Integrating Artifact Detection with Clinical Decision Support Systems (Preprint)
BACKGROUND Clinical decision support systems (CDSS) have the potential to lower patient mortality and morbidity rates. However, signal artifacts present in physiologic data affect the reliability and accuracy of CDSS. Moreover, patient monitors and other medical devices generate false alarms while processing artifactual data. This leads to alarm fatigue due to increased noise levels, staff disruption, and staff desensitization in busy critical care environments. Thereby, adversely affecting the quality of care at the patient bedside. Hence, artifact detection (AD) algorithms play a crucial role in assessing the quality of physiologic data and mitigating the impact of these artifacts. OBJECTIVE Recently, we developed a novel AD framework for integrating AD algorithms with CDSS. The framework was designed with features to support real-time implementation within critical care. In this research, we evaluate the framework and its features in a false alarm reduction study. We develop static framework component models followed by dynamic framework compositions to formulate four CDSS. We evaluate these formulations using neonatal patient data, and validate the six framework features of flexibility, reusability, signal quality indicator standardization, scalability, customizability, and real-time implementation support. METHODS We develop four exemplar static AD components with standardized requirements and provisions interfaces facilitating interoperability of framework components. These AD components are mixed and matched into four different AD compositions to mitigate artifacts. Each AD composition is integrated with a novel static clinical event detection (CED) component to formulate and evaluate dynamic CDSS for arterial oxygen saturation (SpO2) alarms generation. RESULTS With a sensitivity of 80%, the lowest achievable SpO2 false alarm rate is 39%. This demonstrates the utility of the framework in identifying the optimal dynamic composition to serve a given clinical need. CONCLUSIONS The framework features including reusability, signal quality indicator standardization, scalability, and customizability allow for novel CDSS formulations to be evaluated and compared. The optimal solution for a CDSS can then be hard-coded and integrated within clinical workflows for real-time implementation. Flexibility to serve different clinical needs and standardized component interoperability of the framework support the potential for real-time clinical implementation of AD.