splinetrials (version 0.1.1)

ncs_plot_means: Plot Actual and Predicted Response Variable Means by Study Arm.

Description

This function accepts a data set, probably produced by ncs_analysis(), and it uses ggplot2 to produce a panel of plots, one for each study arm. The time variable is along the x-axis, and the response variable is along the y-axis. The actual means of the response variable are points plotted in one color, and the modeled means are plotted in another color. Each point also has its confidence interval plotted.

Usage

ncs_plot_means(
  data,
  arm = "arm",
  time = "time",
  est = "est",
  lower = "lower",
  upper = "upper",
  model_est = "response_est",
  model_lower = "response_lower",
  model_upper = "response_upper"
)

Value

An object returned by ggplot2::ggplot().

Arguments

data

(data frame)
a data frame, probably produced by ncs_analysis(), containing the actual and predicted means. Each row should have a unique combination of arm and time.

arm

(string)
the name of the study arm variable in data. There will be a separate plot produced for each study arm.

time

(string)
the name of the time or visit variable in data. These values correspond to the x-axis.

est, lower, upper

(string)
the name of the variables in data containing the actual response variable's mean and confidence interval bounds. These values correspond to the y-axis.

model_est, model_lower, model_upper

(string)
the name of the variables in data containing the predicted response variable's mean and confidence interval bounds. These values correspond to the y-axis.

Examples

# Create a usable data set out of mmrm::fev_data
fev_mod <- mmrm::fev_data
fev_mod$VISITN <- fev_mod$VISITN * 10
fev_mod$time_cont <- fev_mod$VISITN + rnorm(nrow(fev_mod))
fev_mod$obs_visit_index <- round(fev_mod$time_cont)

# Analysis result data set
ncs_data_results <-
  ncs_analysis(
    data = fev_mod,
    response = FEV1,
    subject = USUBJID,
    arm = ARMCD,
    control_group = "PBO",
    time_observed_continuous = time_cont,
    df = 2,
    time_observed_index = obs_visit_index,
    time_scheduled_continuous = VISITN,
    time_scheduled_baseline = 10,
    time_scheduled_label = AVISIT,
    covariates = ~ FEV1_BL + RACE,
    cov_structs = c("ar1", "us")
  )

ncs_plot_means(ncs_data_results)