1 Introduction

    library(clinUtils)
    library(tools)# toTitleCase
    library(plyr) # for ddply, rbind.fill
    library(pander) # for session info
    library(inTextSummaryTable)

1.1 Data format

The package is demonstrated with a subset of the ADaM datasets from the CDISC Pilot 01 dataset, available in the clinUtils package.

    # load example data
    library(clinUtils)

    # load example data
    data(dataADaMCDISCP01)

    dataAll <- dataADaMCDISCP01
    labelVars <- attr(dataAll, "labelVars")

Typical in-text table for the CSR are included in the following sections.

Please note that the table content e.g. variables, statistics of interest depends strongly on the study at hand and personal preferences.

2 Subject information

2.1 Subject disposition

    # data of interest
    dataDM <- dataAll$ADSL
    
    varDMFL <- grep("FL$", colnames(dataDM), value = TRUE)
    varDMFLLabel <- sub(" Flag$", "", labelVars[varDMFL])
    
    getSummaryStatisticsTable(
        data = dataDM,
        var = varDMFL, varFlag = varDMFL, varGeneralLab = "Analysis Set, N", 
        varLab = varDMFLLabel,
        stats = getStats("n (%)"),
        colVar = "TRT01P",
        labelVars = labelVars,
        colTotalInclude = TRUE, colTotalLab = "All subjects",
        varInclude0 = TRUE,
        title = toTitleCase("Table: subject disposition"),
        file = file.path("tables_CSR", "Table_subjectDisposition.docx")
    )

2.2 Demographics

    # data of interest
    dataDM <- subset(dataAll$ADSL, SAFFL == "Y")
    
    # variables of interest
    # Note: if available: ethnicity is included
    varsDM <- c(
        "SEX", "AGE", "AGEGR1",
        "RACE", "ETHNIC",
        "HEIGHTBL", "WEIGHTBL", 
        "BMIBL", "BMIBLGR1"
    )

    # Sort variables according to corresponding numeric variable
    dataDM$AGEGR1 <- with(dataDM, reorder(AGEGR1, AGEGR1N))
    dataDM$RACE <- with(dataDM, reorder(RACE, RACEN))
    dataDM$TRT01P <- with(dataDM, reorder(TRT01P, TRT01PN))
    
    ## Define set of statistics of interest:
    statsDM <- getStatsData(
        data = dataDM, var = varsDM,
        # different for continuous and categorical variable
        type = c(cont = "median (range)", cat = "n (%)"),
        # for categorical variable, statistic name (here: 'n (%)')
        # should not be included in the table
        args = list(cat = list(includeName = FALSE))
    )

    ## create the table:
    
    getSummaryStatisticsTable(
        data = dataDM, 
        # variables to summarize
        var = varsDM, 
        varGeneralLab = "Parameter",
        # column
        colVar = "TRT01P", colTotalInclude = TRUE, colTotalLab = "All subjects",
        # statistics
        stats = statsDM,
        statsGeneralLab = "",
        labelVars = labelVars,
        # if only one category, should be included in separated row (e.g. RACE: White)
        rowAutoMerge = FALSE,
        rowInclude0 = FALSE, emptyValue = 0,
        title = toTitleCase("Table: Demographic Data (safety Analysis Set)"),
        file = file.path("tables_CSR", "Table_demographicData.docx")
    )

2.3 Baseline disease characteristics

Please note that the content of the table strongly depends on the study.

    # data of interest
    dataBDC <- subset(dataAll$ADSL, SAFFL == "Y")
    
    # create table
    getSummaryStatisticsTable(
        data = dataBDC,
        var = c("DURDIS", "EDUCLVL"), varGeneralLab = "Parameter", 
        colVar = "TRT01P", colTotalInclude = TRUE, colTotalLab = "All subjects",
        stats = getStats("median\n(range)"), statsGeneralLab = "",
        rowAutoMerge = FALSE,
        labelVars = labelVars,
        title = toTitleCase("Table: Baseline Disease Characteristics (safety analysis set)"),
        file = file.path("tables_CSR", "Table_BaselineCharacteristics.docx")
    )

2.4 Medical History and Concomitant Diseases

    dataCM <- subset(dataAll$ADCM, SAFFL == "Y")

    # sort variable according to corresponding numeric variables
    dataCM$TRTA <- with(dataCM, reorder(TRTA, TRTAN))
    
    # Terms should be in lower-case
    dataCM$CMDECOD <- simpleCap(tolower(dataCM$CMDECOD))
    dataCM$CMCLAS <- simpleCap(tolower(dataCM$CMCLAS))
            
    getSummaryStatisticsTable(
        data = dataCM,
        colVar = "TRTA", colTotalInclude = TRUE, colTotalLab = "All subjects",
        rowVar = c("CMCLAS", "CMDECOD"), 
        # include total across generic terms and across ATC4 classes
        rowVarTotalInclude = c("CMCLAS", "CMDECOD"), 
        rowTotalLab = "Any prior and concomitant medication",
        stats = getStats("n (%)"),
        # sort rows based on counts of subjects in the total column 
        rowOrder = "total",
        labelVars = labelVars,
        emptyValue = 0,
        title = toTitleCase(paste("Prior and concomitant therapies",
            "by medication class and generic term (safety analyis set)"
        )),
        file = file.path("tables_CSR", "Table_CM.docx")
    )

3 Efficacy Analyses

The example dataset has has two primary endpoints:

  • ADAS-Cog (11), a.k.a Alzheimer’s Disease Assessment Scale - Cognitive Subscale a metric containing 11 items, available in the ADQSADAS dataset
  • CIBIC+ score a.k.a Video-referenced Clinician’s Interview-based Impression of Change available in the ADQSCIBC dataset
    dataAdasCog11 <- subset(dataAll$ADQSADAS, PARAMCD == "ACTOT")
    dataCIBIC <- subset(dataAll$ADQSCIBC, PARAMCD == "CIBICVAL")
    
    dataEfficacy <- plyr::rbind.fill(dataAdasCog11, dataCIBIC)
    
    dataEfficacy$TRTP <- with(dataEfficacy, reorder(TRTP, TRTPN))
    dataEfficacy$AVISIT <- with(dataEfficacy, reorder(AVISIT, AVISITN))
    
    stats <- getStatsData(
        data = dataEfficacy, 
        var = c("AVAL", "CHG"), 
        type = c("n", "mean (se)", "median (range)")
    )
    
    getSummaryStatisticsTable(
        data = dataEfficacy,
        rowVar = "PARAM",
        colVar = c("TRTP", "AVISIT"),
        var = c("AVAL", "CHG"), 
        stats = stats,
        labelVars = labelVars,
        title = paste("Table: efficacy endpoints", 
            toTitleCase("actual value and changes from baseline per time point"             
        )),
        file = file.path("tables_CSR", "Table_efficacy.docx")
    )

4 Safety Analyses

4.1 Adverse Events

4.1.1 Treatment-emergent summary table

    ## data of interest: safety analysis set and treatment-emergent
    dataTEAE <- subset(dataAll$ADAE, SAFFL == "Y" & TRTEMFL == "Y")
    
    # order treatment and severity categories
    dataTEAE$TRTA <- with(dataTEAE, reorder(TRTA, TRTAN))
    
    ## data considered for the total
    dataTotalAE <- subset(dataAll$ADSL, SAFFL == "Y")
    dataTotalAE$TRTA <- with(dataTotalAE, reorder(TRT01A, TRT01AN))
    
    # TEAE with worst intensity
    # build worst-case scenario
    dataTEAE$AESEV <- factor(dataTEAE$AESEV, levels = c("MILD", "MODERATE", "SEVERE"))
    dataTEAE$AESEVN <- as.numeric(dataTEAE$AESEV)
    dataTEAE <- ddply(dataTEAE, c("USUBJID", "TRTA"), function(x)
        cbind.data.frame(x, 
            WORSTINT = with(x, ifelse(AESEVN == max(AESEVN), as.character(AESEV), NA_character_))
    ))
    dataTEAE$WORSTINT <- factor(dataTEAE$WORSTINT, levels = levels(dataTEAE$AESEV))
    
    ## specify labels for each variable:
    varsAE <- c("TRTEMFL", "AESER", "AESDTH", "AEREL")
    
    # create the table
    getSummaryStatisticsTable(
        data = dataTEAE,
        colVar = "TRTA",
        # define variables to compute statistics on
        var = c("TRTEMFL", "AESER", "WORSTINT", "AESDTH", "AEREL"), 
        varFlag = c("TRTEMFL", "AESER", "AESDTH"),
        varLab = c(TRTEMFL = "Treatment-Emergent", WORSTINT = "Worst-case severity:"),
        varGeneralLab = "Subjects with, n(%):",
        # force the inclusion of lines for variable without count:
        varInclude0 = TRUE,
        # include the total for the worst-case scenario
        varTotalInclude = "WORSTINT",
        # statistics:
        stats = getStats('n (%)'),
        emptyValue = "0",
        labelVars = labelVars,
        # dataset used for the total in the header column (and for percentage as default)
        dataTotal = dataTotalAE,
        # title/export
        title = toTitleCase("Table: Summary Table of Treatment-emergent Adverse Events (safety analysis set)"),
        file = file.path("tables_CSR", "Table_TEAE_summary.docx")
    )

4.1.2 Treatment-emergent incidence table

4.1.2.1 Events occuring in at least one subject

    dataTEAE <- subset(dataAll$ADAE, SAFFL == "Y" & TRTEMFL == "Y")
    
    # order treatment and severity categories
    dataTEAE$TRTA <- with(dataTEAE, reorder(TRTA, TRTAN))
    
    ## data considered for the total
    dataTotalAE <- subset(dataAll$ADSL, SAFFL == "Y")
    dataTotalAE$TRTA <- with(dataTotalAE, reorder(TRT01A, TRT01AN))
    
    getSummaryStatisticsTable(
        data = dataTEAE,
        rowVar = c("AESOC", "AEDECOD"),
        colVar = "TRTA",
        ## total
        # data
        dataTotal = dataTotalAE,
        # row total
        rowVarTotalInclude = c("AESOC", "AEDECOD"), rowTotalLab = "Any TEAE",
        stats = getStats("n (%)"),
        labelVars = labelVars,
        rowVarLab = c('AESOC' = "TEAE by SOC and Preferred Term,\nn (%)"),
        # sort rows based on the total column:
        rowOrder = "total", 
        rowOrderTotalFilterFct = function(x) subset(x, TRTA == "Total"),
        title = paste("Table: Treatment-emergent Adverse Events by System Organ Class",
            "and Preferred Term (Safety Analysis Set)"
        ),
        file = file.path("tables_CSR", "Table_TEAE_SOCPT_atLeast1Subject.docx")
    )

4.1.2.2 Events occuring in at least 25% of all subjects

    getSummaryStatisticsTable(
        data = dataTEAE,
        rowVar = c("AESOC", "AEDECOD"),
        colVar = "TRTA",
        ## total
        # data
        dataTotal = dataTotalAE, 
        # row total
        rowVarTotalInclude = c("AESOC", "AEDECOD"), rowTotalLab = "Any TEAE",
        stats = getStats("n (%)"),
        labelVars = labelVars,
        rowVarLab = c('AESOC' = "SOC and Preferred Term,\nn (%)"),
        # sort rows based on the total column:
        rowOrder = "total", 
        rowOrderTotalFilterFct = function(x) subset(x, TRTA == "Total"),
        title = paste("Table: Treatment-emergent Adverse Events by System Organ Class",
            "and Preferred Term reported in at least 25% of the subjects",
            "in any treatment group (Safety Analysis Set)"
        ),
        file = file.path("tables_CSR", "Table_TEAE_SOCPT_atLeast25PercentsSubject.docx"),
        # include only events occuring in at least 25% for at least one preferred term:
        filterFct = function(x)
            ddply(x, "AESOC", function(x){ # per AESOC to include the total
                ddply(x, "AEDECOD", function(y){
                    yTotal <- subset(y, grepl("Total", TRTA))
                    if(any(yTotal$statPercN >= 25)) y
                })
            })
    )

4.1.3 Treatment-emergent worst-case table

    dataTEAE <- subset(dataAll$ADAE, SAFFL == "Y" & TRTEMFL == "Y")
    
    # order treatment and severity categories
    dataTEAE$TRTA <- with(dataTEAE, reorder(TRTA, TRTAN))
    
    ## data considered for the total
    dataTotalAE <- subset(dataAll$ADSL, SAFFL == "Y")
    dataTotalAE$TRTA <- with(dataTotalAE, reorder(TRT01A, TRT01AN))
    
    # TEAE with worst intensity
    dataTEAE$AESEV <- factor(dataTEAE$AESEV, levels = c("MILD", "MODERATE", "SEVERE"))
    dataTEAE$AESEVN <- as.numeric(dataTEAE$AESEV)
    
    # extract worst-case scenario data (only one record if multiple with same severity)
    dataAEWC <- ddply(dataTEAE, c("AESOC", "AEDECOD", "USUBJID", "TRTA"), function(x){
        x[which.max(x$AESEVN), ]
    })
    # worst-case scenario in lower case
    dataAEWC$WORSTINT <- simpleCap(tolower(dataAEWC$AESEV))
    labelVars["WORSTINT"] <- "Worst-case scenario"

    ## datasets used for the total: 
    # for total: compute worst-case across SOC and across AE term
    # (otherwise patient counted in multiple categories if present different categories for different AEs)
    dataTotalRow <- list(
        # within SOC (across AEDECOD)
        'AEDECOD' = ddply(dataAEWC, c("AESOC", "USUBJID", "TRTA"), function(x){ 
            x[which.max(x$AESEVN), ]
        }),
        # across SOC
        'AESOC' = ddply(dataAEWC, c("USUBJID", "TRTA"), function(x){    
            x[which.max(x$AESEVN), ]
        })
    )
    
    getSummaryStatisticsTable(
        data = dataAEWC,
        ## row variables:
        rowVar = c("AESOC", "AEDECOD", "WORSTINT"), rowVarInSepCol = "WORSTINT",
        # include total across SOC and across AEDECOD
        rowVarTotalInclude = c("AESOC", "AEDECOD"), dataTotalRow = dataTotalRow, 
        rowVarTotalByVar = "WORSTINT", # count for each severity category for the total
        rowTotalLab = "Any TEAE", rowVarLab = c(AESOC = "Subjects with, n(%):", WORSTINT = "Worst-case scenario"),
        # sort per total in the total column
        rowOrder = "total", 
        ## column variables
        colVar = "TRTA", 
        stats = getStats("n (%)"),
        emptyValue = "0",
        labelVars = labelVars,
        dataTotal = dataTotalAE,
        title = toTitleCase(paste("Table: Treatment-emergent Adverse",
            "Events by system organ",
            "and preferred term by worst-case (safety Analysis Set)"
        )),
        file = file.path("tables_CSR", "Table_TEAE_Severity.docx")
    )