Co-prescription of antiseizure medications (ASMs) with direct-acting oral anticoagulants (DOACs) is common due to frequent concurrency and causal links between epilepsy and the indications for anticoagulation. However, enzyme-inducing ASMs (EI-ASMs) may reduce absorption and accelerate metabolism of DOACs, potentially lowering DOAC levels and elevating thromboembolism risk. Further, the global rise in polypharmacy has increased the complexity of evaluating such theoretical drug-drug interactions (DDIs), given growing potential for interactions involving more than two drugs. Leveraging large-scale healthcare data and applying rigorous pharmacoepidemiologic methodologies, the presented research: (1) Employed active comparator, new user design approaches, with high-dimensional propensity score-based controls for observed confounders and proxies for unobserved confounders, to assess the rates of thromboembolic events in discrete epilepsy populations dispensed DOACs+EI-ASMs versus DOACs+non-enzyme inducing ASMs (NEI-ASMs); and, (2) Piloted a semi-automated, high-throughput case-crossover-based strategy for higher-order drug-drug-drug interaction (3DI) screening to explore the role of concomitant medications in adverse outcome occurrence among high susceptibility populations with ongoing exposures to DOACs+EI-ASMs, or with ongoing exposures to a mechanistically-distinct proposed DOAC+enzyme-inhibiting antihypertensive drug (EIH-AHD) interaction. From the DDI investigations, use of EI-ASMs (versus NEI-ASMs) with DOACs was not associated with significant differences in thromboembolic events among nationally representative epilepsy populations with atrial fibrillation (adjusted hazard ratio (aHR)=1.10, 95% CI: 0.82-1.46) or venous thromboembolism (aHR=1.11, CI: 0.81-1.51). From the 3DI screenings, 33 signals were generated based on a p-value threshold, and 9 signals were generated based on a false discovery rate q-value threshold among DOAC+EIH-AHD users with major bleeding events. Among DOAC+EI-ASM users with thromboembolic events, 3 signals were generated based on a p-value threshold, and no signals were generated based on a q-value threshold. The presented real-world evidence sought to contribute to needed foundations for prescribing decisions in the substantial global epilepsy population that requires anticoagulation, as well as to report signals that drive 3DI investigations and to advance paradigms for higher-order drug interaction screenings.