This study conducts design science research (DSR) aimed at addressing the challenge of improving clinical decision support (CDS) in Emergency Departments (EDs). Making timely decisions by physicians to meet the diverse demands within hospital EDs is a challenge as the inadequacies of current Hospital Information Systems (HIS) can host fragmented Electronic Health Record (EHR) data across multiple sub-systems. The advancement of Information Technology (IT), notably machine learning (ML), brings new initiatives for addressing this issue. To tackle this complexity, this research deploys DSR to design a novel system integration architecture. This architecture provides a comprehensive framework that orchestrates HIS integration across multiple strata, encompassing the business, application, and technology layers, to facilitate the integration of ML-based Clinical Decision Support Systems (CDSS) within the ED setting. Two design artefacts are developed as outcomes of DSR: firstly, the development of the system integration architecture, and secondly, the creation of the ML-based CDSS. The former serves as a meticulously designed blueprint for HIS integration, ensuring the effective functioning of ML-based CDSS at the point of care. The latter represents a pioneering CDSS system, harnessing the power of ML algorithms to furnish real-time, context-aware decision support. The impact of the ML-based CDSS on ED efficiency is subjected to rigorous evaluation. This evaluation is conducted by leveraging historical Electronic Health Record (EHR) data within a simulated ED environment. The simulation, calibrated with parameters drawn from a real-world hospital ED setting, yields promising results. These findings underscore the feasibility and the manifold benefits of integrating ML-based CDSS to augment ED efficiency. Furthermore, this research makes a noteworthy contribution to the theoretical underpinnings of information system design. It achieves this by pioneering the development of a novel system integration architecture. This architecture serves as a bridge, alleviating the knowledge gap that traditionally separates HIS integration from ML-based CDSS. The study advances the understanding of strategic design and integration of hospital information systems to support ML and CDSS. Ultimately, this advancement holds the potential to catalyse substantial improvements in patient care and outcomes, especially within the intricate and high-stakes environment of the Emergency Department.