv
+
   run_baseline_data:
     run: stata-mp:latest analysis/001_baseline_data.do
     needs: [generate_dataset]

project.yaml

@@ -15,6 +15,13 @@ actions:
       highly_sensitive:
         cohort: output/dataset_definition.csv
 
+  generate_dataset_midpoint:
+    run: ehrql:v1 generate-dataset analysis/dataset_definition_midpoint.py
+      --output output/dataset_definition_midpoint.csv
+    outputs:
+      highly_sensitive:
+        cohort: output/dataset_definition_midpoint.csv        
+
   # generate_dataset_data_avail:
   #   run: ehrql:v1 generate-dataset analysis/dataset_definition_data_avail.py
   #     --output output/dataset_definition_data_avail.csv
@@ -32,6 +39,15 @@ actions:
   #       log1: logs/data_avail_tables.log   
   #       data1: output/tables/data_check_*.csv
 
+  run_baseline_data_midpoint:
+    run: stata-mp:latest analysis/000_baseline_data_midpoint.do
+    needs: [generate_dataset_midpoint]
+    outputs:
+      moderately_sensitive:
+        log1: logs/baseline_data_midpoint.log   
+        table1: output/tables/reference_table_rounded.csv
+        table2: output/tables/reference_mean_age_rounded.csv
+
   run_baseline_data:
     run: stata-mp:latest analysis/001_baseline_data.do
     needs: [generate_dataset]