Virtual Emergency Department Keeps Four in Five Patients Away From Hospital EDs, Landmark Data Show

Emergency departments around the world are drowning in demand, but one of the boldest experiments in digital medicine just delivered its strongest evidence yet that care delivered entirely by video can keep patients safely out of the hospital. In the first population-scale study of its kind, researchers linked more than 100,000 consultations with Australia’s Victorian Virtual Emergency Department to statewide hospital records and found that 81.5 percent of patients avoided a physical emergency department within 48 hours of their virtual visit. The findings, published in The Lancet Regional Health – Western Pacific, offer the clearest picture yet of whether virtual emergency care genuinely diverts demand or simply shifts problems downstream.

The Victorian Virtual Emergency Department, or VVED, began life in October 2020 as a pandemic-era initiative at Northern Health in Melbourne, designed to relieve pressure on a crowded physical emergency department by treating lower-acuity patients at home. Demand grew so quickly that by February 2022 it had expanded into a 24/7, state-wide service staffed around the clock by emergency nurses, emergency physicians, nurse practitioners and general practitioners. Today it ranks among the largest virtual emergency departments in the world, with more than 800,000 consultations logged by March 2026. During the study period, spanning July 2022 to June 2023, it handled 100,422 presentations, roughly 5.3 percent of all public emergency department presentations recorded across Victoria in the same window.

What sets the new analysis apart is its methodological scale. Previous evaluations of virtual emergency care have largely been confined to pilot programmes, single-site audits, or patient satisfaction surveys, leaving open the crucial question of what actually happens to patients after the video call ends. To answer that, the research team used the Centre for Victorian Data Linkage to connect VVED records to the Victorian Emergency Minimum Dataset, which captures every public emergency department presentation in the state, and the Victorian Admitted Episodes Dataset, covering all hospital admissions. The result was a linked cohort of 100,422 consultations from 80,161 individual patients, tracked across eight distinct referral pathways, from self-referrals through paramedic, ambulance, aged care, and urgent care centre referrals.

The headline numbers are striking. Three-quarters of all patients, 76 percent, were managed entirely at home following their virtual consultation. Only 15 percent were directed to a physical emergency department, and a mere 0.3 percent required direct hospital admission. Among those sent home, just 8.2 percent turned up at a physical emergency department within 48 hours, rising to 11.4 percent within seven days, with hospital admissions of 4.1 and 6.6 percent respectively. Those figures compare favourably with international benchmarks: re-presentation rates after discharge from physical emergency departments range from about 5 percent in Australia to more than 30 percent overseas, and British and Canadian virtual urgent care models have reported unplanned attendance rates of 10.5 and 13.7 percent respectively.

The researchers also found evidence that the service is changing the shape of demand across the entire state system. Recent analysis of more than 11 million Victorian emergency presentations between 2018 and 2024 documented a decline in low-acuity triage category 4 and 5 attendances, with avoidance strategies such as the VVED cited as a likely contributor. Victoria was also the only Australian state to record a fall in ambulance incidents in 2024–25, a 5.4 percent drop attributed in part to active management of lower-acuity cases, with the VVED now one of Ambulance Victoria’s largest alternative care pathways. Crucially, the low unplanned re-presentation rate suggests the avoidance reflects appropriate clinical selection rather than dangerous under-triage.

When the team dug into which patients were most likely to bounce back to a hospital, diagnosis emerged as the decisive factor. In multivariable mixed-effects logistic regression covering 65,860 presentations managed in the community, the only variable independently associated with unplanned emergency presentation within 48 hours was the primary ICD-10 diagnostic category. Conditions amenable to assessment by history and visual inspection performed especially well: skin and subcutaneous tissue disorders showed 73 percent lower odds of re-presentation compared with non-specific symptoms, ear conditions such as suspected middle ear infection showed 70 percent lower odds, and respiratory illnesses and minor injuries also fared significantly better. Patients coded under vague symptom categories, the authors note, may simply have had illnesses that had not yet declared themselves, with their higher return rates reflecting appropriate safety-netting advice rather than failures of virtual care.

The picture grows more nuanced for clinically complex patients arriving through healthcare professional referral pathways, including paramedics, residential aged care facilities, and urgent care centres. These referrals involve older, frailer patients with multiple chronic conditions, and when their condition deteriorated, hospital admission was often the appropriate outcome. Among patients sent home from these pathways, admission rates reached 6.3 percent for ambulance referrals and 6.9 percent for aged care referrals within 48 hours. Yet even here, three-quarters were successfully managed at home without any emergency attendance or admission, suggesting the virtual service can safely handle a substantial share of clinically complex cases that would otherwise trigger a hospital transfer.

One surprising finding may reshape how virtual services communicate with patients. Linked data revealed that a third of patients advised to attend a physical emergency department never showed up within 48 hours, a non-compliance rate remarkably consistent with Australian telephone triage services, where 34 percent of callers told to attend immediately also failed to comply within 24 hours. The authors speculate that perceptions of illness severity, trust in medical advice, transport barriers, symptom improvement, or preferences for community-based alternatives may all influence what patients do after the screen goes dark. They caution that presentations to private emergency departments, not captured in the linked datasets, may also account for part of the gap, and they flag this group as a priority for safety research.

The study’s limitations are acknowledged candidly. It captures a single year during the service’s implementation phase, excludes private emergency department use, and relies on administrative coding that may contain inaccuracies. Residual record linkage error, potentially non-random across population subgroups, remains possible. The cohort was also predominantly Australian-born and English-speaking, underrepresenting culturally diverse communities relative to Victoria’s broader population, where roughly 30 percent of residents were born overseas. The authors note that future evaluations should embed Aboriginal and Torres Strait Islander data governance, and that longer-term outcomes including mortality, patient decision-making, equity of access, and rigorous health economic analysis all remain to be established at scale.

Even so, the implications for health systems worldwide grappling with unsustainable emergency demand are hard to ignore. At a population scale exceeding 100,000 consultations, an avoidance rate above 80 percent represents a meaningful reduction in physical emergency attendances, achieved with unplanned re-presentation rates equal to or better than those following discharge from physical emergency departments and rival remote care models. The authors conclude that virtual emergency departments can serve as a scalable complement to traditional emergency services, provided policymakers design referral pathways carefully for higher-risk groups such as frail older adults in residential aged care. As emergency systems confront ageing populations, rising chronic disease, and chronic overcrowding, Victoria’s experiment suggests the emergency department of the future may increasingly begin not in a corridor, but on a screen.

Subject of Research: Population-level outcomes of a state-wide virtual emergency department in Victoria, Australia, using linked emergency and hospital administrative data

Article Title: Outcomes of a state-wide virtual emergency department in Victoria, Australia: a linked data analysis of emergency department utilisation and hospital use

Article References: Sher, L. M., Jessup, R. L., Boyd, J. H., Hutton, J., Suleiman, M., Lawrence, J., Miller, S., Nehme, Z., Bevan, D., & Semciw, A. I. (2026). Outcomes of a state-wide virtual emergency department in Victoria, Australia: a linked data analysis of emergency department utilisation and hospital use. The Lancet Regional Health – Western Pacific, 74, Article 101985. https://doi.org/10.1016/j.lanwpc.2026.101985

Image Credits: AI Generated

DOI: 10.1016/j.lanwpc.2026.101985

Keywords: virtual emergency department, telehealth, emergency medicine, data linkage, hospital utilisation, Victoria, digital health, health services research, ED avoidance, patient outcomes, Outcomes, state-wide