Disposition of patients utilising the virtual emergency department service in southeast region of Melbourne (SERVED-1)

Objective : Supported by the state gov-ernment, three health networks par-tnered to initiate a virtual ED (VED), as part of a broader roll-out of emergency telehealth services in Victoria. The aim of the present study (Southeast Region Virtual Emergency Department-1 [SERVED-1]) was to report the initial 5-month experience and included all patients assessed through the service over the ﬁ rst 5 months (1 February 2022 to 30 June 2022). Methods : VED consults occurred after referral from paramedics in the pre-hospital setting. Electronic medical records were retrospectively reviewed for demographic, presenting complaint and outcome data. The primary outcome was the count of VED consultations. The secondary outcome was the proportion of patients where physical ED attendance was avoided within 72 h. The proportion of physical ED attendances avoided sub-grouped by primary presenting complaints were reported. Results : There were 1748 patients who had a VED consultation, of which 1261 (72.1%; 95% con ﬁ - dence interval [CI] 70.0 – 74.2) patients had physical presentation to an ED avoided in the 72 h following the consult. There was a signi ﬁ cant increase in consultations over the 5-month period (incidence rate ratio 1.27; 95% CI 1.23 – 1.31, P < 0.001) that was consistent in the three health services. The most common presenting complaints were COVID-19 and shortness of breath, and physical presentation was avoided most often among younger patients and those with COVID-19. Conclusions : Initial experience dem-onstrated a signi ﬁ cant increase in adoption of the service and an overall avoidance of physical ED attendance by a majority of patients. These results support ongoing VED consultations, complemented by follow up and health economic evaluations.

Objective: Supported by the state government, three health networks partnered to initiate a virtual ED (VED), as part of a broader roll-out of emergency telehealth services in Victoria. The aim of the present study (Southeast Region Virtual Emergency Department-1 [SERVED-1]) was to report the initial 5-month experience and included all patients assessed through the service over the first 5 months (1 February 2022 to 30 June 2022). Methods: VED consults occurred after referral from paramedics in the pre-hospital setting. Electronic medical records were retrospectively reviewed for demographic, presenting complaint and outcome data. The primary outcome was the count of VED consultations. The secondary outcome was the proportion of patients where physical ED attendance was avoided within 72 h. The proportion of physical ED attendances avoided sub-grouped by primary presenting complaints were reported. Results: There were 1748 patients who had a VED consultation, of which 1261 (72.1%; 95% confidence interval [CI] 70.0-74.2) patients had physical presentation to an ED avoided in the 72 h following the consult. There was a significant increase in consultations over the 5-month period (incidence rate ratio 1.27; 95% CI 1.23-1.31, P < 0.001) that was consistent in the three health services. The most common presenting complaints were COVID-19 and shortness of breath, and physical presentation was avoided most often among younger patients and those with COVID-19. Conclusions: Initial experience demonstrated a significant increase in adoption of the service and an overall avoidance of physical ED attendance by a majority of patients. These results support ongoing VED consultations, complemented by follow up and health economic evaluations.

Introduction
A persistent increase in ED attendances above population growth 1 has contributed to overcrowded EDs, limiting capacity to provide optimal emergency care for consumers. The pandemic has exacerbated pressure with staff shortages, ward closures, disruption to community services and isolation and infection control procedures. 2,3 It is now recognised a substantial proportion of presentations could be safely and efficiently managed outside of EDs, 4,5 if appropriate alternatives were available to access care. This approach is supported by The Australian Commission for Safety and Quality in Healthcare charter of rights, which emphasises that consumers must get the right care, in the right place, at the right time. 6 During the pandemic, telehealth (specifically consultation through an audiovisual platform), was utilised by some EDs as a means to reduce patientprovider contact and thus reduce transmission risk. 7,8 Prior to this, telehealth had been utilised in a number of ways, including enabling remote consultation with inpatient teams and to facilitate the transfer of patients between EDs. In the prehospital setting, it has also been used to seek guidance from receiving hospitals regarding the initiation of treatment by paramedics. [9][10][11] Focusing on telehealth-enabled physician audiovisual consult pre-hospital, studies have centred on the care of patients requiring time critical interventions (such as stroke or major trauma). Relatively little has been published on the potential use of this technology to avoid patient transport to the ED. 12 These can include lower acuity presentations or patients who have the potential to be followed up and managed through communitybased services (such as 'residential inreach'). By providing alternative access to healthcare, telehealth can improve patient experience through provision of definitive care or linking with appropriate follow up. 13 In January 2022, the Southeast Melbourne Metropolitan Health Service Partnership inclusive of Alfred Health, Monash Health and Peninsula Health, supported by the Victorian Department of Health, collaborated with Ambulance Victoria to develop and implement a virtual ED (VED) model of care focussing on addressing ED demand. This was concurrent to other emergency telehealth services being developed elsewhere in the state. This was launched at a time when, because of COVID-19, there was substantial emergency care demand and diminished access to primary care with many of those who were vulnerable potentially being reluctant to seek care because of the associated infection risk. 14 The service was designed to support the appropriate referral of patients through paramedics in the community to facilitate a virtual consultation with a senior emergency doctor. This allowed patients to receive high-quality care in the location most appropriate to their needs. Definitive treatment was to be provided with the assistance of the paramedics or alternative care coordinated using community-based services. Secondary systemic benefits that were anticipated were a reduction in the number of presentations including ambulance presentations to the ED with resultant increased availability of ambulance paramedic crews to respond to emergencies.
The aim of the present study (Southeast Region Virtual Emergency Department-1 [SERVED-1]) was to report on the initial 5-month experience of the VED service. In particular, the trend in number of patients who utilised the service, demographics, presenting complaints and the proportion of patients in whom physical ED attendance was avoided by facilitating timely care through an alternative, more desirable avenue.

Setting and intervention
The health services involved encompass seven EDs serving the large population within Melbourne's southeast. 15 These EDs are embedded in hospitals that range from metropolitan major tertiary referral centres to urban district centres covering both suburban and semi-rural parts of Melbourne. Combined, these receive on average 342 ambulance presentations per day. Peninsula Health covers the southern most parts of Melbourne and receives over 90 000 patients per annum. Monash Health in the largest health network, covering the high growth areas of southeast Melbourne and receives approximately 200 000 cases per year. Alfred Heath is based in the inner southeast of Melbourne and receives about 110 000 ED presentations per year.
The VED commenced on 27 January 2022 with initial operating hours of 1200 to 2100, 7 days per week. This report includes all patients assessed through the VED between 1 February and 30 June 2022. One VED team covered each individual health network, each one being staffed by at least one trained VED clerical staff member and one emergency physician (or multidisciplinary team including general practitioners, ED and general physicians in the case of Peninsula Health). Within Alfred Health a care coordinator (senior nurse or allied health provider) was also present for consultations. The Alfred and Monash VED teams were not expected to undertake any other duties while on shift; however, the Peninsula VED was incorporated into the regular duties of the community care team.
Within Victoria, low-acuity calls to the ambulance service (Ambulance Victoria [AV]) are subject to secondary triage by a paramedic or nurse. Alternative means of service are arranged to address patient needs where possible (e.g. referral to locum doctor) and only those determined to require emergency treatment and transport are dispatched an ambulance. 16 The VED acted as an adjunct to this existing protocol with initial patient access to the VED occurring through referrals from paramedics on attendance at the scene if considered appropriate. Paramedics were required to complete an electronic form that confirmed that the patient met prespecified eligibility criteria for a VED consult. Once completed, patient encounters were registered virtually by a VED clerical member via a dedicated audiovisual platform. The VED clerical staff member ensured the patient/ ambulance crew had access to a secure and confidential consult that was conducted on the 'Healthdirect' telehealth platform. 17 Generally, the telehealth consults began with a handover from the AV crew of the clinical state, pertinent aspects of past medical history and any clinical examination (including vital signs and ECG) that may have already been performed. The aim was to develop a collaborative patientfocussed treatment and disposition plan. The Alfred VED care coordinators, who were present during consultations, could clarify aspects of the patient's care requirements (particularly social history) and assist in explaining referral and follow-up plans to the patient. Additional discussions were able to be had within a consultation, with inpatient teams, general practitioners or carers/family. Disposition options from VED were transport to hospital or to remain at the place of residence. To facilitate management on site, paramedics were able to administer therapies and prescriptions could be sent to local pharmacies. Follow up could be arranged through referrals to general practitioners, outpatient clinics, or through a number of hospital coordinated, community-based services (e.g. residential in-reach or COVID community follow up). 18 Should patients not be transported, paramedics would reinforce follow-up plans prior to departure. As a means of safety-netting, patients not transported had a telephone follow up by a care coordinator the following day. This was conducted to assess for any clinical deterioration, determine if there had been any problems with following through with the discharge plan and to reinforce advice about when to return to hospital.

Study design and analysis
This was a retrospective cohort study including all patients assessed through the service during the study period. Data were captured through review of the electronic medical record system within the respective health networks for all VED consultations between 1 February and 30 June 2022.
The number of VED consults (primary outcome), initial disposition advice and physical attendance to EDs at 72 h after the consultation were presented as counts and proportions. An avoided ED attendance was recorded when a patient did not attend the physical ED within the network in the 72 h following a VED consult.
Descriptive statistics were used to summarise the profiles of ED presentations and presenting complaints were summarised using the classification proposed by the Alfred Registry for Emergency Care Project. 19 Patients presenting with multiple presenting complaints had only what was considered to the single most significant primary complaint coded. Any patient diagnosed with COVID-19 presenting with a symptom potentially attributable to COVID-19 (e.g. shortness of breath) were allocated a presenting complaint of 'COVID-19'. Patients who presented with these symptoms who had not been diagnosed with COVID-19 (even if test was pending at the time of consultation) were allocated the primary complaint (i.e. shortness of breath) rather than 'COVID-19'. Incidence rates (95% confidence intervals [CIs]) of physical ED attendances avoided were calculated for each monthly period based on the total number of consults in each calendar month. A Poisson regression model was used to test for a dose-response effect of a change in ED avoided incidence per month over the 5 months by assuming a linear change in the logarithm of the rate with increasing months. The incidence rate ratio (IRR) and 95% CIs were calculated; a P-value <0.05 was considered statistically significant. All analyses were performed using Stata v 15.1, College Station, TX, USA.
The present study was approved by the Alfred Human Research Committee (Project ID: 79/22), with site specific authorisation granted by Monash and Peninsula Health.

Results
There were 1748 patients who had VED consultation and were included in the present study. The distribution of patients by hospital, disposition plan by VED and presentation to the ED within 72 h of consultation are displayed in Figure 1. There were 1261 (72.1%; 95% CI 70.0-74.2) patients who did not present to the ED after a VED consult (whether they were advised to or not). There were 1346 patients who, after a VED consult, were advised to avoid a presentation to an ED. Among these, 1225 (90.9%; 95% CI 89.2-92.3) did not physically present to an ED within 72 h of the consultation.
Demographics and presenting complaints of patients who were consulted by VED are presented in Table 1. Most patients were aged ≥65 years. There was equal distribution of consultations by day of the week. The most common presenting complaint categories were COVID-19 and shortness of breath. Most patients were living at home at the time of the consultation.
The monthly count of VED consultations is presented in Figure 2. There was an overall increase in the number of VED consultations per   1.42-1.68).
The proportion of VED presentations where physical ED presentations were avoided, sub-grouped by health service is displayed in Figure 3. There was no significant change in the proportion of ED presentations that were avoided during the 5-month study period.
The proportion of patients who had ED presentations avoided, subgrouped by demographics and presenting complaints are displayed in Figure 4. The most common variables associated with avoided ED presentation were young age and presenting complaint of COVID-19. The variables associated with lowest proportion of avoided ED visit were older age, presenting complaint of abdominal pain and reduced GCS. During the 7-day period after the VED consult, there were a further 16 patients who were assessed by the VED that presented to EDs.

Discussion
The VED service in southeast Melbourne resulted in avoidance of ED attendance for the majority of patients. Over the first few months of this service there was a significant increase in consultations. That was likely because of increasing awareness and acceptance of the service by paramedics.
Our service performed favourably in comparison to other services trialled elsewhere. One small study in a rural region in the Czech Republic achieved a non-transport rate of 17.7% utilising at scene mandatory video consults. 20 A larger study in Houston, again utilising audiovisual consults in the pre-hospital setting, was able to reduce ambulance transfers to ED to 18% following paramedic and VED consultation at scene. 21 However, in the case of the latter study, many patients who were not transported by ambulance were instead given a pre-paid taxi voucher for transport to EDs, translating to a 51% diversion rate. It is difficult to determine reasons for the difference in the diversion rate between our VED service and those detailed above. One possible reason, based on our experience is that VED consultation works best as a continuum of care, and the ability to supplement through access to other services such as in-reach services and community pathways allows more patients to be suitably managed outside the hospital. In addition, compared to some other VED models that which incorporate community self-referral, 22 a key aspect of our service was exclusively paramedic-facilitated consults for all patients. This enabled recording of vital signs and physical examination by a trained health practitioner.
The ability to facilitate care outside of hospital is dependent not only on the hospital units and ambulance services, but numerous other components of the healthcare system including general practitioners, outpatient clinics  and residential care homes, with uptake required by all relevant stakeholders. It is our expectation that as attitudes shift, we will see a greater adoption of the VED as part of community-based care and thus increased utilisation. Paired with this, further integration of virtual technologies into other parts of the healthcare system will likely see the VED forming an important component of the remote delivery of community care. Further work is required to assess the feasibility of the VED through health economic evaluation of ambulance time and admission costs saved against the cost of operating the service. An evaluation of the satisfaction of patients and healthcare providers utilising the service as well as a review of adverse outcomes will be essential. It should be noted that the virtual ED service partly functions by redirecting care, not always providing definitive management: ongoing assessment is required while expanding the service to ensure other components of the health system are not inappropriately burdened while diverting physical care away from the ED.
In this initial report of the service, the total number of ED presentations avoided was small, relative to the total number of ambulance presentations. It is expected that there will be an increase in VED consults, and the change in incidence of VED consults being a key performance indicator of the service. Even with the small numbers, the cumulative patient-times avoided in physical EDs were likely to be substantial. It is acknowledged that VED cannot substitute for a significant proportion of ED attendances, but appears to provide effective care to a selected population. The health economic analysis of this service will be the focus of future research.
In addition to the small sample size, it is worth noting that there were some other limitations of the present study. This pilot was undertaken during the pandemic and a significant proportion of patients assessed through the virtual ED were actively infected with COVID-19, many of whom were able to be safely managed at home through COVID community pathways. While the VED proved useful, this may not be a long term problem. Second, we were not able to obtain data from patients if they presented to a hospital outside of our own networks following virtual ED consultation. This is likely the explanation as to why fewer patients seemed to arrive at our ED than were advised to present to hospital. This is likely to be a small number but future evaluations will address this through phone follow up at 7 days post consultation.
We also assume that all patients would have otherwise been transported by the ambulance service. It is possible that a small proportion of patients reviewed by paramedics may have been advised to remain at home anyway should the VED service not have been available. 23

Conclusions
This initial report on the service suggests it is a useful intervention, avoiding a majority of patients from attending ED and potentially reducing ED and hospital overcrowding. The proportion of ED attendances avoided was consistent through the 5-month period. Avoidance rates were high among younger patients and those with COVID-19 and present priority target sub-groups for expansion of the service beyond paramedic facilitated consults.