causal_e_value is the VanderWeele-Ding E-value RR* + sqrt(RR*(RR* - 1)); unmeasured_conf_bias the Ding-VanderWeele bounding factor RR_UD RR_UY / (RR_UD + RR_UY - 1); causal_mediation_baron_kenny and
mediation_analysis the Baron-Kenny three-regression decomposition
(morie_baron_kenny), the latter after residualising baseline covariates;
frontdoor_adjustment (and front_door) Pearl's front-door formula for
discrete data; frontdoor_criterion Pearl's three front-door conditions
checked by d-separation in the moral ancestral graph; mediation_formula
Pearl's natural direct and indirect effects from cell means.
Usage
causal_e_value(RR, ci_lower = NULL, ci_upper = NULL, rare_outcome = TRUE)
unmeasured_conf_bias(RR_UD, RR_UY, RR_obs = NULL)
causal_mediation_baron_kenny(X, M, Y)
mediation_analysis(Y, T, M, X = NULL)
frontdoor_adjustment(x, z, y, at = NULL)
front_door(Y, X, M)
frontdoor_criterion(dag, X, Y, Z)
mediation_formula(x, m, y, x1 = NULL, x0 = NULL)Arguments
- RR
Risk ratio.
- ci_lower, ci_upper
Optional confidence limits.
- rare_outcome
Recorded flag (odds or hazard ratios approximate RR).
- RR_UD, RR_UY
Confounder-treatment and confounder-outcome risk ratios.
- RR_obs
Observed risk ratio.
- X, M, Y, T, y, m, x, z
Treatment, mediator, outcome (and covariates
Xinmediation_analysis).- at
Treatment levels to intervene on.
- dag
Named list of children or two-column edge matrix.
- Z
Candidate front-door set.
- x1, x0
Treatment levels contrasted.
References
VanderWeele, T. J. and Ding, P. (2017). Sensitivity analysis in observational research: introducing the E-value. Annals of Internal Medicine 167, 268-274. Ding, P. and VanderWeele, T. J. (2016). Sensitivity analysis without assumptions. Epidemiology 27, 368-377. Baron, R. M. and Kenny, D. A. (1986). The moderator-mediator variable distinction. Journal of Personality and Social Psychology 51, 1173-1182. Pearl, J. (2009). Causality, 2nd ed. Cambridge University Press. Pearl, J. (2001). Direct and indirect effects. Proceedings of UAI 17, 411-420.