Version 250314
===================================
REMINDER
ICDPICR Version 2.0.5 IS BEING TESTED
Major bugs and flaws may still exist
Please report issues to david.clark@tufts.edu
or at github/clark-david/icdpicr2/issues
===================================
#1 #CLEAR WORKSPACE, SET WORKING DIRECTORY, #LOAD REQUIRED PACKAGES,
IF NOT LOADED ALREADY rm(list=ls())
setwd(“/Users/davideugeneclark/Documents/icdpicr”)
require(tidyverse) require(janitor) require(broom) require(skimr)
#2a #GET TQIP DATA
#Import raw 2020 TQIP data d0<-read_csv(“PUF AY
2020/CSV/PUF_TRAUMA.csv”)
d1<-rename(d0,INC_KEY=inc_key,TQPISS=ISS)
#Demonstrate that INC_KEY is unique identifier d1<-group_by(d1,INC_KEY) d1<-mutate(d1,seq=row_number()) d1<-ungroup(d1) tabyl(d1,seq)
#Save recorded outcome data and ISS d2<-select(d1,INC_KEY,HOSPDISCHARGEDISPOSITION,TQPISS) write_csv(d2,“tqipmort2020.csv”)
#Obtain ICD-10-CM codes for each TQIP patient d0<-read_csv(“PUF AY 2020/CSV/PUF_ICDDIAGNOSIS.csv”) d1<-rename(d0,INC_KEY=Inc_Key,withdot=ICDDIAGNOSISCODE)
#Identify duplicates d2<-group_by(d1,INC_KEY,withdot) d2<-mutate(d2,dxseq=row_number()) d2<-mutate(d2,dxrep=max(dxseq)) d2<-ungroup(d2) d2test<-filter(d2,dxrep>1) d2test<-arrange(d2test,INC_KEY,withdot,dxseq,dxrep) #View(d2test) #Drop duplicates d3<-filter(d2,dxseq==1) d3<-select(d3,INC_KEY,withdot) d3<-arrange(d3,INC_KEY,withdot)
#Extract parts of each ICD-10-CM code #Convert to format without dot
(e.g., S060X0A not S06.0X0A) #Restrict to codes with first digit “S” or
“T” # and seventh digit “A”,“B”,“C”, or blank
d4<-arrange(d3,INC_KEY,withdot)
d4<-mutate(d4,predot=str_sub(withdot,1,3))
d4<-mutate(d4,postdot=str_sub(withdot,5,8))
d4<-mutate(d4,icdcm=str_c(predot,postdot))
d4<-mutate(d4,digit1=str_sub(icdcm,1,1))
d4<-mutate(d4,digit7=str_sub(icdcm,7,7)) d4<-filter(d4,digit1==“S”
| digit1==“T”) d4<-filter(d4,digit7==“A” | digit7==“B” | digit7==“C”
| digit7==““)
d5<-select(d4,INC_KEY,icdcm) write_csv(d5,“tqipdcode2020.csv”)
#Merge TQIP patient diagnoses and outcomes
d1<-read_csv(“tqipdcode2020.csv”)
d2<-read_csv(“tqipmort2020.csv”)
d3<-full_join(d1,d2,by=“INC_KEY”)
#Drop if outcome unspecified or if transferred to another acute care hospital #Generate “died” outcome, and save result d4<-filter(d3,HOSPDISCHARGEDISPOSITION>=2 & HOSPDISCHARGEDISPOSITION<=14) d4<-mutate(d4,died=if_else( HOSPDISCHARGEDISPOSITION==5,1,0)) tabyl(d4,HOSPDISCHARGEDISPOSITION,died) write_csv(d4,“tqipmerged2020.csv”)
#2b
#GET NIS DATA
#Import raw 2020 NIS cases
d0<-read_fwf(“NIS_2020/NIS_2020_Core.ASC”, col_positions=fwf_positions( c(1,10,23,35,39,55,62,69,76,83,90,97,104,111,118,125,132,139,146,153,160,167,174,181,188, 195,202,209,216,223,230,237,244,251,258,265,272,279,286,293,300,307,314,321,328,335,521,531), c(3,11,24,36,41,61,68,75,82,89,96,103,110,117,124,131,138,145,152,159,166,173,180,187,194, 201,208,215,222,229,236,243,250,257,264,271,278,285,292,299,306,313,320,327,334,336,530,535), c(“AGE”,“DIED”,“DISPUNIFORM”,“ELECTIVE”,“HCUP_ED”, “I10_DX1”,“I10_DX2”,“I10_DX3”,“I10_DX4”,“I10_DX5”,“I10_DX6”,“I10_DX7”,“I10_DX8”, “I10_DX9”,“I10_DX10”,“I10_DX11”,“I10_DX12”,“I10_DX13”,“I10_DX14”,“I10_DX15”,“I10_DX16”, “I10_DX17”,“I10_DX18”,“I10_DX19”,“I10_DX20”,“I10_DX21”,“I10_DX22”,“I10_DX23”,“I10_DX24”, “I10_DX25”,“I10_DX26”,“I10_DX27”,“I10_DX28”,“I10_DX29”,“I10_DX30”,“I10_DX31”,“I10_DX32”, “I10_DX33”,“I10_DX34”,“I10_DX35”,“I10_DX36”,“I10_DX37”,“I10_DX38”,“I10_DX39”,“I10_DX40”, “INJURY”,“INC_KEY”,“LOS”) )) View(d0)
#Include only admissions that were not elective and originated in Emergency Department tabyl(d0,ELECTIVE,HCUP_ED) d1<-filter(d0,ELECTIVE==0,HCUP_ED!=0) tabyl(d1,ELECTIVE,HCUP_ED) #Include only admissions with a principal AHRQ “injury” diagnosis (S00-T34 and some others) tabyl(d1,INJURY) d2<-filter(d1,INJURY==1) tabyl(d2,INJURY)
#Verify that INC_KEY is a unique identifier d2<-group_by(d2,INC_KEY) d2<-mutate(d2,seq=row_number()) d2<-ungroup(d2) tabyl(d2,seq)
write_csv(d2,“NISraw2020.csv”)
#Reshape to “long” format d3<-gather(d2,I10_DX1,I10_DX2,I10_DX3,I10_DX4,I10_DX5,I10_DX6,I10_DX7,I10_DX8, I10_DX9,I10_DX10,I10_DX11,I10_DX12,I10_DX13,I10_DX14,I10_DX15,I10_DX16, I10_DX17,I10_DX18,I10_DX19,I10_DX20,I10_DX21,I10_DX22,I10_DX23,I10_DX24, I10_DX25,I10_DX26,I10_DX27,I10_DX28,I10_DX29,I10_DX30,I10_DX31,I10_DX32, I10_DX33,I10_DX34,I10_DX35,I10_DX36,I10_DX37,I10_DX38,I10_DX39,I10_DX40, key=“original”,value=“icdcm”)
#Identify duplicates d3a<-group_by(d3,INC_KEY,icdcm) d3a<-mutate(d3a,dxseq=row_number()) d3a<-mutate(d3a,dxrep=max(dxseq)) d3a<-ungroup(d3a) #Drop duplicates (mostly NA) d3b<-filter(d3a,dxseq==1) d3b<-select(d3b,INC_KEY,icdcm,DIED,DISPUNIFORM,LOS) d3b<-arrange(d3b,INC_KEY,icdcm) View(d3b)
#Extract parts of each ICD-10-CM code #Restrict to codes with first
digit “S” or “T” # and seventh digit “A”,“B”,“C”, or blank
d4<-arrange(d3b,INC_KEY,icdcm)
d4<-mutate(d4,digit1=str_sub(icdcm,1,1))
d4<-mutate(d4,digit7=str_sub(icdcm,7,7)) d4<-filter(d4,digit1==“S”
| digit1==“T”) d4<-filter(d4,digit7==“A” | digit7==“B” | digit7==“C”
| digit7==““)
#Verify that “DIED” corresponds correctly to “DISPUNIFORM==20” #Rename DIED to died, drop if died not 0 or 1 #Drop if transferred to another acute-care hospital #Create dummy TQPISS tabyl(d4,DIED,DISPUNIFORM) d4<-rename(d4,died=DIED) d5<-filter(d4,died>=0) d5<-filter(d5,DISPUNIFORM!=2) d5<-mutate(d5,longLOS=if_else(died==1 | LOS>=10, 1,0)) d5<-mutate(d5,TQPISS=0)
#Discard temporary analytic variables and save result d5<-select(d5,INC_KEY,icdcm,died,TQPISS) write_csv(d5,“nisreshaped2020.csv”)
#3 #PREPARE TQIP OR NIS DATA FOR REGRESSION ANALYSIS
d0<-read_csv(“tqipmerged2020.csv”) d0<-read_csv(“nisreshaped2020.csv”)
#Inspect for otherwise valid codes not ending in “A”,“B”, or
“C”
d1<-filter(d0,str_sub(icdcm,7,7)==““) tabyl(d1,icdcm) #There are
some, especially T31x, T31xx, T32x, T32xx (Burns)
#Restrict to cases with at least one valid icdcm Anatomic Injury Code # as defined by National Trauma Data Standard # or codes T20-T32 (denoting burns/corrosions) d4<-mutate(d0,validcode=case_when( str_sub(icdcm,1,1)==“S” & str_sub(icdcm,7,7)==“A” ~ 1, str_sub(icdcm,1,1)==“S” & str_sub(icdcm,7,7)==“B” ~ 1, str_sub(icdcm,1,1)==“S” & str_sub(icdcm,7,7)==“C” ~ 1, str_sub(icdcm,1,1)==“T” & str_sub(icdcm,2,3)==“07” ~ 1, str_sub(icdcm,1,1)==“T” & str_sub(icdcm,2,3)==“14” ~ 1, str_sub(icdcm,1,1)==“T” & str_sub(icdcm,2,2)==“2” ~ 1, str_sub(icdcm,1,1)==“T” & str_sub(icdcm,2,3)==“30” ~ 1, str_sub(icdcm,1,1)==“T” & str_sub(icdcm,2,3)==“31” ~ 1, str_sub(icdcm,1,1)==“T” & str_sub(icdcm,2,3)==“32” ~ 1, str_sub(icdcm,1,5)==“T79.A” & str_sub(icdcm,7,7)==“A” ~ 1, TRUE ~ 0 ) )
#Inspect whether valid codes properly identified d4test<-group_by(d4,icdcm) d4test<-mutate(d4test,dxseq=row_number()) d4test<-filter(d4test,dxseq==1) d4test<-ungroup(d4test) d4test<-arrange(d4test,icdcm) #View(d4test)
#Identify valid cases (with at least one valid code and died==0 or 1)
#Discard invalid cases and invalid codes
d5<-group_by(d4,INC_KEY) d5<-mutate(d5,idseq=row_number())
d5<-mutate(d5,validcase=max(validcode)) d5<-ungroup(d5)
tabyl(d5,validcase,validcode)
d6<-filter(d5,validcase==1,validcode==1) d6<-filter(d6,died>=0)
tabyl(d6,validcase,validcode)
#Discard temporary analytic variables, and save results d7<-select(d6,INC_KEY,icdcm,died,TQPISS) write_csv(d7,“tqip2020cm.csv”) write_csv(d7,“nis2020cm.csv”)