Title: | Assessing Heterogeneity in Surrogacy Using Censored Data |
---|---|
Description: | Provides functions to assess and test for heterogeneity in the utility of a surrogate marker with respect to a baseline covariate using censored (survival data), and to test for heterogeneity across multiple time points. More details will be available in the future in: Parast, L., Tian L, Cai, T. (2023) "Assessing Heterogeneity in Surrogacy Using Censored Data." Under Review. |
Authors: | Layla Parast |
Maintainer: | Layla Parast <[email protected]> |
License: | GPL |
Version: | 1.0 |
Built: | 2025-02-25 02:44:31 UTC |
Source: | https://github.com/laylaparast/hetsurrsurv |
Example data
data("example.data")
data("example.data")
A list with 10 elements representing 2000 observations from a treatment group and 1500 observations from a control group:
s1
the surrogate marker in the treatment group
s0
the surrogate marker in the control group
w1
the baseline covariate of interest in the treatment group
w0
the baseline covariate of interest in the control group
d1
the event indicator in the treatment group
d0
the event indicator in the control group
x1
the observed event time in the treatment group
x0
the observed evenet time in the control group
w1_cat
the discrete baseline covariate of interest in the treatment group
w0_cat
the discrete baseline covariate of interest in the treatment group
data(example.data) names(example.data)
data(example.data) names(example.data)
Assesses heterogeneity in the utility of a surrogate marker with respect to a baseline covariate using censored (survival data) by estimates the proportion of treatment effect explained by the surrogate marker as a function of a baseline covariate, w
R.main.estimate(xone, xzero, deltaone, deltazero, sone, szero, wone, wzero, w.grd, myt, landmark, type = "cont", var = FALSE, test = FALSE, extrapolate = T, h.0 = NULL, h.1 = NULL, h.w = NULL, h.s = NULL, h.w.1 = NULL)
R.main.estimate(xone, xzero, deltaone, deltazero, sone, szero, wone, wzero, w.grd, myt, landmark, type = "cont", var = FALSE, test = FALSE, extrapolate = T, h.0 = NULL, h.1 = NULL, h.w = NULL, h.s = NULL, h.w.1 = NULL)
xone |
x1, observed event time in the treated group |
xzero |
x0, observed event time in the control group |
deltaone |
delta1, event indicator in the treated group |
deltazero |
delta0, event indicator in the control group |
sone |
s1, surrogate marker in the treated group |
szero |
s0, surrogate marker in the control group |
wone |
w1, baseline covariate in the treated group |
wzero |
w0, baseline covariate in the control group |
w.grd |
grid for w where estimation will be provided |
myt |
t of interest |
landmark |
t0, landmark time |
type |
options are "cont" or "discrete"; type of baseline covariate, default is "cont" |
var |
TRUE or FALSE, if variance/standard error estimates are wanted |
test |
TRUE or FALSE, if test for heterogeneity is wanted wanted |
extrapolate |
TRUE or FALSE |
h.0 |
bandwidth |
h.1 |
bandwidth |
h.w |
bandwidth |
h.s |
bandwidth |
h.w.1 |
bandwidth |
A list is returned:
w.values |
grid for w where estimation is provided |
R.s.w |
The propoportion of treatment effect explained as a function of the baseline covariate, w |
delta.w |
The treatment effect as a function of the baseline covariate, w |
delta.s.w |
The residual treatment effect as a function of the baseline covariate, w |
sd.R |
Standard error estimate of R.s.w |
sd.delta |
Standard error estimate of delta.w |
sd.delta.s |
Standard error estimate of delta.s.w |
pval.omnibus |
p-value from the omnibus test for heterogeneity |
pval.con |
p-value from the conservative omnibus test for heterogeneity |
Layla Parast
Parast, L., Tian L, Cai, T. (2023) "Assessing Heterogeneity in Surrogacy Using Censored Data." Under Review.
data(example.data) names(example.data) R.main.estimate(xone=example.data$x1, xzero=example.data$x0, deltaone=example.data$d1, deltazero=example.data$d0, sone=log(example.data$s1), szero=log(example.data$s0), wone=log(example.data$w1), wzero=log(example.data$w0), w.grd=log(seq(0.1,0.9, length=25)), myt=1, landmark=0.5) R.main.estimate(xone=example.data$x1, xzero=example.data$x0, deltaone=example.data$d1, deltazero=example.data$d0, sone=log(example.data$s1), szero=log(example.data$s0), wone=example.data$w1_cat, wzero=example.data$w0_cat, myt=1, landmark=0.5,type = "discrete", w.grd = c(1,2,3,4)) #computationally intensive R.main.estimate(xone=example.data$x1, xzero=example.data$x0, deltaone=example.data$d1, deltazero=example.data$d0, sone=log(example.data$s1), szero=log(example.data$s0), wone=log(example.data$w1), wzero=log(example.data$w0), w.grd=log(seq(0.1,0.9, length=25)), myt=1, landmark=0.5, test=TRUE)
data(example.data) names(example.data) R.main.estimate(xone=example.data$x1, xzero=example.data$x0, deltaone=example.data$d1, deltazero=example.data$d0, sone=log(example.data$s1), szero=log(example.data$s0), wone=log(example.data$w1), wzero=log(example.data$w0), w.grd=log(seq(0.1,0.9, length=25)), myt=1, landmark=0.5) R.main.estimate(xone=example.data$x1, xzero=example.data$x0, deltaone=example.data$d1, deltazero=example.data$d0, sone=log(example.data$s1), szero=log(example.data$s0), wone=example.data$w1_cat, wzero=example.data$w0_cat, myt=1, landmark=0.5,type = "discrete", w.grd = c(1,2,3,4)) #computationally intensive R.main.estimate(xone=example.data$x1, xzero=example.data$x0, deltaone=example.data$d1, deltazero=example.data$d0, sone=log(example.data$s1), szero=log(example.data$s0), wone=log(example.data$w1), wzero=log(example.data$w0), w.grd=log(seq(0.1,0.9, length=25)), myt=1, landmark=0.5, test=TRUE)
Tests for heterogeneity across multiple timepoints
test.multiplet(t.mult, xone, xzero, deltaone, deltazero, sone, szero, wone, wzero, w.grd, landmark, extrapolate = T, h.0 = NULL, h.1 = NULL, h.w = NULL, h.s = NULL, h.w.1 = NULL,type = "cont")
test.multiplet(t.mult, xone, xzero, deltaone, deltazero, sone, szero, wone, wzero, w.grd, landmark, extrapolate = T, h.0 = NULL, h.1 = NULL, h.w = NULL, h.s = NULL, h.w.1 = NULL,type = "cont")
t.mult |
Vector of time points |
xone |
x1, observed event time in the treated group |
xzero |
x0, observed event time in the control group |
deltaone |
delta1, event indicator in the treated group |
deltazero |
delta0, event indicator in the control group |
sone |
s1, surrogate marker in the treated group |
szero |
s0, surrogate marker in the control group |
wone |
w1, baseline covariate in the treated group |
wzero |
w0, baseline covariate in the control group |
w.grd |
grid for w where estimation will be provided |
landmark |
t0, landmark time |
extrapolate |
TRUE or FALSE |
h.0 |
bandwidth |
h.1 |
bandwidth |
h.w |
bandwidth |
h.s |
bandwidth |
h.w.1 |
bandwidth |
type |
options are "cont" or "discrete"; type of baseline covariate, default is "cont" |
A list is returned:
pval.multi |
p-value for omnibus test |
pval.con.multi |
p-value for conservative omnibus test (only applicable for continuous W) |
Layla Parast
Parast, L., Tian L, Cai, T. (2023) "Assessing Heterogeneity in Surrogacy Using Censored Data." Under Review.
data(example.data) names(example.data) #computationally intensive test.multiplet(t.mult = c(1,1.25,1.5), xone=example.data$x1, xzero=example.data$x0, deltaone=example.data$d1, deltazero=example.data$d0, sone=log(example.data$s1), szero=log(example.data$s0), wone=log(example.data$w1), wzero=log(example.data$w0), w.grd=log(seq(0.1,0.9, length=25)), landmark=0.5)
data(example.data) names(example.data) #computationally intensive test.multiplet(t.mult = c(1,1.25,1.5), xone=example.data$x1, xzero=example.data$x0, deltaone=example.data$d1, deltazero=example.data$d0, sone=log(example.data$s1), szero=log(example.data$s0), wone=log(example.data$w1), wzero=log(example.data$w0), w.grd=log(seq(0.1,0.9, length=25)), landmark=0.5)