Overview
This chapter draws on two papers by Vindrola-Padros et al.,32,36 published in Lancet eClinicalMedicine: https://doi.org/10.1016/j.eclinm.2021.100965 and https://doi.org/10.1016/j.eclinm.2021.100799. These are open access articles under the term of the Creative Commons Attribution 4.0 International License (CC BY 4.0), which permits use, distribution and reproduction in any medium, provided that the original work is properly cited (see https://creativecommons.org/licenses/by/4.0/).
What was already known
- Previous research has explored the use of remote home monitoring models for non-COVID-19 conditions.
- COVID-19 remote home monitoring services were implemented rapidly around the world, but implementation processes were rarely documented.
What this chapter adds
- The COVID-19 remote home monitoring services varied in relation to the health-care settings and mechanisms used for patient triage, monitoring and escalation.
- Good communication within clinical teams, culturally appropriate information for patients/carers and the combination of multiple approaches for patient monitoring (app and paper-based) were considered facilitators in implementation.
- We could not reach substantive conclusions regarding patient safety and the identification of early deterioration due to lack of standardised reporting and missing data.
Aims of the phase 1 evaluation
In July 2020, we began a rapid mixed-methods evaluation to identify the key characteristics of COVID-19 remote home monitoring models operating during the first wave of the COVID-19 pandemic, explore the experiences of staff implementing these models, understand the use of data for monitoring progress against outcomes, document variability in staffing and resource allocation, document patient numbers and impact, draw out lessons learned for the development of models for the 2020–21 winter period and identify areas for further research. To achieve these aims, phase 1 was divided in two main workstreams: (1) a rapid review of the literature, and (2) a rapid mixed-methods evaluation to document implementation and identify the lessons learned during wave 1 of the pandemic. Findings from both these phases have been published,32,36 so we provide a summary here.
A rapid systematic review of the evidence
Remote home monitoring models have been implemented in the United States, Australia, Canada, the Netherlands, Ireland, China and UK, with some variation in the frequency of patient monitoring, modality (a combination of telephone or video calls and use of applications or online portals), patient admission criteria, staffing used for patient monitoring and level of clinical oversight and use of pulse oximetry. However, there is limited published literature on the models of care developed to implement remote home monitoring across different health-care contexts during the COVID-19 pandemic, the experiences of staff implementing these models and patients receiving care, the use of data for monitoring progress, resources required, as well as the impact of these models on clinical, process and economic outcomes. The aim of this review was to analyse the implementation and impact of remote home monitoring models for confirmed or suspected COVID-19 patients, identifying their main components, processes of implementation, target patient populations, impact on outcomes, costs and lessons learned.
Methods
Design
We followed an established rapid review method.48 This method uses an adapted systematic review approach, with the aim to reduce time taken to carry out the review (e.g. using a large multidisciplinary team to review abstracts/full texts and extract data, having a percentage of excluded articles reviewed by a second reviewer, and using software for data extraction).48
We used the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement.75 The protocol was registered with PROSPERO (CRD42020202888, registered 6 August 2020).
Search strategy
We used a phased search approach.48 This involved gradually adding search terms based on the keywords used in the identified literature (see Vindrola-Padros et al.32 for search terms). We used the following databases: MEDLINE, CINAHL PLUS, EMBASE, TRIP, medRxiv and Web of Science (searches carried out by CV on 9 July 2020 and updated on 21 August 2020, 21 September 2020 and 5 February 2021). To identify additional publications, we manually screened reference lists of included articles. We removed duplicates.
Study selection, inclusion and exclusion criteria
One researcher (CVP) screened titles, with researchers cross-checking exclusions in abstract (KS) and full-text phases (KS/MS). Disagreements were discussed until consensus was reached. To be included, studies needed to: (1) focus on monitoring of confirmed or suspected patients with COVID-19, (2) focus on pre-hospital monitoring, monitoring after emergency department presentation and step-down wards for early discharge, (3) focus on monitoring at home (excluding monitoring done while the patient is in health-care facilities), and (4) be published in English.
Data extraction and management
We analysed articles using a data extraction form developed in REDCap. We extracted data on design, populations, process, clinical outcomes and economic impact. The form was independently piloted by two researchers (CV/KS) and finalised. Data extraction was checked by three researchers (TG/CSJ/ST).
Data synthesis
The information entered in free text boxes was exported from REDCap and analysed using framework analysis.76 We did not assess quality of studies.
Results
We present a summary of findings below (see Vindrola-Padros et al. for further information).32
Search results
The initial search yielded 902 articles. After screening, 155 articles remained for full-text review. After full-text review (n = 11), identifying additional references from reference lists (n = 3) and updated searches (n = 13), 27 articles met the inclusion criteria (see Report Supplementary Material 332 for PRISMA flow chart).
Characteristics of included remote home monitoring models
Remote home monitoring models were implemented in different countries, including the United States (n = 11), UK (n = 9), Canada (n = 2), Netherlands (n = 2), China (n = 1), Ireland (n = 1), Brazil (n = 1), Australia (n = 1). Different study designs were used: descriptive studies (n = 12), evaluations (n = 6), observational studies (n = 7), feasibility study (n = 1), news feature (n = 1). See Vindrola-Padros et al.32 for further details.
Aims and designs of remote home monitoring models
The primary aim of remote home monitoring models was to enable early identification of deterioration for patients self-managing COVID-19 symptoms at home (including those who had not been admitted to hospital as well as those who had been discharged).
A secondary aim of the models was to reduce the rate of hospital infection and demand for beds in the acute care sector (including preventing admitting those suitable for home management and discharging those admitted to hospital earlier with care from a remote team).
Most of the remote home monitoring models were led by teams in secondary care (n = 23), with some led by primary care (n = 3) and some led by both primary and secondary care (n = 2). Thirteen of the models functioned as pre-admission wards (preventing admission of patients to hospital) and five were designed for patients admitted to hospital who could be discharged and monitored at home (step down wards), and 10 were both (with separate pathways).
Patient populations appropriate for remote monitoring
Most of the models established broad criteria for patient eligibility, defining the patient group as adult (over 18 years) patients with COVID-19 symptoms (suspected and confirmed cases). Some models limited referrals to confirmed cases. Some models excluded certain groups (e.g. patients over 65 years with significant comorbidities, or pregnant women). The size of patient cohorts varied considerably (ranging from 12 patients to 6853; see Vindrola-Padros et al.32 for further details).
Stages of remote home monitoring
The articles described five main stages in remote home monitoring for COVID-19: (1) referral and triage to determine eligibility (including recording patient demographics, clinical variables, health data for risk assessment and vital signs data, and in some cases a risk assessment); (2) onboarding of patient to remote home monitoring service (provision of information to patient and/or carer on monitoring process, mechanisms for escalation and self-care); (3) monitoring [including recording of observations, communication of the information via a paper-based system and telephone or digital technologies such as apps or online forms, inclusion of pulse oximetry (in some cases), assessment of the information by the medical team]; (4) escalation (if required according to pre-established thresholds); and (5) discharge from the pathway (once their symptoms improved or patient opted out of the pathway).
Expected outcomes
Outcomes of remote home monitoring models were grouped in three main categories: (1) process outcomes related to the remote home monitoring pathway (e.g. time from swab to assessment, time to escalation, ambulance attendance/emergency activation), (2) process outcomes related to secondary care (e.g. LOS) and (3) patient outcomes, including clinical outcomes such as emergency department attendance/reattendance, hospital admission, ICU admission, readmission, mortality, ventilation needs and experience (e.g. patient satisfaction).
Impact on outcomes
It was difficult to carry out an analysis of the impact of remote home monitoring across all examples because not all articles reported data on the same outcomes. In terms of clinical outcomes, mortality rates were low, admission or readmission rates ranged from 0 to 29%, and emergency department attendance or reattendance ranged from 4% to 36%. Six of the models reported data on patient feedback and indicated high satisfaction rates.
Information about remote home monitoring process outcomes was limited (n = 6 articles), but time, from swab to assessment ranged from 2 to 3.7 days and virtual LOS from 3.5 days to 13 days (see Vindrola-Padros et al.).32
Economic impact
Very few of the selected studies for this rapid review provided a descriptive form of economic analysis. The amount spent per patient on remote monitoring varied by country and type of costs included in the analysis.
Strengths and limitations of review
The last search was carried on 5 February 2021, so any articles published after this date were not included. However, the inclusion of preprints helped to mitigate this issue. Although we employed multiple broad search terms, it is possible that we missed articles that did not use these terms. Owing to the variability in study designs and the descriptive nature of the articles we did not assess these articles for quality using standardised tools for assessment. However, we feel it is important to note that we found several cases of missing data and inconsistencies in the reporting of evaluations that would lead to low quality ratings.
Future research areas identified
The review highlighted that further research should focus on exploring patient experience, implementation, outcome evaluation and cost-effectiveness.
A rapid mixed-methods evaluation of the early implementation of remote home monitoring services
A number of remote home monitoring models were set-up during the first wave of the COVID-19 pandemic in England with a high degree of variability in relation to the mechanisms implemented for patient assessment and triage, monitoring and escalation.
Despite previous research on the use of remote home monitoring models for other conditions and their widespread use during the COVID-19 pandemic, studies on their implementation and impact are rare. This study aimed to identify the key characteristics of these services, explore staff experiences, understand the use of data for monitoring progress against outcomes, and document variability in staffing/resource allocation.
Methods
Design
A multisite mixed-methods study that combined qualitative and quantitative approaches to analyse the implementation and impact of remote home monitoring models during the first wave of the COVID-19 pandemic (July to September 2020) in England (see Vindrola-Padros et al.).36
Sample and recruitment
We recruited staff with experience delivering these models across eight pilot sites in England. We purposively recruited sites to represent a range of characteristics including the type of remote home monitoring service, different population sizes, areas of the country and monitoring approaches (see Vindrola-Padros et al.).36
Data collection
We conducted interviews with 22 staff. Most of these interviews were carried out by pairs of researchers. Interviews were conducted using a topic guide, were audio-recorded and lasted 30–60 minutes. We also collected data on staffing models and resource allocation (number of staff, their function/seniority, the number of hours worked, all other resources used in setting up and running pilots and associated costs).
Data analysis
Detailed notes from interviews were added to RAP sheets.70 Multiple researchers cross-checked data to ensure consistency. The team met weekly to discuss emerging findings and develop themes. Set up costs were reported as mean cost per site, and for running costs, mean costs of patient were calculated for the pilot period (see Vindrola-Padros et al.).36
Findings
We present a summary of findings below (see Vindrola-Padros et al.36).
Models implemented
The models varied in relation to the health-care settings and mechanisms used for patient triage, monitoring and escalation. The main components of models were triage, admission to service (including provision of pulse oximeter), patient provides readings through phone calls or app, medical team monitor observations and escalate cases, and discharge at 14 days or when symptoms have improved.
Barriers and facilitators to implementation
Staff described high levels of patient engagement. Implementation was embedded in existing staff workloads and budgets. Implementation facilitators included: good communication within clinical teams, culturally appropriate information for patients/carers and the combination of multiple approaches for patient monitoring (app and paper-based).
Barriers to implementation included: unclear referral criteria and processes, difficulties carrying out non-verbal assessments over telephone, lack of culturally appropriate patient information, lack of administrative support and workforce availability.
Cost
The mean cost per monitored patient varied from £400 to £553, depending on the model. Mean costs were higher in pre-hospital than early discharge from hospital model.
Strengths and limitations
The study was limited in the sense that we could not determine the effectiveness of these models due to lack of comparator data.
Future research areas identified
We identified the sustainability of the models and patient experience (considering the extent to which some of the models exacerbate existing inequalities in access to care) as future areas of research.
Dissemination and use of findings
The findings from the rapid review and the rapid mixed-methods evaluation on the implementation of COVID-19 remote home monitoring models across eight sites in England were disseminated widely and used to inform decisions in relation to the roll-out of services during autumn/winter 2020–21. The main channels for the dissemination of findings included journal articles,13,32,36 blogs,77–79 a slide set80 and presentations (see Project website documentation – Dissemination).
Impact
A senior director at NHSEI provided the following feedback on how findings from phase 1 of our evaluation influenced the national roll-out of COVID-19 remote home monitoring services in England:
- Setting very simple entry criteria [65 and over OR clinically extremely vulnerable (CEV)].
- Ensured that we did not go down a tech-only or tech-first route.
- Produced translations of all the patient facing guidance and an easy-read version at the outset. It is possible that we may have missed the importance of this point.
- National SOP was focused on patient self-escalation and our workforce calculations were informed by this work.
- Created a minimum standard for out-of-hours to ensure that the links were clear.
- We required services to be established within two months and had confidence that we could encourage all CCGs to set up by 31 December.
- Ensured that we were able to get pulse oximeters out to CCGs within three working days.
- Provide CCGs with a feed of patients that have tested positive and identify people within the cohort (65 years and over or CEV).
Informing phase 2 of the evaluation
The findings from phase 1 of the evaluation were also used to inform the design of phase 2, including the selection of outcome measures, cost analysis and identification of key areas of focus such as: the sustainability of these models during the second and third waves of the pandemic, patient and carer experiences with the services and how these varied according to the type of model (including analogue vs. digital models), factors that could have created or exacerbated inequalities in health and access to care and the costs involved in delivering these services.