In DHLab-CGU/emr: An analysis tool for diagnosis and procedure records

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Usage of dxpr Package

knitr::opts_chunk$set(collapse = T, comment = "#>")
options(tibble.print_min = 4L, tibble.print_max = 4L)
remotes::install_github("DHLab-TSENG/dxpr")
library(dxpr) 
library(data.table)

I. Data integration

head(sampleDxFile)

1. Data standardization

# ICD to uniform short format
ICD_Short <- icdDxDecimalToShort(dxDataFile = sampleDxFile,
                                 icdColName = ICD,
                                 dateColName = Date, 
                                 icd10usingDate = "2015/10/01")

head(ICD_Short$ICD)

tail(ICD_Short$Error)

sampleDxFile$Decimal <- ICD_Short$ICD 

head(sampleDxFile)

2. Data grouping

# ICD to level 2 CCS description of multiple-level CCS
CCSlvl_description <- icdDxToCCSLvl(dxDataFile = sampleDxFile,
                                    idColName = ID,
                                    icdColName = ICD,
                                    dateColName = Date, 
                                    icd10usingDate = "2015-10-01", 
                                    CCSLevel = 2,
                                    isDescription = TRUE)

head(setorder(CCSlvl_description$groupedDT, "ID"))

head(CCSlvl_description$summarised_groupedDT)

## ICD to pheWAS
pheWAS <- icdDxToPheWAS(dxDataFile  = sampleDxFile,
                        idColName = ID,
                        icdColName = ICD,
                        dateColName = Date, 
                        icd10usingDate = "2015-10-01", 
                        isDescription = FALSE)

head(setorder(pheWAS$groupedDT, "ID"))

head(setorder(pheWAS$summarised_groupedDT, "ID"))

# ICD to Comorbidity: Elixhauser
ELIX <- icdDxToComorbid(dxDataFile  = sampleDxFile,
                        idColName = ID,
                        icdColName = ICD,
                        dateColName = Date, 
                        icd10usingDate = "2015-10-01", 
                        comorbidMethod = ELIX)

head(setorder(ELIX$groupedDT, "ID"))

head(ELIX$summarised_groupedDT)

# ICD to Comorbidity: AHRQ
AHRQ <- icdDxToComorbid(dxDataFile  = sampleDxFile,
                        idColName = ID,
                        icdColName = ICD,
                        dateColName = Date, 
                        icd10usingDate = "2015-10-01", 
                        comorbidMethod = ahrq)

head(setorder(AHRQ$groupedDT, "ID"))

head(AHRQ$summarised_groupedDT)

# ICD to Comorbidity: Charlson
Charlson <- icdDxToComorbid(dxDataFile  = sampleDxFile,
                        idColName = ID,
                        icdColName = ICD,
                        dateColName = Date, 
                        icd10usingDate = "2015-10-01", 
                        comorbidMethod = Charlson)

head(setorder(Charlson$groupedDT, "ID"))

head(Charlson$summarised_groupedDT)

# ICD to self-defined grouping method: precise matching
groupingTable <- data.table(Group = rep("Chronic kidney disease",6),
                            ICD = c("N181","5853","5854","5855","5856","5859"))

groupingTable

CustomGroup <- icdDxToCustom(dxDataFile  = sampleDxFile,
                             idColName = ID,
                             icdColName = ICD, 
                             dateColName = Date, 
                             customGroupingTable = groupingTable)

head(setorder(CustomGroup$groupedDT, "ID"))

head(CustomGroup$summarised_groupedDT)

# ICD to self-defined grouping method: fuzzy matching
grepTable <- data.table(Group = "Chronic kidney disease",
                          grepIcd = "^585|^N18")

grepTable

CustomGrepGroup <- icdDxToCustomGrep(dxDataFile  = sampleDxFile,
                                     idColName = ID,
                                     icdColName = ICD, 
                                     dateColName = Date, 
                                     customGroupingTable = grepTable)

head(setorder(CustomGrepGroup$groupedDT, "ID"))

II. Data wrangling

1. Case selection

# Select cases by ICD condition "^585"
Case <- selectCases(dxDataFile = sampleDxFile,
                    idColName = ID,
                    icdColName = ICD, 
                    dateColName = Date, 
                    icd10usingDate = "2015/10/01",
                    groupDataType = icd,
                    caseCondition = "^585",
                    caseCount = 2,  
                    periodRange = c(30, 365))

head(Case)

tail(Case)

# Select cases by AHRQ comorbidity condition  "Renal"
Case <- selectCases(dxDataFile = sampleDxFile,
                    idColName = ID,
                    icdColName = ICD, 
                    dateColName = Date, 
                    icd10usingDate = "2015/10/01",
                    groupDataType = ahrq,
                    isDescription = FALSE,
                    caseCondition = "Renal",
                    caseCount = 1,
                    periodRange = c(30, 365))

head(Case)

# Select cases by CCS Level 2 condition "Diseases of the urinary system"
Case <- selectCases(dxDataFile = sampleDxFile,
                    idColName = ID,
                    icdColName = ICD, 
                    dateColName = Date, 
                    icd10usingDate = "2015/10/01",
                    groupDataType = ccslvl2,
                    isDescription = TRUE,
                    caseCondition = "Diseases of the urinary system",
                    caseCount = 1,
                    periodRange = c(30, 365))

head(Case)

2. Condition era generation

# get era by the comorbidity condition ELIX
Era <- getConditionEra(dxDataFile = sampleDxFile,
                       idColName = ID,
                       icdColName = ICD,
                       dateColName = Date,
                       icd10usingDate = "2015/10/01",
                       groupDataType = ELIX,
                       isDescription = TRUE,
                       gapDate = 30)

head(Era)

3. Analysis ready data generation

# generate analysis ready data mapped with case and Control Group
CHARLSON <- icdDxToComorbid(dxDataFile = sampleDxFile,
                            idColName = ID,
                            icdColName = ICD,
                            dateColName = Date,
                            icd10usingDate = "2015-10-01",
                            comorbidMethod = CHARLSON)

head(CHARLSON$groupedDT)

selectedCaseFile <- selectCases(dxDataFile = sampleDxFile,
                                idColName = ID,
                                icdColName = ICD,
                                dateColName = Date,
                                icd10usingDate = "2015/10/01",
                                groupDataType = ccs,
                                caseCondition = "11",
                                isDescription = FALSE,
                                caseCount = 1)    

head(selectedCaseFile)

groupedDataWide <- groupedDataLongToWide(dxDataFile  = CHARLSON$groupedDT,
                                         idColName = ID,
                                         categoryColName = Comorbidity,
                                         dateColName = Date,
                                         reDup = TRUE,
                                         numericOrBinary = N,
                                         selectedCaseFile = selectedCaseFile)

tail(groupedDataWide)

III. Visualization

# Bar chart of diagnostic categories – case and control groups
selectedCaseFile <- selectCases(dxDataFile = sampleDxFile,
                                idColName = ID,
                                icdColName = ICD,
                                dateColName = Date,
                                icd10usingDate = "2015/10/01",
                                groupDataType = ccs,
                                caseCondition = "11",
                                isDescription = FALSE,
                                caseCount = 1) 

groupedDataWide <- groupedDataLongToWide(dxDataFile  = ELIX$groupedDT,
                                         idColName = ID,
                                         categoryColName = Comorbidity,
                                         dateColName = Date,
                                         reDup = TRUE,
                                         numericOrBinary = B,
                                         count = 2,
                                         selectedCaseFile = selectedCaseFile)

plot <- plotDiagCat(groupedDataWide = groupedDataWide,
                    idColName = ID,
                    groupColName = selectedCase,
                    topN = 10,
                    limitFreq = 0.01,
                    pvalue = 0.05)
plot$graph 

plot$sigCate

IV. Procedure

head(samplePrFile)

1. Code classification

# CCS single level classification
CCS <- icdPrToCCS (prDataFile = samplePrFile,
                   idColName = ID,
                   icdColName = ICD, 
                   dateColName = Date,
                   icd10usingDate = "2015-10-01",
                   isDescription = TRUE) 

head(CCS$groupedDT)

# ICD to CCS multiple level 1
CCSlvl <- icdPrToCCSLvl(prDataFile = samplePrFile,
                        idColName = ID,
                        icdColName = ICD, 
                        dateColName = Date,
                        icd10usingDate = "2015-10-01",
                        CCSLevel = 2,
                        isDescription = FALSE)

head(CCSlvl$groupedDT)

# Procedure class
ProcedureClass <- icdPrToProcedureClass(prDataFile = samplePrFile,
                                        idColName = ID,
                                        icdColName = ICD, 
                                        dateColName = Date, 
                                        icd10usingDate = "2015-10-01", 
                                        isDescription = FALSE)

head(ProcedureClass$groupedDT)