Quantitative bias analysis in practice: Review of software for regression with unmeasured confounding

Abstract

AbstractFailure to appropriately account for unmeasured confounding may lead to erroneous conclusions. Quantitative bias analysis (QBA) can be used to quantify the potential impact of unmeasured confounding or how much unmeasured confounding would be needed to change a study’s conclusions. Currently, QBA methods are not routinely implemented, partly due to a lack of knowledge about accessible software. We review the latest developments in QBA software between 2011 to 2021 and compare five different programs applicable when fitting a linear regression:treatSens, causalsens, sensemakr, EValue, andkonfound. We illustrate application of these programs to two datasets and provide code to assist analysts in future use of these software programs. Our review found 21 programs with most created post 2016. All are implementations of a deterministic QBA, and the majority are available in the free statistical software environment R. Many programs include features such as benchmarking and graphical displays of the QBA results to aid interpretation. Out of the five programs we compared,sensemakrperforms the most detailed QBA and includes a benchmarking feature for multiple unmeasured confounders. The diversity of QBA methods presents challenges to the widespread uptake of QBA among applied researchers. Provision of detailed QBA guidelines would be beneficial.