Skip to contents

Create Love Plot for SMD Visualization

Source: R/gba_utils.R
create_love_plot.Rd

Creates a Love plot (also known as a covariate balance plot) showing standardized mean differences for baseline variables with threshold reference lines. This visualization is commonly used in HTA dossiers and propensity score analysis to assess covariate balance.

Usage

create_love_plot(
  balance_assessment,
  threshold = 0.1,
  show_ci = TRUE,
  title = "Standardized Mean Differences",
  xlab = "Standardized Mean Difference",
  color_by_type = TRUE,
  sort_by = c("abs_smd", "smd", "name", "none"),
  colors = NULL,
  point_size = 3,
  base_size = 11
)

Arguments

balance_assessment

A BalanceAssessment object from assess_baseline_balance(), or a data frame with columns variable, smd, and optionally ci_lower, ci_upper, var_type.

threshold

Numeric. SMD threshold for reference lines (default: 0.1). Vertical lines are drawn at +/- threshold. When balance_assessment is a BalanceAssessment object, its stored threshold is used regardless of this parameter.

show_ci

Logical. Show confidence intervals as error bars (default: TRUE).

title

Character. Plot title (default: "Standardized Mean Differences").

xlab

Character. X-axis label (default: "Standardized Mean Difference").

color_by_type

Logical. Color points by variable type (continuous vs categorical) when available (default: TRUE).

sort_by

Character. How to sort variables: "abs_smd" (default), "smd", "name", or "none".

colors

Named character vector of colors for "continuous", "categorical", and "threshold" elements. NULL uses defaults.

point_size

Numeric. Size of points (default: 3).

base_size

Numeric. Base font size (default: 11).

Value

A ClinicalPlot object containing a ggplot2 Love plot.

Details

The Love plot displays each baseline covariate on the y-axis with its SMD on the x-axis. Vertical reference lines at +/- threshold help identify imbalanced covariates.

Points falling outside the threshold lines indicate potential baseline imbalance that may require adjustment in the analysis.

References

Love, T. E. (2004). Demonstrating balance between groups in observational studies. Presentation at the Cleveland Chapter of the American Statistical Association.

See also

assess_baseline_balance(), add_smd_to_table()

Examples

if (FALSE) { # \dontrun{
# From BalanceAssessment object
balance <- assess_baseline_balance(
  data = adsl,
  trt_var = "TRT01P",
  continuous_vars = c("AGE", "WEIGHT"),
  categorical_vars = c("SEX", "RACE")
)
love_plot <- create_love_plot(balance)

# From data frame
smd_data <- data.frame(
  variable = c("Age", "Sex", "Weight"),
  smd = c(0.05, -0.12, 0.08),
  ci_lower = c(-0.02, -0.20, 0.01),
  ci_upper = c(0.12, -0.04, 0.15)
)
create_love_plot(smd_data, threshold = 0.1)
} # }