🤖 AI Summary
This paper addresses the practical violation of conditional unconfoundedness—a core assumption in causal machine learning—by systematically analyzing the bias mechanisms of causal forests and X-learners under unmeasured confounding. It introduces negative control outcomes (NCOs) as a novel, pragmatic diagnostic tool for detecting unobserved confounding, marking the first application of NCOs for this purpose in causal ML. Through extensive simulation studies across diverse confounding structures, sample sizes, and degrees of treatment effect heterogeneity, the study demonstrates that unconfoundedness violations induce spurious heterogeneity detection. Although NCOs do not strictly satisfy theoretical identification conditions, they robustly flag subgroups severely affected by unmeasured confounding, thereby substantially improving the credibility of individualized treatment effect estimation. This work advances the integration of NCOs from sensitivity analysis into standard causal modeling pipelines, providing methodological foundations for robust causal machine learning.
📝 Abstract
Observational studies developing causal machine learning (ML) models for the prediction of individualized treatment effects (ITEs) seldom conduct empirical evaluations to assess the conditional exchangeability assumption. We aimed to evaluate the performance of these models under conditional exchangeability violations and the utility of negative control outcomes (NCOs) as a diagnostic. We conducted a simulation study to examine confounding bias in ITE estimates generated by causal forest and X-learner models under varying conditions, including the presence or absence of true heterogeneity. We simulated data to reflect real-world scenarios with differing levels of confounding, sample size, and NCO confounding structures. We then estimated and compared subgroup-level treatment effects on the primary outcome and NCOs across settings with and without unmeasured confounding. When conditional exchangeability was violated, causal forest and X-learner models failed to recover true treatment effect heterogeneity and, in some cases, falsely indicated heterogeneity when there was none. NCOs successfully identified subgroups affected by unmeasured confounding. Even when NCOs did not perfectly satisfy its ideal assumptions, it remained informative, flagging potential bias in subgroup level estimates, though not always pinpointing the subgroup with the largest confounding. Violations of conditional exchangeability substantially limit the validity of ITE estimates from causal ML models in routinely collected observational data. NCOs serve a useful empirical diagnostic tool for detecting subgroup-specific unmeasured confounding and should be incorporated into causal ML workflows to support the credibility of individualized inference.