G-computation, propensity score-based methods, and targeted maximum likelihood estimator for causal inference with different covariates sets: a comparative simulation study

Abstract

controlling for confounding bias is crucial in causal inference. Distinct methods are currently employed to mitigate the effects of confounding bias. Each requires the introduction of a set of covariates, which remains difficult to choose, especially regarding the different methods. We conduct a simulation study to compare the relative performance results obtained by using four different sets of covariates (those causing the outcome, those causing the treatment allocation, those causing both the outcome and the treatment allocation, and all the covariates) and four methods: g-computation, inverse probability of treatment weighting, full matching and targeted maximum likelihood estimator. our simulations are in the context of a binary treatment, a binary outcome and baseline confounders. the simulations suggest that considering all the covariates causing the outcome led to the lowest bias and variance, particularly for g-computation. The consideration of all the covariates did not decrease the bias but significantly reduced the power. We apply these methods to two real-world examples that have clinical relevance, thereby illustrating the real-world importance of using these methods. We propose an R package RISCA to encourage the use of g-computation in causal inference. The randomised controlled trial (RCT) remains the primary design for evaluating the marginal (population average) causal effect of a treatment, i.e., the average treatment effect between two hypothetical worlds where: i) everyone is treated and ii) everyone is untreated 1. Indeed, a well-designed RCT with a sufficient sample size ensures the baseline comparability between groups, thus allowing the estimation of a marginal causal effect. Nevertheless, it is well established that RCT is performed under optimal circumstances (e.g., over-representation of treatment-adherent patients, low frequency of morbidity), which may be different from real-life practices 2. Observational studies have the advantage of limiting the issue of external validity, but treated and untreated patients are often non-comparable, leading to a high risk of confounding bias. To reduce such confounding bias, the vast majority of observational studies have been based on multivariable models (mainly linear, logistic, or Cox models), allowing for the direct estimation of conditional (subject-specific) effects, i.e., the average effect across sub-populations of subjects who share the same characteristics. Several

Prognostication for prelabor rupture of membranes and the time of delivery in nationwide insured women: development, validation, and deployment

“…To verify our assumption of PROM causality, we applied one of generalized (G) methods, which is IPW, for each causal factor. 63,64 This method is design for time-varying exposure. 65,66 However, we also conducted outcome regression for the causal inference since this method is one of the methods that are more common in use although do not work in general.…”

G-computation and machine learning for estimating the causal effects of binary exposure statuses on binary outcomes

“…As noted by VanderWeele and Shpitser 24 , investigators can identify the causes of exposure statuses or outcomes as potential covariates.Unfortunately, full knowledge of causal relationships is often unavailable. There is a growing literature about the best set of covariates to consider, and it recommends including all the covariates that cause the outcome 11,21,25 . The corresponding data-driven selection procedure for GC is straightforward since it corresponds to the predictors of the Q-model.…”

“…A large number simulation-based studies have compared several ML methods to obtain PSs 1,[6][7][8][9][10] . While the corresponding PS-based results were very encouraging, GC was compared to PS-based methods in the context of classical regression models and showed several advantages in terms of statistical power [11][12][13][14] and robustness of the estimates regardless of the set of included covariates 11 . However, simulation-based studies related to the use of ML for predicting outcomes in GC are infrequent.…”

Systematic human learning by literature and data mining for feature selection in machine learning

“…To verify of PROM causality, we applied one of the generalized (G) methods, i.e., IPW, for each causal factor. 7,8 This method was designed for time-varying exposures. 9,10 However, we also conducted outcome regressions for causal inferences, since this is one of the more commonly methods although it does not work in general.…”

“…Then, one of the generalized (G) methods, i.e., inverse probability weighting (IPW), is used to verify each causal factor using available data. 7,8 This method was designed for time-varying exposures which are typically available data from electronic health records. 9,10 Eventually, all causal factors are included in a prediction model to describe the explainability.…”