When I try to run code below in R software, got error reported as "Error in…
- User
When I try to run code below in R software, got error reported as "Error in model.frame.default(data = new_patients, formula = Surv(survival_time, : variable lengths differ (found for 'age')" . Please help debug the code. The original code reporting error are like blow:
Load necessary libraries
library(survival)
Simulate data for 200 patients
set.seed(123) n <- 200 age <- rnorm(n, mean = 60, sd = 10) gender <- sample(0:1, n, replace = TRUE) survival_time <- rexp(n, rate = 0.1 * (1 + gender) * exp(-0.02 * age)) death_event <- rbinom(n, 1, prob = 0.8)
Create a data frame
data <- data.frame(age, gender, survival_time, death_event)
datagender, levels = c(0,1), labels = c("Male", "Female"))
Specify the time knot points for piece-wise hazard
knots <- quantile(survival_time, probs = seq(0, 1, length.out = 4), na.rm = TRUE)[-1] intervals <- quantile(survival_time, probs = seq(0, 1, length.out = 5)) data$time_interval <- as.numeric(as.factor(findInterval(survival_time, intervals)))
Fit the piece-wise exponential hazard model
piecewise_model <- coxph(Surv(survival_time, death_event) ~ age + gender + strata(time_interval), data = data)
Summary of piece-wise exponential hazard model
summary(piecewise_model)
Plot the predicted hazard function
predicted_hazard <- predict(piecewise_model, type = "expected", p = 0.5) plot(survival_time, predicted_hazard, ylim = range(predicted_hazard), xlab = "Survival Time", ylab = "Predicted Hazard", main = "Piecewise Hazard")
Simulate new data for 5 remaining patients
new_patients <- data.frame( age = c(55, 63, 70, 75, 62), gender = factor(c('Male', 'Female', 'Male', 'Female', 'Male'), levels = c("Male", "Female")), time_interval = c(2, 3, 4, 1, 2) )
Predict survival times for new patients
predicted_survival <- predict(piecewise_model, newdata = new_patients, type = "expected") print(predicted_survival)
- Assistant
It seems like the 'time_interval' column is missing in the 'new_patients' data frame. You need to calculate the time_interval for new_patients similar to how it has been created for the 'data' data frame. Here's the modified code:
# Load necessary libraries library(survival) # Simulate data for 200 patients set.seed(123) n <- 200 age <- rnorm(n, mean = 60, sd = 10) gender <- sample(0:1, n, replace = TRUE) survival_time <- rexp(n, rate = 0.1 * (1 + gender) * exp(-0.02 * age)) death_event <- rbinom(n, 1, prob = 0.8) # Create a data frame data <- data.frame(age, gender, survival_time, death_event) data$gender <- factor(data$gender, levels = c(0,1), labels = c("Male", "Female")) # Specify the time knot points for piece-wise hazard knots <- quantile(survival_time, probs = seq(0, 1, length.out = 4), na.rm = TRUE)[-1] intervals <- quantile(survival_time, probs = seq(0, 1, length.out = 5)) data$time_interval <- as.numeric(as.factor(findInterval(survival_time, intervals))) # Fit the piece-wise exponential hazard model piecewise_model <- coxph(Surv(survival_time, death_event) ~ age + gender + strata(time_interval), data = data) # Summary of piece-wise exponential hazard model summary(piecewise_model) # Plot the predicted hazard function predicted_hazard <- predict(piecewise_model, type = "expected", p = 0.5) plot(survival_time, predicted_hazard, ylim = range(predicted_hazard), xlab = "Survival Time", ylab = "Predicted Hazard", main = "Piecewise Hazard") # Simulate new data for 5 remaining patients new_patients <- data.frame( age = c(55, 63, 70, 75, 62), gender = factor(c('Male', 'Female', 'Male', 'Female', 'Male'), levels = c("Male", "Female")) ) # Calculate time_interval for new patients new_patients$time_interval <- as.numeric(as.factor(findInterval(survival_time, intervals))) # Predict survival times for new patients predicted_survival <- predict(piecewise_model, newdata = new_patients, type = "expected") print(predicted_survival)Now, this code should work without any errors.