updated read_data.R to return fp_adults as well

data_extraction_formatting_clinical.R

@@ -10,10 +10,15 @@ getwd()
 # Input: source data
 #====================
 source("read_data.R")
-
+source("reg_cols_extraction.R")
+########################################################################
+#==========
+# 
+#==========
 # extract the flu positive population
 fp_adults = adult_df[adult_df$flustat == 1,]
 
+########################################################################
 table(adult_df$ia_exac_copd)
 table(adult_df$ia_exac_copd==1 & adult_df$asthma == 1) # check this is 4 
 
@@ -22,37 +27,9 @@ table(fp_adults$ia_exac_copd==1 & fp_adults$asthma == 1) # check this is 3
 # clear unnecessary variables
 rm(all_df)
 rm(adult_df)
+
+
 ########################################################################
-cols_to_extract = c("mosaic"
-                    , "ia_exac_copd"
-                    , "death"
-                    #, "obese2" #inc peaeds, but once you subset data for adults, its the same!
-                    , "obesity"
-                    , "flustat"
-                    , "sfluv"
-                    , "h1n1v"
-                    , "age"
-                    , "gender"
-                    , "asthma"
-                    , "vl_pfu_ul_npa1"
-                    , "los"
-                    , "onset2final"
-                    , "onsfindeath"
-                    , "onset_2_initial"
-                    , "o2_sat_admis"
-                    , "o2_sat_suppl"
-                    , "ethnicity"
-                    , "smoking"
-                    , "ia_cxr"
-                    , "max_resp_score"
-                    , "T1_resp_score"
-                    , "com_noasthma"
-                    , "T2_resp_score"
-                    , "inresp_sev"
-                    , "steroid")
-
-#npa_data = 
-
 reg_data = fp_adults[, cols_to_extract]
 
 # sanity checks

read_data.R

@@ -43,7 +43,6 @@ all(meta_data_cols%in%colnames(all_df))
 
 metadata_all = all_df[, meta_data_cols]
 
-
 #==============
 # adult patients
 #==============
@@ -56,6 +55,12 @@ if (table(adult_df$adult == 1)[[1]] == nrow(adult_df) ){
   cat ("FAIL: adult df number mismatch!")
 }
 
+#==============
+# FLU positive: adult patients
+#==============
+# extract the flu positive population
+fp_adults = adult_df[adult_df$flustat == 1,]
+
 #============
 # hc
 #============