R/mul.model.R

In vandy10s/mnma: Multivariate Network Meta-Analysis (NMA) within a Bayesian Framework

#' @title Create a data object for multivariate network meta-analysis (MNMA)
#' @description The function creates a data frame for bugs operation, using outcome 1, outcome 2, and treatments inputs
#' @param data1 A data frame for outcome 1. First 2 (or 3) columns contain the information about number of events for each arm and the rest 2 (or 3) colums contain the sample size for each arm. Number of columns is from 2 to 3, depending on the number of maximum arms in the data.
#' @param data2 A data frame for outcome 2, which has the same number of rows and columns.
#' @param TX A data frame for tratments, which has maximum 3 treatments in 3 columns. The number of rows should coincide with data1 and data2.
#' @param noa Maximum number of arms in the trial studies.
#' @param ref reference treatment numbers in the analysis
#' @param parameters 'simple' or 'details' list of parameters.
#' @return mnma.model object for mnma.run operation
#' @keywords internal


mul.model <- function(data1, data2, TX ,noa="3arm", ref=1, parameters="simple"){


  if (noa=="3arm"){
    # transfomrm as events and failures then Add 0.5 for 0 cells -----------------------------
    out1 <- mnma:::add.half.3arm(data1)
    out2 <- mnma:::add.half.3arm(data2)
    # ----------------------------------------------------------------------------------------
    # check the number of treatments
    NT <- mnma:::tx.check(TX)
    #-----------------------------------------------------------------------------------------
    # sorting by na information
    data <- mnma:::sortbyna(out1,out2,TX)
    #head(data)
    #-----------------------------------------------------------------------------------------
    NS  <- nrow(data)
    N2h <- sum(is.na(data$T3))
    N1O <- sum(data$noo==1)  # number of missing studies
    # ----------------------------------------------------------------------------------------
    # Split the dataset
    dat1 <- data[,c(1:6)]     # outcome1
    dat2 <- data[,c(7:12)]    # outcome2
    T1<-data$T1
    T2<-data$T2
    T3<-data$T3
    #--------------------------------------------------------------------------------------------
    # Data transformation
    #--------------------------------------------------------------------------------------------
    ref  =  ref
    out1 <- mnma:::cont.trans(dat1, NS=NS, N2h=N2h)
    out2 <- mnma:::cont.trans(dat2, NS=NS, N2h=N2h)
    y    <- mnma:::gen.y(out1$y, out2$y, NS=NS, N2h=N2h)
    var1 <- out1$var
    var2 <- out2$var

    input <- list(NS=NS,N2h=N2h,NT=NT,ref=ref,T1=T1,T2=T2,T3=T3,y=y,var1=var1,var2=var2)

    if (parameters=="simple"){
      parameters <- c( "dOUT1","dOUT2","rho1","psi1","psi2")
    } else if (parameters=="detail"){
      parameters <- c( "ddOUT1","ddOUT2", "rho1","psi1","psi2")
    } else if (parameters!="detail" & parameters!="simple"){
      stop("Choose either 'simple' or 'detail' for the list of parameters")
    }

    source <- list(data = input, parameters = parameters, noa=noa)
    class(source) <- "mnma.model"
    return(source)
  }

  #====================================================================================================

  if (noa=="2arm"){

    # warnings about the dimensions


    # transfomrm as events and failures then Add 0.5 for 0 cells -----------------------------
    out1 <- mnma:::add.half.2arm(data1)
    out2 <- mnma:::add.half.2arm(data2)
    # ----------------------------------------------------------------------------------------
    # check the number of treatments
    NT <- mnma:::tx.check(TX)
    #-----------------------------------------------------------------------------------------
    # sorting by na information
    data <- mnma:::sortbyna.2arm(out1,out2,TX)
    #head(data)
    #-----------------------------------------------------------------------------------------
    NS  <- nrow(data)
    N2h <- nrow(data)
    N1O <- sum(data$noo==1)  # number of missing studies
    # ----------------------------------------------------------------------------------------
    # Split the dataset
    dat1 <- data[,c(1:4)]     # outcome1
    dat2 <- data[,c(5:8)]    # outcome2
    T1<-data$T1
    T2<-data$T2
    # T3<-data$T3
    #--------------------------------------------------------------------------------------------
    # Data transformation
    #--------------------------------------------------------------------------------------------
    ref  =  ref
    out1 <- mnma:::cont.trans.2arm(dat1, NS=NS)
    out2 <- mnma:::cont.trans.2arm(dat2, NS=NS)
    y    <- mnma:::gen.y.2arm(out1$y, out2$y, NS=NS)
    var1 <- out1$var
    var2 <- out2$var

    input <- list(NS=NS,N2h=N2h,NT=NT,ref=ref,T1=T1,T2=T2,y=y,var1=var1,var2=var2)

    if (parameters=="simple"){
      parameters <- c( "dOUT1","dOUT2","rho1","psi1","psi2")
    } else if (parameters=="detail"){
      parameters <- c( "ddOUT1","ddOUT2", "Eff.O1", "Eff.O2","rho1","psi1","psi2")
    } else if (parameters!="detail" & parameters!="simple"){
      stop("Choose either 'simple' or 'detail' for the list of parameters")
    }

    source <- list(data = input, parameters = parameters, noa=noa)
    class(source) <- "mnma.model"
    return(source)
  }
}