Deep ensemble multitask classification of emergency medical call incidents combining multimodal data improves emergency medical dispatch

Abstract

The objective of this work was to develop a predictive model to aid non-clinical dispatchers to classify emergency medical call incidents by their life-threatening level (yes/no), admissible response delay (undelayable, minutes, hours, days) and emergency system jurisdiction (emergency system/primary care) in real time. We used a total of 1 244 624 independent incidents from the Valencian emergency medical dispatch service in Spain, compiled in retrospective from 2009 to 2012, including clinical features, demographics, circumstantial factors and free text dispatcher observations. Based on them, we designed and developed DeepEMC 2 , a deep ensemble multitask model integrating four subnetworks: three specialized to context, clinical and text data, respectively, and another to ensemble the former. The four subnetworks are composed in turn by multi-layer perceptron modules, bidirectional long short-term memory units and a bidirectional encoding representations from transformers module. DeepEMC 2 showed a macro F1score of 0.759 in life-threatening classification, 0.592 in admissible response delay and 0.757 in emergency system jurisdiction. These results show a substantial performance increase of 12.5%, 17.5% and 5.1%, respectively, with respect to the current in-house triage protocol of the Valencian emergency medical dispatch service. Besides, DeepEMC 2 significantly outperformed a set of baseline machine learning models, including naive bayes, logistic regression, random forest and gradient boosting (α=0.05). Hence, DeepEMC 2 is able to: 1) capture information present in emergency medical calls not considered by the existing triage protocol, and 2) model complex data dependencies not feasible by the tested baseline models. Likewise, our results suggest that most of this unconsidered information is present in the free text dispatcher observations. To our knowledge, this study describes the first deep learning model undertaking emergency medical call incidents classification. Its adoption in medical dispatch centers would potentially improve emergency dispatch processes, resulting in a positive impact in patient wellbeing and health services sustainability.

Deep continual multitask out-of-hospital incident severity assessment from changing clinical features

“…The initial experience comprised data solely from the years spanning from 2009 to 2012 and, subsequently, each ensuing experience corresponded to a specific year (2014)(2015)(2016)(2017)(2018)(2019). This partitioning strategy was selected due to forthcoming architectural and training variations within the Clinical network of the DeepECM 2 model [11]. This network was trained collectively on the entire 2009-2012 data batch, rather than in a year-by-year manner.…”

“…Considering the outcomes detailed in [11], where Deep Learning models exhibited superior performance compared to other Machine Learning approaches, the focus in this work remains on models of a similar nature. However, the present study excludes the adoption of recurrent architectures or other sequential models like the Transformer [24].…”

“…In the context of the Valencian Region, a deep multitask ensemble model named DeepEMC 2 was developed recently [11]. This model is constituted by four main deep neural networks, each one specialized on a specific data type: the Context network , the Clinical network , the Text network and the Ensemble network .…”

Artificial Intelligence in Emergency Medicine: Viewpoint of Current Applications and Foreseeable Opportunities and Challenges

“…Most of the current dispatch algorithms are rule based or encompass a human review of rule-based algorithms [ 39 ]. To date, 2 retrospective studies have shown that statistical machine learning and deep learning can improve or outperform rule-based algorithms [ 51 , 52 ]. Further validation and impact studies are needed to improve the current dysfunctional EMD triage, and AI should be considered for enhancing the dispatch algorithms.…”

Clinical Decision Support Systems for Palliative Care Referral: Design and Evaluation of Frailty and Mortality Predictive Models

“…The author contributed to the development of predictive models to aid dispatching emergency medical calls by classifying them in terms of life-threatening level, admissible response delay and emergency system jurisdiction [Ferri et al, 2021]. In addition, the author is participating in the CANCERLESS project, which is expected to develop micro-simulation models to evaluate the best policies to help the homeless population with cancer prevention and treatment.…”