In umich-cphds/SAMBA: Selection and Misclassification Bias Adjustment for Logistic Regression Models

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)
  version.github <- as.vector(read.dcf('DESCRIPTION')[, 'Version'])
  version.github <- gsub('-', '.', version.github)
  version.cran <- ""
  # get the version from the CRAN package mirror
  # cran.dcf.con <- url("https://raw.githubusercontent.com/cran/SAMBA/master/DESCRIPTION")
  # version.cran <- read.dcf(file = cran.dcf.con)[, "Version"]
  # close(cran.dcf.con)
  if (version.cran == version.github) {
      color <- "success"
  } else {
      color <- "informational"
  }

SAMBA

Health research using data from electronic health records (EHR) has gained popularity, but misclassification of EHR-derived disease status and lack of representativeness of the study sample can result in substantial bias in effect estimates and can impact power and type I error for association tests. Here, the assumed target of inference is the relationship between binary disease status and predictors modeled using a logistic regression model. SAMBA implements several methods for obtaining bias-corrected point estimates along with valid standard errors as proposed in Beesley and Mukherjee (2020), currently under review.

Installation

SAMBA can be downloaded from Github via the R Package devtools

devtools::install_github("umich-cphds/SAMBA", build_opts = c())

Vignette

Once you have SAMBA installed, you can type

vignette("UsingSAMBA")

in R to bring up a tutorial on SAMBA and how to use it.

Questions

For questions and comments about the implementation, please contact Alexander Rix ([e-mail]). For questions about the method, contact Lauren Beesley ([e-mail]).

Reference

Statistical inference for association studies using electronic health records: handling both selection bias and outcome misclassification Lauren J Beesley, Bhramar Mukherjee medRxiv 2019.12.26.19015859

umich-cphds/SAMBA documentation built on Feb. 10, 2020, 6:56 a.m.