A 2-Year Retrospective Clinical Evaluation of a Novel Virtual Ward Model

Introduction

Globally we are experiencing demographic shifts toward older populations, ¹ along with a rise in chronic diseases^2,3 and multimorbidity.^{4 -6} Hospitals in the UK, like the US, have seen a rise in older, multimorbid patients often experiencing prolonged hospital stays.^{7 -9} These delays have a downstream effect on hospital bed occupancy and overcrowding in acute emergency care areas. National Health Service (NHS) virtual wards (VW), similar to US Hospital at Home (HaH), help expand capacity by allowing inpatients to be cared for outside the brick-and-mortar hospital.

The promise of VWs and HaH lies in their ability to deliver safe and effective healthcare for patients at home, reducing burden on brick-and-mortar institutions, supporting care provision closer to home, and maintaining or increasing patient satisfaction. VWs allow patients to receive hospital-grade care in the comfort of their own homes, ¹⁰ linking acute hospital care with community-based healthcare. VWs can integrate with primary care by supporting chronic disease management, enhancing care continuity, and reducing unnecessary hospitalizations.

The National Health Service (NHS) in the UK has extended funding for VWs for 2025 and 2026, following a successful evaluation of 22 000 VW admissions in South-East England.^11,12 In the US, the Acute Hospital at Home Care Waiver, created by the Centers for Medicare & Medicaid Services to cope with the COVID-19 pandemic, ¹³ has allowed VWs to expand, along with improved technology. ¹⁴ Continued funding, however, depends on further expansion, and evaluation of this care model both in the UK and overseas.

VW services described in the literature have typically focused on COVID-19, frailty, and respiratory care as primary use cases driving VWs forward.^{15 -17} These often used “step-down” pathways, moving patients in a single direction from an acute bed in a hospital to a VW. ¹⁸ There are limited publications describing VW care models that include step-up pathways in the literature. These tended to be small, focused on single use cases, but show positive results.^{19 -21} Step-up pathways could fundamentally change healthcare by diverting patients that would otherwise occupy a hospital bed, to receiving the care they need at home. For the right patient, this could mean a reduced risk of complications, ²² more movement, ²³ and increased comfort ²⁴ through monitoring or recovery.

Wrightington, Wigan, and Leigh Teaching Hospitals NHS Foundation Trust (WWL) developed a VW service that integrates with primary and community care to support patients across the care continuum with both “step-up” and “step-down” pathways and associated referral sources with very few exclusion criteria (Supplemental 1). Every admission was assessed on its merits based on clinical need and clinician assessment. Referral sources for “step-up” care included Advanced Care Practitioner, Emergency Department (ED) physician, General Practitioner (GP) and other community clinicians (cardiac or respiratory nurse, matron, React team GP). The Community React Team is a responsive service dealing with urgent issues, usually within 2 h, to reduce hospital attendance and admissions. The team can receive referrals from other services and provide interventions including intravenous antibiotics. ²⁵ The numbers 111 and 119 also provide medical help and advice for urgent but not life-threatening situations. The Community React Team, 111, and 119 will be referred to as “urgent medical help services” in this paper. Referral sources for “step-down” care included discharging consultants, ward staff, and VW Band 7 Coordinator on an acute site.

An example of a patient on a “step-up” pathway would be one referred directly into a VW by their GP, without ever having to leave their home. It would also be a patient admitted to the VW from the ED before they take up a hospital bed. In this paper, we refer to step-up and step-down pathways in specific ways, meaning to, and from the VW. The step-down pathway is when a patient is moved from an acute bed in a hospital to a VW. The step-up pathway is diverting a patient before they require a hospital bed.

Step-up patients came from their homes, communities, care facilities, or GP surgeries, unlike step-down patients that came to the VW from an acute bed in a hospital. Admission to VW included remote monitoring, individualized care plans, education around equipment and clinical team contact, and when to contact emergency medical services (ED) out of hours to promote autonomy and shared care responsibility (Supplemental 1).

Methods

Retrospective, single-center study of patients admitted to the WWL NHS Foundation Trust VW service from January 14, 2022 to January 31, 2024 and had not requested their data be removed.

Results

The WWL clinical dataset included 1835 admissions to the VW during the evaluation window; 50 patients requested their data be excluded. Of the 1835 WWL VW admissions, we had a 93% match rate to CH admissions data. A total of 134 admissions could not be matched to CH data, leaving 1701 admissions successfully matched to the CH database. There were 38% of referrals from step-up pathways (ambulatory care = 31%, primary care = 7%) and 62% from the step-down pathway (100% inpatients). Most specialty referrals were from thoracic and acute medicine (77%), and the remainder from cardiology, general surgery, endocrinology, gastroenterology, stroke, and geriatric medicine.

Admissions over time by referral pathway (step-up or step-down care) are shown in Figure 1 reflecting a growth in VW admissions, and particularly those for the step-up pathway. Average daily census by month is shown in Figure 2 reflecting growth in the average number of patients cared for each day on the VW. Of the 1701 admissions, the large majority were single admissions (n = 1524 or 90%). However, there were some repeat admissions: 155 patients (9%) had a second admission to the VW, 19 patients (1%) had a third admission to the VW, and 3 patients had 4 admissions.

OPEN IN VIEWER

Figure 1.

Number of admissions to the virtual ward per month by referral source.

OPEN IN VIEWER

Figure 2.

Average Daily Census on the virtual ward.

Admissions were 54% female. Most admissions were 65 years or older (64%) with an overall mean age of 67 (16 SD) years ranging from 16 to 98 years old (Figure 3). Six patients, aged 16 or 17 years, were admitted to the VW based on clinical need, though the overall inclusion criteria for the VW was age 18.

OPEN IN VIEWER

Figure 3.

Age at admission to the virtual ward.

Reported comorbidities for admissions were diabetes (20%), hypertension (34%), coronary heart disease (2%), chronic kidney disease (8%), and other long-term conditions (80%). Only 14% of admissions had no comorbid conditions, 43% had at least 1, and the remaining 43% had more than 1.

Clinical outcomes, adherence metrics, and alarms can be found reported in Table 1 by referral pathway (step-down or step-up) and for the total sample.

OPEN IN VIEWER

Table 1.

Virtual Ward Outcomes.

	Step-down	Step-up	Total
Clincial outcomes, adherence, and alarms	n = 1051	n = 650	n = 1701
VW length of stay in days	8 [IQR 6-13]	8 [IQR 5-13]	8 [IQR 5-13]
Estimated bed days saved	7 [IQR 7-7]	7 [IQR 7-7]	7 [IQR 7-7]
Escalations to the Emergency Department	112 (11%)	67 (11%)	179 (11%)
Escalations to the hospital	130 (12%)	79 (12%)	209 (12%)
Hospital bed days if admitted to the hospital	4 [IQR 1-9]	3 [IQR 1-8]	4 [IQR 1-8]
Signposted to urgent medical help services	22 (2%)	7 (1%)	29 (1.8%)
Adherence
Wearable	92 [IQR 84-96] %	92 [IQR 84-97] %	92 [IQR 84-96] %
Blood pressure	72 [IQR 56-85] %	72 [IQR 56-85] %	72 [IQR 56-85] %
Survey	60 [IQR 28-83] %	60 [IQR 25-83] %	60 [IQR 26-83] %
Total alarms/patient-day	2.6 [IQR 1.5-4.4]	2.9 [IQR 1.5-4.9]	2.7 [IQR 1.5-4.6]

There were 38 incidents during the 2-year evaluation period. These incidents were related to VW patients, systems, and processes. Some of the patient factors included hypersensitivity reactions to the wearable device resulting in minor irritation, laboratory issues with blood test results going to the GP rather than to the VW staff, and a patient using their own blood pressure cuff rather than the provided cuff that transmitted blood pressure readings to the clinician-facing dashboard, resulting in absent blood pressure readings. An example of a process issue was a discharge medication not being documented in a discharge letter. There was one out-of-hospital cardiac arrest and on full review of the case, no omissions of care were found with regards to the VW. Of 1835 admissions in the VW there were 17 expected deaths in the cohort (0.9%).

Like the overall VW sample, 91% to 92% of responses to patient experience surveys were from unique admissions. However, patients admitted more than once had the opportunity to submit survey data more than once; the remaining 8% to 9% were from repeaters to the VW. Response rates for surveys were between 35% and 56% of the total admissions sample (Table 2). Patients reported positive experiences in the VW via the 3 surveys (Table 2). The Ease of Use survey showed high levels of agreement that the kit was easy to use (Supplemental 2). About 77% of responses to the Friends and Family Test were between 8 and 10, indicating high satisfaction with the VW service provided by WWL.

OPEN IN VIEWER

Table 2.

Patient Experience.

Survey	Sample size (% of total sample)	Days from admission to taking survey Median [IQR] (range)	Score a
Ease of use	959 (56)	4 [IQR 4-4] days (3, 16)	5.7 ± 1.4
Patient experience	595 (35)	3 [IQR 3-3] days (3, 29)	94% rated as “excellent” or “good”
Friends and family test	933 (55)	4 [IQR 4-4] days (3, 15)	8.5 ± 2.3

Patient Experience score is the percentage of responses rated as “excellent” or “good” (eg, care received from the VW).

Friends and Family Test score is a mean value of responses from 0 to 10 with 10 being most likely to recommend the Current Health kit.

a

Ease of use score is a mean of the Likert-scale responses per patient from 1 = strongly disagree to 7 = strongly agree. A score of 7 indicates that patients strongly agree with statements about the Current Health kit being easy to use.

In the WWL Patient Experience survey, the VW experience (94%), the care received from the VW (98%), and the CH kit and wearable (95%) were rated as “excellent” or “good.” CH kits were rated as “easy” or “very easy” to use in 93% of survey responses, which aligned with findings from the Ease of Use survey results. The main benefits as reported by patients were “being cared for in their own home” and “how good it was to be in their own homes knowing they were safe” because they were being monitored. Many patients mentioned what comfort it was to be able to get good sleep, have control over their food and downtime, be free from hospital-acquired illness, and to know their care at home left a bed open for someone that needed it more than them.

Being with family, using my own bathroom, my food and getting my own drinks. Getting more rest and knowing there was someone in the background if I needed help, plus it’s freeing up a bed at the hospital.

Female, 72 years old

Home environment definatly [sic] aids recovery, less stressful in regards to in hospital having to continuously ask staff for everything which is not fair on staff as they are all so busy.

Male, 61 years old

Patients were also asked how their VW care could be improved and whether they found anything difficult. Many patients responded that there was nothing that needed to be improved. Some patients felt that it could be lonely, were worried about their safety if something happened when nobody was nearby, and they would have liked more help with the equipment.

Just anxiety over any symptoms and that you may become unwell on your own without help around.

Female, 39 years old

I find it ok, but I do also like being in the care of nurses or doctors in the ward too. When I was at the hospital the other day I found it comfortable and staff were very helpful. I didn’t get long periods of being left alone considering the workload they have.

Female, 67 years old

Discussion

Providing care at home through WWL’s novel VW model is safe and effective, even for complex patients that are high acuity and have comorbidities. WWL developed their VW within integrated care services using multidisciplinary community rapid response teams per the guidance of the British Geriatric Society. ³⁰ Furthermore, they created a VW with enhanced clinical scope by improving accessibility and inclusivity. They expanded and scaled the VW to more than the traditional use cases of respiratory and frailty patients. It meant that 38% of the VW patients that would otherwise have occupied a bed in a hospital were diverted to the VW and of those, 7% never had to leave their home.

Patient satisfaction scores suggest the technology was easy to use and the VW service was well liked by patients and aligned with person-centered care. Survey response rates were between 35% and 56%, and likely those patients that answered the surveys also engaged better with the VW service and technology. However, the very high adherence rates to the wearable device across this sample support a positive relationship between patients and the technology.

In our outcomes, we included metrics on alarming. Alarms are related to patient safety, alarm fatigue, and clinical workflow. If alarms are underutilized, early warning signs of patient deterioration could be missed. If alarms are overutilized, patients, caregivers, or HCPs may become desensitized. Alarming for VWs and HaH is different than alarming in a hospital setting. ³¹ By providing alarming data on this VW service we provide insight into clinical burden and potential alarm fatigue. ³² This could help clinicians manage expectations and staffing based on expected alarm rates.

In a meta-analysis and systematic review of RCTs providing step-down VW care, Chauhan and McAlister ³³ found that VWs were associated with fewer ED visits, lower health care costs, and shorter lengths of stay, compared to standard post-discharge care. The WWL VW had slightly higher median LOS at 8 days compared to 6.3 mean days in a national study of over 5000 patients of US hospital care at home programs. ³⁴ The WWL VW had escalations to the ED at 11%, higher than 6.2% in the US study but similar to a smaller study of a COVID-19 VW that reported 9%. ¹⁵ This shows us that the WWL VW appears comparable in its metrics to other VW-type services. Escalation to hospital care should not be considered a “failure” - more as evidence that the VW was accepting patients that were truly acute, rather than fit for discharge home.

The VW at WWL was safe for patients with only 17 expected deaths (1%) in the 2-year timeframe of this study. Of those 17 deaths, some were from families that asked for the reassurance and support of monitoring a family member during end-of-life care at home.

Average hospital death rates in the UK are 8.5%, though there is considerable variation. ³⁵ Other VWs report a mortality rate of 1% to 2%, ³⁶ in line with our findings. While it may be that fewer acutely ill patients are admitted to the VW than remain in the hospital, the matched case control study by Jalilian et al ³⁷ still found that WWL VW patients had 63% lower mortality rates than matched hospital inpatients. Levine et al ³⁴ in a national review of acute hospital at home care, found a mortality of 0.5% during care and a 30-day post-discharge mortality of 3.2%. A rapid evidence synthesis found that VW care had mortality rates as good or better than hospitalized patients ³⁸ while another systematic review found a reduced mortality rate for heart failure VW patients compared to hospitalized patients but no reduction in mortality for other diagnoses. ³³ While there have not been direct comparisons to hospitalized patients, Bates et al ³⁹ identified adverse events in 1 in 4 hospital admissions with almost a quarter of those being preventable.

Though the VW admits high acuity patients, it will not be ideal for every patient. Patient acuity level of the VW was not measured; therefore, it is unknown how similar the acuity level of the VW population is to patients cared for solely in hospital. The simplified and inclusive clinical pathways WWL used for their VW meant that admitting clinicians had to have the experience and autonomy to include more challenging patients. This could limit use by clinical staff with less experience. The WWL VW team also found it challenging to get clinician buy-in and promote VW awareness, especially for the step-up pathway. To secure clinician buy-in, the WWL VW team regularly communicated the uptake of the VW service from various specialty teams, creating a leaderboard. These successes were shared widely through a newsletter, executive blogs and hospital trust updates. Small group engagement sessions and one-to-one discussions were also utilized to secure uptake and engagement.

It took time and communication to improve clinical acceptability of the service and technology and promote its uptake among GPs. WWL started small, had a few engaged clinicians (ie, clinical champions), ⁴⁰ showed early results of safety and acceptability, targeted clinicians happy to adopt these VW models, and shared success metrics. They also improved the service model based on feedback from clinical teams and patients. The increased trend for uptake of the VW in general and of step-up care in particular suggests that these tactics were successful (Figure 1).

VWs, enabled by remote monitoring and incorporating step-up care pathways could help manage complex cases across multidisciplinary teams. In lower resource settings, simplified models of remote consultations with intermittent vital sign monitoring could benefit patients requiring prolonged care when a hospital is far or lacking capacity. They can bridge the gap between primary and secondary care, though success relies (in part) on effective integration into a shared electronic patient record. Further research would be useful to assess outcomes for step-up patients, and to identify opportunities for primary care specific pathways.

Researchers and clinicians could refine patient selection criteria and patient stratification. First, to identify patients most likely to benefit from VW services. Secondly, to ensure patients receive the appropriate level of care and monitoring based on their clinical acuity, possibly moving from continuous remote monitoring to intermittent monitoring as they recover.

We recommend that beyond the measures of clinical outcomes, adherence, safety, and patient satisfaction available for this evaluation, measures of cost, patient acuity, 30-day readmission, and the number of hospital bed days a patient had before being stepped down to VW care should be included in VW evaluations. To be meaningful, analyses should be powered and collected once the VW is established. ¹²

As the aging population grows, infrastructure is strained, and healthcare costs continue to soar, many health systems could benefit from step-up care pathways that keep patients safe and out of hospital beds. This evaluation contributes to the limited evidence base on VW outcomes that include step-up pathways. We have shown here it is safe, effective, and well-liked by patients.

Conclusion

In this application of an expanding VW service we have shown the service is safe and liked by patients, yielded good clinical outcomes, and had no negative impact on adverse events or mortality in this patient sample. The VW service was innovative in its acceptance of a broad range of patients and in developing the step-up pathway. We have shown how to successfully integrate a VW service with primary care and utilize an efficient step-up and step-down approach. This care model could reduce burden on hospitals by reducing length of stay and bypassing hospital admission altogether.

Supplemental Material

Supplemental Material