Mirror Slope Plots for Two-Arm Clinical Trial Data

Creates mirrored slope plots comparing two treatment arms over a single observation interval. Both arms share a common x-axis and converge at a central endpoint, making the between-arm gap the primary visual quantity. Returns standard 'ggplot2' objects for full extensibility. The display was designed by Kati Maharry and is described in Maharry and Sullivan (2026, manuscript in preparation); it was first used in the supplementary appendix of McDonald and colleagues (2026) .

mirrorslope

Mirror slope plots for two-arm clinical trial data

A mirror slope plot compares two treatment arms over a single observation interval. Both arms share one x-axis and converge at a central endpoint, making the between-arm gap at the endpoint the primary visual quantity. The reference arm is drawn on the left, the comparator on the right, and both mean lines meet at the centre.

mirror_slope_plot() returns a standard ggplot2 object, so the full ggplot2 ecosystem remains available for customisation.

Installation

# install.packages("devtools")
devtools::install_github("Sullivanstatistics/mirrorslope")

Quick start

library(mirrorslope)

# A long analysis dataset: one row per subject per visit
df <- data.frame(
  USUBJID = rep(paste0("S", 1:40), each = 2),
  TRTP    = rep(c("Placebo", "Active"), each = 40),
  AVISIT  = rep(c("Baseline", "Week 24"), times = 40),
  AVAL    = c(rbind(rep(0, 40),
                    c(rnorm(20, -10, 5), rnorm(20, -6, 5))))
)

df |>
  prep_mirror_data(
    trt_levels    = c("Placebo", "Active"),
    time_baseline = "Baseline",
    time_endpoint = "Week 24",
    value_label   = "% Change from Baseline"
  ) |>
  add_summary(
    estimates  = c("Placebo" = -10.2, "Active" = -5.9),
    stat_label = "LS-Mean"
  ) |>
  add_difference(estimate = 4.3, ci_lower = 1.2, ci_upper = 7.4,
                 p_value = 0.011) |>
  mirror_slope_plot(
    trt_colors      = mirror_palette("colorblind"),
    show_difference = TRUE,
    show_slowing    = TRUE,
    title           = "Percent Change from Baseline at Week 24"
  )

The pipeline

The plot is built in composable, pipe-friendly stages. Each returns a mirror_data object that stores everything available; the plot call exposes only what is requested.

Function	Role
`prep_mirror_data()`	Validate and structure the analysis data
`add_summary()`	Attach group-level estimates and baselines
`add_precision()`	Attach per-arm SE or confidence intervals (optional)
`add_difference()`	Attach the between-arm difference and p-value (optional)
`mirror_slope_plot()`	Render the figure

Helpers: mirror_palette() for colours (including a colourblind-safe pair) and theme_mirror() for a publication theme.

Documentation

Three vignettes are included:

Introduction to mirrorslope — the core walkthrough.
Clinical workflow with ADaM data — a realistic data-preparation pattern, including merging population flags from a subject-level dataset.
Customising the plot — theming, palettes, and combining panels.

Model-agnostic by design

The package does not fit statistical models. Estimates, intervals, and p-values are computed upstream (for example from a mixed model for repeated measures) and passed in. This keeps the tool independent of any particular analysis method and keeps every figure traceable to the values supplied. The mirror_data structure conforms to a versioned specification shared with a companion SAS macro so that both tools produce equivalent figures.

Citation and credit

The mirrored slopegraph was designed by Kati Maharry (Capricor Therapeutics). The display is described in Maharry K, Sullivan A. The Mirrored Slopegraph: A Back-to-Back Composition for Comparing Two-Arm Change-from-Baseline Trajectories (2026, manuscript in preparation), and was first used in the supplementary appendix of McDonald CM et al., Lancet 2026;408:721-733, doi:10.1016/S0140-6736(26)01385-1. Run citation("mirrorslope") for a citation entry.