table1中的P值列-使用table1函数

Question

NOTE - im referring to "table1" function, not "tableOne"

As you might know table1 package summarizes the data into table by putting up some variable against all the others (like how many males vs females - smoked, got the drug, etc..)

Very good guidelines about the table1 function here.

https://cran.r-project.org/web/packages/table1/vignettes/table1-examples.html

In lowest part of link site above you can see the writer making p-value column by using trick since the command is not built into table1. The thing is that its no problem with his code since he is only dealing with "male" and "female" (2 variables).. and im trying to do it with 4 variables.

Im fitting it so it can use my variables, im getting error and need help

ERROR

"argument is not numeric or logical: returning NACalling var(x) on a factor x is deprecated and will become an error. Use something like 'all(duplicated(x)[-1L])' to test for a constant vector.Error in if (stderr < 10 * .Machine$double.eps * max(abs(mx), abs(my))) stop("data are essentially constant") : missing value where TRUE/FALSE needed"

a <- c(0,13.3,14.0,14.7,15.8,999.0)

dat1$RDWquintiles<-cut(dat1$RDW_pre,a,include.lowest = TRUE)

z <- levels(dat1$RDWquintiles)
dat1$RDWquintiles <- factor(dat1$RDWquintiles,levels=c(z,factor("P-value")),labels=c("0,13.3","13.3,14","14,14.7","14.7,15.8","15.8,999","P-values" ))

dat1$RDWquintiles

table(dat1$RDWquintiles)


rndr <- function(x, name, ...) {
    if (length(x) == 0) {
        y <- dat1[[name]]
        s <- rep("", length(render.default(x=y, name=name, ...)))
        if (is.numeric(y)) {
            p <- t.test(y, dat1$RDWquintiles)$p.value
        } else {
            p <- chisq.test(table(y, droplevels(dat1$RDWquintiles)))$p.value
        }
        s[2] <- sub("<", "&lt;", format.pval(p, digits=3, eps=0.001))
        s
    } else {
        render.default(x=x, name=name, ... ,)
    }
}

rndr.strat <- function(label, n, ...) {
    ifelse(n==0, label, render.strat.default(label, n, ...))
}

table1(~ Age + Sex + LOS + Open + adm_PE |RDWquintiles,data=dat1, droplevels=F, render=rndr, render.strat=rndr.strat, overall=F)

Answer 1

You might need to check your data. I don't have your dataset so I can't reproduce your error. The only issue I see with having 4 groups to compare instead of 2 is that you cannot use a T-test for continuous variables, you would need to choose an appropriate test like ANOVA or Kruskal-Wallis to derive the P-value (in general, the P-value depends on the test used; selecting an appropriate test is an important consideration, but outside the scope of the current question).

Here I have adapted the example from the vignette (I don't repeat the whole code, just the relevant changes):

lalonde$age.quartiles <- eqcut(lalonde$age, 4, varlabel="Age")
lalonde$age.quartiles <- factor(lalonde$age.quartiles,
    levels=c(levels(lalonde$age.quartiles), "P-value"))

rndr <- function(x, name, ...) {
    if (length(x) == 0) {
        y <- lalonde[[name]]
        s <- rep("", length(render.default(x=y, name=name, ...)))
        if (is.numeric(y)) {
            p <- kruskal.test(y ~ lalonde$age.quartiles)$p.value
        } else {
            p <- chisq.test(table(y, droplevels(lalonde$age.quartiles)))$p.value
        }
        s[2] <- sub("<", "&lt;", format.pval(p, digits=3, eps=0.001))
        s
    } else {
        render.default(x=x, name=name, ...)
    }
}

table1(~ black + hispan + married + nodegree + re74 + re75 + re78 | age.quartiles,
    data=lalonde, droplevels=F, render=rndr, render.strat=rndr.strat, overall=F)

Note: this is not part of the original question, but I thought I would mention that the table1 package contains the function eqcut which can be conveniently used to create equal-sized groups such as tertiles, quartiles, etc.