Emergency Department Utilization in Cancer Care

Emergency department (ED) utilisation among patients with cancer has emerged as a critical component of oncological care, reflecting both the complex clinical trajectories of malignancy and the challenges of symptom management. Across diverse health systems, cancer-related ED visits account for approximately 3–4 per cent of all attendances, with rates rising in parallel with increasing cancer incidence and survival. Presentations range from pain, dyspnoea and febrile episodes to treatment-related complications such as neutropenia, sepsis and dehydration. Nearly two-thirds of these visits lead to hospital admission, often with prolonged lengths of stay and substantial in-hospital mortality. Risk factors for ED attendance include advanced stage, recent systemic therapy, poor performance status and sociodemographic determinants. In response, multidisciplinary initiatives aim to optimise outpatient symptom control, define preventable presentations and deploy predictive tools. Emerging approaches combine real-time electronic health record data with machine learning to identify high-risk individuals and prompt pre-emptive interventions. Such strategies promise to reduce avoidable ED use, enhance patient experience and alleviate the burden on acute care services while maintaining safety across the cancer continuum.

Research from Nature Portfolio

Recent studies have demonstrated the potential of artificial intelligence in forecasting 30-day ED visits among oncology patients. One report describes a variational autoencoder k-nearest neighbours algorithm applied to large-scale electronic health record data, achieving an area under the receiver-operating characteristics curve of approximately 0.80 and maintaining equitable performance across demographic and disease subgroups during prospective monitoring. A national cross-sectional analysis of ED utilisation in Korea revealed a rising age- and sex-standardised incidence of cancer-related visits from 522 to 642 per 100 000 population between 2015 and 2019. Lung, liver and colorectal malignancies accounted for the largest proportions, with pneumonia, gastroenteritis and fever among the principal reasons for attendance; over half of these visits resulted in hospital admission and nearly 10 per cent in-hospital mortality. Another investigation of systemic therapy-related complications in the United States, using a national inpatient sample, found an 8.1 per cent annual increase in hospitalisations due to treatment toxicity between 2005 and 2016. Anaemia, neutropenia and sepsis were the leading causes, collectively incurring billions of pounds in charges and substantial in-hospital mortality, underscoring the need for early symptom recognition and coordinated urgent care pathways.

Emergency Department Utilization in Cancer Care publication trend

The graph below shows the total number of articles in emergency department utilization in cancer care across all publications each year (not limited to Nature Index journals).

Technical terms

Oncologic emergency: An acute health condition arising directly from cancer or its treatment that requires immediate evaluation and intervention.

Predictive model: A statistical or algorithmic tool designed to estimate the probability of a clinical event based on patient-specific data.

Variational autoencoder: A neural network framework for unsupervised learning of latent data representations and generation of complex datasets.

k-nearest neighbours algorithm: A non-parametric method that classifies or predicts outcomes by comparing a data point to its closest peers in feature space.

Area under the receiver-operating characteristics curve (AUC): A metric that quantifies a model’s discriminative ability to separate different outcome classes, with higher values indicating better performance.

Systemic therapy: Cancer treatment administered through the bloodstream to target malignant cells throughout the body, including chemotherapy, targeted agents and immunotherapy.

Performance status: A clinical scale assessing a patient’s functional capacity and ability to carry out daily activities, often guiding treatment choices.

References

1. Ensuring fair, safe, and interpretable artificial intelligence-based prediction tools in a real-world oncological setting. Communications Medicine (2023).
2. Risk Prediction of Emergency Department Visits in Patients With Lung Cancer Using Machine Learning: Retrospective Observational Study. JMIR Medical Informatics (2023).
3. Recognizing the emergency department’s role in oncologic care: a review of the literature on unplanned acute care. Emergency Cancer Care (2022).
4. Epidemiologic trends in cancer-related emergency department utilization in Korea from 2015 to 2019. Scientific Reports (2021).
5. Socio-demographic and disease related characteristics associated with unplanned emergency department visits by cancer patients: a retrospective cohort study. BMC Health Services Research (2019).
6. EPICANCER—Cancer Patients Presenting to the Emergency Departments in France: A Prospective Nationwide Study. Journal of Clinical Medicine (2020).
7. Hospitalization rates for complications due to systemic therapy in the United States. Scientific Reports (2021).