Machine learning in clinical practice: Evaluation of an artificial intelligence tool after implementation

Abstract

ObjectiveArtificial intelligence (AI) has gradually found its way into healthcare, and its future integration into clinical practice is inevitable. In the present study, we evaluate the accuracy of a novel AI algorithm designed to predict admission based on a triage note after clinical implementation. This is the first of such studies to investigate real‐time AI performance in the emergency setting.MethodsThe novel AI algorithm that predicts admission using a triage note was translated into clinical practice and integrated within St Vincent's Hospital Melbourne's electronic emergency patient management system. The data were collected from 1 January 2021 to 17 August 2022 to evaluate the diagnostic accuracy of the AI system after implementation.ResultsA total of 77 125 ED presentations were included. The live AI algorithm has a sensitivity of 73.1% (95% confidence interval 72.5–73.8), specificity of 74.3% (73.9–74.7), positive predictive value of 50% (49.6–50.4) and negative predictive value of 88.7% (88.5–89) with a total accuracy of 74% (73.7–74.3). The accuracy of the system was at the lowest for admission to psychiatric units (34%) and at the highest for gastroenterology and medical admission (84% and 80%, respectively).ConclusionOur study showed the diagnostic evaluation of a real‐time AI clinical decision‐support tool became less accurate than the original. Although real‐time sensitivity and specificity of the AI tool was still acceptable as a decision‐support tool in the ED, we propose that continuous training and evaluation of AI‐enabled clinical support tools in healthcare are conducted to ensure consistent accuracy and performance to prevent inadvertent consequences.

Modeling the influence of attitudes, trust, and beliefs on endoscopists’ acceptance of artificial intelligence applications in medical practice

“…With the vast amount of AI software being developed, it will also be important for healthcare professionals to know about the software that will be utilized. New AI and systems are constantly being developed and applied in various fields of healthcare, such as in breast cancer medical imaging (31,32), tracking patient's medical history (33), and patient management systems (34). In educating both present and future healthcare practitioners, the acceptance of AI can be cultivated.…”

MoCab: A Framework for the Deployment of Machine Learning Models across Health Information Systems

“…The concept of continuous training is crucial in the realm of machine learning, particularly in maintaining the accuracy of models in dynamic environments. This process involves periodically updating the model using transfer learning with new data to account for possible changes in the underlying data patterns or operational context [5]. The MoCab Model Retraining Center synthesizes the retraining process into six modules [26]: Scheduler, Data Retrieval Parser, Data Transformation Bundler, Model Trainer, Model Evaluator, and Model Register.…”

“…The increasing adoption of predictive models in clinical practice has highlighted the need for a unified management system capable of streamlining workflows [4], [5]. Integrating health information systems (HISs) with predictive models is imperative to automate and streamline prediction processes.…”

“…The precision of clinical decision support (CDS) systems is based on a regular update of predictive models with new data. This is essential to counteract the degradation of performance caused by outdated information [5]. The shift from traditional evidence-based medicine to "medicine-based evidence" underscores the importance of considering a broader data context surrounding patients [8], [9].…”

“…The shift from traditional evidence-based medicine to "medicine-based evidence" underscores the importance of considering a broader data context surrounding patients [8], [9]. In this context, continuous training of machine learning models assumes vital importance [5], [10], [11], empowering these models to adapt to emerging trends and evolving patterns, thus ensuring precise predictions [12].…”