confoundr: Diagnostics for Confounding of Time-Varying and Other Joint Exposures

Implements three covariate-balance diagnostics for time-varying confounding and selection-bias in complex longitudinal data, as described in Jackson (2016) <doi:10.1097/EDE.0000000000000547> and Jackson (2019) <doi:10.1093/aje/kwz136>. Diagnostic 1 assesses measured confounding/selection-bias, diagnostic 2 assesses exposure-covariate feedback, and diagnostic 3 assesses residual confounding/selection-bias after inverse probability weighting or propensity score stratification. All diagnostics appropriately account for exposure history, can be adapted to assess a particular depth of covariate history, and can be implemented in right-censored data. Balance assessments can be obtained for all times, selected-times, or averaged across person-time. The balance measures are reported as tables or plots. These diagnostics can be applied to the study of multivariate exposures including time-varying exposures, direct effects, interaction, and censoring.

Version: 1.2
Depends: R (≥ 3.1.0)
Imports: grid, gridExtra (≥ 2.3), scales (≥ 0.5.0), Rmpfr (≥ 0.7-0), dplyr (≥ 0.7.5), tidyr (≥ 0.8.1), ggplot2 (≥ 2.2.1), magrittr (≥ 1.5), stringr (≥ 1.3.1), purrr (≥ 0.2.5), rlang (≥ 0.2.1), stats
Suggests: covr, knitr, rmarkdown, testthat, broom (≥ 0.4.4)
Published: 2019-09-20
Author: John W. Jackson [aut, cre], Erin Schnellinger [ctb], Linda Valeri [ctb]
Maintainer: John W. Jackson <john.jackson at>
License: GPL-3
NeedsCompilation: no
Citation: confoundr citation info
Materials: README
CRAN checks: confoundr results


Reference manual: confoundr.pdf
Vignettes: Demo Vignette
Extended Vignette
Package source: confoundr_1.2.tar.gz
Windows binaries: r-devel:, r-release:, r-oldrel:
OS X binaries: r-release: confoundr_1.2.tgz, r-oldrel: confoundr_1.2.tgz
Old sources: confoundr archive


Please use the canonical form to link to this page.