I’m using the AdjustedSurv package in R to create adjusted survival curves with the IPTW_cox method. However, I’m running into an issue with setting the x-axis limits for the plot.
This is the code that I’m using:
# Apply g-computation with confidence intervals for adjusted survival curves
adjsurv_selected <- adjustedsurv(data = df_selected,
variable = "AAA_size",
ev_time = "Surv_years",
event = "DEAD",
method = "iptw_cox",
treatment_model = glm_mod)
# Plot the adjusted survival curves
p1 <- plot(adjsurv_selected,
conf_int = FALSE,
xlab = "Years",
ylab = "Adjusted Survival Probability",
risk_table = TRUE,
risk_table_stratify = TRUE,
risk_table_title = "AAA size",
risk_table_ylab = "AAA size",
title = "Adjusted Survival Curve Stratified by AAA Size",
legend.labs = levels(df_selected$AAA_size),
legend.title = "AAA size",
xlim = c(0,5), # This doesn't seem to work
ylim = c(0, 1))
set.seed(42) # For reproducibility
# Create a minimal dataset
data <- data.frame(
AAA_size = factor(rep(c("Medium", "Large"), each = 5)),
AGE = sample(50:80, 10, replace = TRUE),
GENDER = factor(rep(c("Male", "Female"), each = 5)),
renal_dysf = sample(c(0, 1), 10, replace = TRUE),
prior_chf_log = sample(c(0, 1), 10, replace = TRUE),
DIABETES = sample(c(0, 1), 10, replace = TRUE),
prior_hypertension = sample(c(0, 1), 10, replace = TRUE),
prior_copd = sample(c(0, 1), 10, replace = TRUE),
bmi_cat = sample(c("Normal", "Overweight", "Obese"), 10, replace = TRUE),
prior_smoking = sample(c(0, 1), 10, replace = TRUE),
prior_mi = sample(c(0, 1), 10, replace = TRUE),
preop_anemia = sample(c(0, 1), 10, replace = TRUE),
branches = sample(1:3, 10, replace = TRUE),
prior_asause = sample(c(0, 1), 10, replace = TRUE),
prior_p2y = sample(c(0, 1), 10, replace = TRUE),
prior_statinuse = sample(c(0, 1), 10, replace = TRUE),
physvol = sample(c(0, 1), 10, replace = TRUE),
Surv_years = runif(10, 0, 5),
DEAD = sample(c(0, 1), 10, replace = TRUE)
)
# Output the dataset with dput
dput(head(data))
# Load necessary libraries
library(survival)
library(survey)
# Fit a logistic regression model
glm_mod <- glm(AAA_size ~ AGE + GENDER + renal_dysf +
prior_chf_log + DIABETES + prior_hypertension +
prior_copd + bmi_cat + prior_smoking +
prior_mi + preop_anemia + branches + prior_asause +
prior_p2y + prior_statinuse + physvol,
data = data, family = "binomial")
# Apply g-computation with confidence intervals for adjusted survival curves
adjsurv_selected <- adjustedsurv(data = data,
variable = "AAA_size",
ev_time = "Surv_years",
event = "DEAD",
method = "iptw_cox",
treatment_model = glm_mod)
# Plot the adjusted survival curves
p1 <- plot(adjsurv_selected,
conf_int = FALSE,
xlab = "Years",
ylab = "Adjusted Survival Probability",
risk_table = TRUE,
risk_table_stratify = TRUE,
risk_table_title = "AAA size",
risk_table_ylab = "AAA size",
title = "Adjusted Survival Curve Stratified by AAA Size",
legend.labs = levels(data$AAA_size),
legend.title = "AAA size",
xlim = c(0,5),
ylim = c(0, 1))
I noticed that while there is an argument for ylim, there doesn’t seem to be a similar argument for xlim. It appears that the x-axis is automatically set to the longest observed time in my dataset.
Despite specifying xlim = c(0, 5), I’m unable to adjust the x-axis limits, which are automatically set based on the longest observed time in my data.
Is there a way to manually set the x-axis limits in the plot.adjustedsurv() function or work around this limitation? Any guidance or suggestions would be greatly appreciated.
I want to set the X-lim to 5 years.
Thanks in advance!
New contributor
IsaFleurvG is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.
4
The issue is caused by the risk.table options, e.g.
Load data and create a model:
data <- data.frame(
AAA_size = factor(rep(c("Medium", "Large"), each = 5)),
AGE = sample(50:80, 10, replace = TRUE),
GENDER = factor(rep(c("Male", "Female"), each = 5)),
renal_dysf = sample(c(0, 1), 10, replace = TRUE),
prior_chf_log = sample(c(0, 1), 10, replace = TRUE),
DIABETES = sample(c(0, 1), 10, replace = TRUE),
prior_hypertension = sample(c(0, 1), 10, replace = TRUE),
prior_copd = sample(c(0, 1), 10, replace = TRUE),
bmi_cat = sample(c("Normal", "Overweight", "Obese"), 10, replace = TRUE),
prior_smoking = sample(c(0, 1), 10, replace = TRUE),
prior_mi = sample(c(0, 1), 10, replace = TRUE),
preop_anemia = sample(c(0, 1), 10, replace = TRUE),
branches = sample(1:3, 10, replace = TRUE),
prior_asause = sample(c(0, 1), 10, replace = TRUE),
prior_p2y = sample(c(0, 1), 10, replace = TRUE),
prior_statinuse = sample(c(0, 1), 10, replace = TRUE),
physvol = sample(c(0, 1), 10, replace = TRUE),
Surv_years = runif(10, 0, 5),
DEAD = sample(c(0, 1), 10, replace = TRUE)
)
# Load necessary libraries
library(ggplot2)
library(survival)
library(survey)
library(adjustedCurves)
# Fit a logistic regression model
glm_mod <- glm(AAA_size ~ AGE + GENDER + renal_dysf +
prior_chf_log + DIABETES + prior_hypertension +
prior_copd + bmi_cat + prior_smoking +
prior_mi + preop_anemia + branches + prior_asause +
prior_p2y + prior_statinuse + physvol,
data = data, family = "binomial")
# Apply g-computation with confidence intervals for adjusted survival curves
adjsurv_selected <- adjustedsurv(data = data,
variable = "AAA_size",
ev_time = "Surv_years",
event = "DEAD",
method = "iptw_cox",
treatment_model = glm_mod)
#> Warning in predict.lm(object, newdata, se.fit, scale = 1, type = if (type == :
#> prediction from rank-deficient fit; attr(*, "non-estim") has doubtful cases
Without the risk table (x limit 2 years):
# Plot the adjusted survival curves
plot(adjsurv_selected,
conf_int = FALSE,
xlab = "Years",
ylab = "Adjusted Survival Probability",
#risk_table = TRUE,
#risk_table_stratify = TRUE,
#risk_table_title = "AAA size",
#risk_table_ylab = "AAA size",
title = "Adjusted Survival Curve Stratified by AAA Size",
legend.labs = levels(data$AAA_size),
legend.title = "AAA size",
#xlim = c(0,5),
#ylim = c(0, 1)
) +
coord_cartesian(x = c(0, 2))
With the risk table (x limits ignored):
# Plot the adjusted survival curves
p1 <- plot(adjsurv_selected,
conf_int = FALSE,
xlab = "Years",
ylab = "Adjusted Survival Probability",
risk_table = TRUE,
risk_table_stratify = TRUE,
risk_table_title = "AAA size",
risk_table_ylab = "AAA size",
title = "Adjusted Survival Curve Stratified by AAA Size",
legend.labs = levels(data$AAA_size),
legend.title = "AAA size"
) +
scale_x_continuous(limits = c(0,2))
#> Loading required namespace: cowplot
#> Scale for x is already present.
#> Adding another scale for x, which will replace the existing scale.
p1
Created on 2024-09-18 with reprex v2.1.0
A potential solution is to change the limits ‘manually’, e.g.
# Plot the adjusted survival curves
p1 <- plot(adjsurv_selected,
conf_int = FALSE,
xlab = "Years",
ylab = "Adjusted Survival Probability",
risk_table = TRUE,
risk_table_stratify = TRUE,
risk_table_title = "AAA size",
risk_table_ylab = "AAA size",
title = "Adjusted Survival Curve Stratified by AAA Size",
legend.labs = levels(data$AAA_size),
legend.title = "AAA size"
)
#> Loading required namespace: cowplot
p1$coordinates$limits$x <- c(0, 0.6)
p1
Created on 2024-09-18 with reprex v2.1.0
Could you adapt this to your use-case?
2