R/smoothHR.R

In smoothHR: Smooth Hazard Ratio Curves Taking a Reference Value

smoothHR <- function(data, time=NULL, time2=NULL, status=NULL, formula=NULL, coxfit, status.event=NULL) {
	ctype <- "FALSE";
	modelfit <- "TRUE";
	mydata2 <- deparse( substitute(data) );
	if ( missing(coxfit) ) {modelfit <- "FALSE";}
	if ( !missing(coxfit) ) {
		if ( !inherits(coxfit, "coxph") ) {stop("Argument coxfit must be of class coxph");}
	}
	if ( !is.data.frame(data) ) {stop("data must be of class data.frame");}
	if (modelfit == "TRUE") {
		if ( !inherits(coxfit, "coxph") ) {stop("Object coxfit must be of class coxph");}
		if ( is.null(coxfit$x) ) {stop("The argumment x in the coxph object is missing");}
		fit <- coxfit;
	}
	#if ( !missing(formula) & !missing(coxfit) ) {stop("....only one is requested");}
	if ( missing(data) ) {stop("The argumment data is missing");}
	if (missing(time) & modelfit == "FALSE") {stop("The argumment time is missing");}
	if (missing(status) & modelfit == "FALSE") {stop("The argumment status is missing");}
	if (missing(formula) & modelfit == "FALSE") {stop("The argumment formula is missing");}
	if (!missing(time2) & modelfit == "FALSE") {ctype <- "TRUE";}
	mydata <- data;
	if (modelfit == "FALSE") {
		p0 <- match(names(data),time, nomatch=0);
		p1 <- which(p0 == 1);
		ntime <- data[,p1];
		p2 <- match(names(data), status, nomatch=0);
		p3 <- which(p2 == 1);
		nstatus <- data[,p3];
		if (ctype == "TRUE") {
			p4 <- match(names(data), time2, nomatch=0);
			p5 <- which(p4 == 1);
			ntime2 <- data[,p5];
		}
		if ( !missing(time) ) {time <- ntime;}
		if ( !missing(time2) ) {time2 <- ntime2;}
		if ( !missing(status) ) {status <- nstatus;}
		#if ( !missing(status) & ( nlevels( as.factor(nstatus) ) != 2 | (levels( as.factor(nstatus) )[1] != "0" | levels( as.factor(nstatus) )[2] != "1") ) ) {stop("status must be of class factor with levels '0' and '1' with '1' for the event");}
		if ( !missing(time) & !is.numeric(ntime) & !is.integer(ntime) ) {stop("time must be of class numeric or integer");}
		if ( !missing(time2) ) {
			if ( !is.numeric(ntime2) & !is.integer(ntime2) ) {stop("time2 must be of class numeric or integer");}
		}
		if ( !missing(status) & missing(status.event) ) {status.event <- max(nstatus);}
		fmla <- attr(terms(formula), "term.labels");
		ncov <- length(fmla);
		colvar <- rep(0, ncov);
		for (k in 1:ncov) {
			if ( fmla[k] %in% names(data) ) {
				colvar[k] <- which(names(data) == fmla[k]);
			} else {
				for ( j in 1:ncol(data) ) {
					if ( any( grep(names(data)[j], fmla[k]) ) ) {colvar[k] <- j;}
				}
			}
		}
		if ( any(colvar == 0) ) {stop("'formula' must contain the right variables");}
		if (ctype == "TRUE") {
			covar <- as.formula( paste( " Surv(ntime,ntime2,nstatus)~ ", paste(fmla, collapse = "+") ) );
			covar2 <- as.formula( paste( " Surv(ntime,ntime2,nstatus==", status.event, ")~ ", paste(fmla, collapse = "+") ) );
			fit <- coxph(covar2, data = data, x=T);
		} else {
			covar <- as.formula( paste( " Surv(ntime,nstatus)~ ", paste(fmla, collapse = "+") ) );
			covar2 <- as.formula( paste(" Surv(ntime,nstatus==", status.event, ")~ ", paste(fmla, collapse = "+") ) );
			fit <- coxph(covar2, data=data, x=TRUE);
		}
	}
	phtest <- cox.zph(fit);
	dimt <- dim(phtest$table);
	p_value <- phtest$table[dimt[1],dimt[2]];
	if (p_value <= 0.05) {cat("Warning: a global p-value of ", p_value, "was obtained when testing the proportional hazards assumption", "\n");}
	a1 <- c();
	if (modelfit == "TRUE") {
		for(k in 1:dim(mydata)[2]) {
			if ( any( grep(names(mydata)[k], fit$call) ) ) {a1 <- c(a1,k);}
		}
		mydata <- mydata[,a1];
	}
	if (modelfit == "FALSE") {
		if (ctype == "TRUE") {a1 <- c(p1, p5, p3);}
		else {a1 <- c(p1, p3);}
		for (k in 1:dim(mydata)[2]) {
			if ( any( grep(names(mydata)[k], formula) ) ) {a1<-c(a1,k);}
		}
		mydata <- mydata[,a1];
	}
	if (modelfit == "TRUE") {
		if (as.name( toString(fit$call[[3]]) ) != mydata2) {cat("Warning: check if argument 'data' is the same used in coxfit", "\n");}
	}
	mydata <- na.omit(mydata);
	nv <- fit$nevent;
	if (ctype == "TRUE") {
		myd <- as.data.frame( table(mydata[,3]) );
		w.status.event <- which(myd[,1] == status.event);
		if (nv != table( as.factor(mydata[,3]) )[[w.status.event]]) {cat("Warning: check if argument 'data' is the same used in coxfit", "\n");}
	}
	object <- list(dataset=mydata, coxfit=fit, phtest=phtest);
	class(object) <- "HR";
	return(object);
} # smoothHR