key: cord-0791064-nbotxv35
authors: Vindrola-Padros, C.; Singh, K.; Sidhu, M. S.; Georghiou, T.; Sherlaw-Johnson, C.; Tomini, S. M.; Inada-Kim, M.; Kirkham, K.; Streetly, A.; Fulop, N. J.
title: Remote home monitoring (virtual wards) during the COVID-19 pandemic: a living systematic review
date: 2020-10-09
journal: nan
DOI: 10.1101/2020.10.07.20208587
sha: ceffa3008dfc0fb5002ae81cc72c832499f65f5c
doc_id: 791064
cord_uid: nbotxv35

Objectives: The aim of this review was to analyse the implementation and impact of remote home monitoring models (virtual wards) during COVID-19, identifying their main components, processes of implementation, target patient populations, impact on outcomes, costs and lessons learnt. The review will be kept live through regular updates. Design: The review was designed as a living systematic review to capture a rapidly evolving evidence base. We used the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement. Setting: The review included remote home monitoring models led by primary and secondary care across seven countries. Participants: 17 examples of remote home monitoring models were included in the review. Main outcome measures: Impact of remote home monitoring on virtual length of stay, escalation, Emergency department attendance, reattendance, admission, readmission and mortality. Results: The primary aim of the remote home monitoring models was the early identification of deterioration for patients self-managing COVID-19 symptoms at home. Most models were led by secondary care. Broad criteria for the eligible patient population were used and confirmation of COVID-19 was not required (in most cases). Monitoring was carried via online platforms, paper-based systems with telephone calls or (less frequently) through wearable sensors. We could not reach conclusions regarding patient safety and the identification of early deterioration due to lack of standardised reporting across articles and missing data. None of the articles reported any form of economic analysis, beyond how the resources were used. Conclusions: The review pointed to variability in the implementation of the models, in relation to healthcare sector, monitoring approach and selected outcome measures. Lack of standardisation on reporting prevented conclusions on the impact of remote home monitoring on patient safety or early escalation during COVID-19. Future research should focus on staff and patient experiences of care and potential inequalities in access to these models. Attention needs to be paid to the processes used to implement these models, the evaluation of their impact on patient outcomes through the use of comparators, the use of risk-stratification tools, and cost-effectiveness of the models and their sustainability. Protocol registration: The review protocol was published on PROSPERO (CRD: 42020202888).

COVID-19 has rapidly spread across the world, leading to high rates of mortality and unprecedented pressure on healthcare systems. Delays in the presentation of patients with COVID-19 has led to patients arriving as emergencies with very low oxygen saturations, often without accompanying breathlessness ('silent hypoxia') 1 . These delayed presentations of severe COVID-19 led to extended hospital admissions for patients, often requiring invasive treatment and potential admission to intensive care units (ICU) or death 2 . Remote home monitoring models (sometimes referred to as 'virtual wards') have been established to: 1) avoid unnecessary hospital admissions (appropriate care at the appropriate place), and 2) escalate cases of deterioration at an earlier stage to avoid invasive ventilation and ICU admission 3 . Some of these models have integrated the use of pulse oximetry to monitor oxygen levels and identify and treat cases of 'silent hypoxia' 2 .

Remote home monitoring models have been implemented in the US, 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 models used for patient monitoring and level of clinical oversight, and use of pulse oximetry [4] [5] [6] [7] [8] .

There is a paucity of published literature on the models of care developed to implement remote home monitoring across different healthcare 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 address these gaps by analysing the remote home monitoring models implemented during COVID-19, their main components, processes of implementation, target patient populations and lessons learned. We sought to identify evaluations of these models and their outcomes. Due to the rapidly evolving evidence base on the use of remote home monitoring models during COVID-19, we have designed this review as a living systematic review.

We followed the review method proposed by Tricco et al. 9 . The rapid review method follows a systematic review approach but proposes adaptations to some of the steps to reduce the amount of time required to carry out the review. We used a large multidisciplinary team to review abstracts and full texts, and extract data; in lieu of dual screening and selection, a percentage of excluded articles was reviewed by a second reviewer, and software was used for data extraction and synthesis 9 .

We used the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement 10 to guide the reporting of the methods and findings. The review protocol was registered with PROSPERO (CRD: 42020202888, registered 6 August 2020). This manuscript is the original version of the review and it will be updated on a quarterly basis over a period of two years.

The review sought to answer the following questions:

1. What are the aims and designs of remote home monitoring models? 2. What are the main stages involved in remote home monitoring for COVID- 19? 3 . What are the patient populations considered appropriate for remote monitoring? 4. How is patient deterioration determined and flagged? 5. What are the expected outcomes of implementing remote home monitoring?

. CC-BY 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 9, 2020. . https://doi.org/10.1101/2020. 10 .07.20208587 doi: medRxiv preprint 6 . What is their impact on outcomes and costs? 7. What are the benefits and limitations of implementing these models?

We used a phased search approach 9 . We carried out a series of search phases where we gradually added search terms based on the keywords used in the literature we identified. Appendix 1 includes the strategies used for these waves, including the final search strategy. We searched for literature indexed in the following databases: MEDLINE, CINAHL PLUS, EMBASE, TRIP, medRxiv and Web of Science. Initial searches were carried out on 9 July 2020 and updated on 21 August 2020 and 21 September 2020. Results were combined into Mendeley and duplicates were removed. The reference lists of included articles were manually screened to identify additional relevant publications.

One researcher (CVP) screened the articles in the title phase, and additional researchers (KS) cross-checked exclusions in the abstract and full-text phases (KS, MS). Disagreements were discussed until consensus was reached. The inclusion criteria used for study selection was: 1) focus on the monitoring of confirmed or suspected patients with COVID-19), 2) focus on pre-hospital monitoring, monitoring after Emergency Department (ED) presentation and step-down wards for early discharge, 3) focus on monitoring at home (excluding monitoring done while the patient is in healthcare facilities), and 4) published in English. Due to the rapidly expanding evidence-base on COVID-19, we included a wide range of publications (i.e. feature articles, descriptions of services) and did not limit the selection to evaluations of remote home monitoring.

The included articles were analysed using a data extraction form developed in REDCap (Research Electronic Data Capture) that extracted data on: the design and general characteristics of the model, patient populations, main reported process and clinical outcomes and its potential economic impact. The form was developed after the initial screening of full-text articles. It was then piloted independently by two researchers using a random sample of five articles (CV and KS). Disagreements were discussed until consensus was reached. The data extraction form was finalised based on the findings from the pilot. Data extraction was cross-checked by three researchers (TG, CSJ and ST).

Data were exported from REDCap and the main article characteristics were synthesised. The information entered in free text boxes was exported from REDCap and analysed using framework analysis 11 . The initial categories for the framework were informed by our research questions but we were also sensitive to topics emerging from the data.

Due to the descriptive nature of the articles and lack of data in relation to study design, we did not assess the quality of the studies.

The initial search yielded 902 articles ( Figure 1 ). These were screened based on the title and abstract and type of article, resulting in 155 articles for full-text review. Full-text review of these articles led to 11 articles that met the inclusion criteria (reasons for exclusion can be found in Figure 1 ). Three additional articles were identified by reviewing the bibliography and two articles were identified in an updated search carried out on 21 September 2020, . CC-BY 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 9, 2020. . https://doi.org/10.1101/2020.10.07.20208587 doi: medRxiv preprint ultimately leading to 16 articles (including 17 examples of remote home monitoring) included in the review. We excluded articles that focused on monitoring that took place within hospital settings (i.e. ICU) or for other non-COVID-19 related conditions.

Eight of the remote home monitoring models were implemented in the US, three in the UK, two in Canada, and one each in The Netherlands, China, Ireland and Australia. Seven of the articles described the service, two were identified as evaluations, five as observational studies, one as a feasibility study and one (containing the example of two models) was a news feature (with a limited description of the services). Twelve of the examples were published in peer-reviewed journals and five were published in the form of preprints. The main characteristics of the included remote home monitoring examples are summarised in Table 1 .

The primary aim of the remote home monitoring models was to enable the 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). The programme theory guiding these models was that if patients were able to take the required regular observations whilst remaining at home and communicate these to the healthcare professionals responsible for their care, then cases of deterioration could be identified early and acted upon. These actions could include changing their treatment protocol, referring them to primary care or to the emergency department for assessment and potential admission to hospital. A secondary aim of the models was their use to reduce the rate of hospital infection and demand for beds in the acute care sector, where admission to . CC-BY 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 9, 2020. . https://doi.org/10.1101/2020.10.07.20208587 doi: medRxiv preprint hospital could be prevented for patients considered suitable to be managed at home and those who had been admitted to hospital could be discharged earlier but continue under the remote care of a medical team (a team that varied in composition depending on the model).

Most of the remote home monitoring models included in the review (15 examples out of 17) were led by teams in secondary care. Two examples were primary care led. Nine of the models functioned as pre-admission wards, in the sense that they sought to prevent the admission of patients to hospital or to identify cases of deterioration early (so those who should be referred could be admitted to hospital with lower rates of acuity). Three of the models functioned as "step-down" wards, that is, they were designed for patients who had been admitted to hospital (including ICU) where the medical team had identified that they could be discharged and safely monitored at home until their symptoms improved. Five models functioned as pre-admission and step-down wards, organised according to two separate pathways.

Most of the models had broad criteria for the eligible patient population, defining the patient group as adult (over 18 yrs.) patients with COVID-19 symptoms (suspected and confirmed cases). Five of the models limited referrals to COVID-19 cases confirmed through testing 6 12-15 . The model described by Hutchings et al. 6 excluded patients over 65 years with significant comorbidities. Shah et al. 16 excluded pregnant women and only included patients with SpO2 above 92% at initial assessment. We did not find any examples targeting socially and economically disadvantaged groups (although some models included support from social workers and mental health professionals) 4 17 . It is important to highlight that the size of the patient cohorts varied considerably (see Table 2 for patient numbers) and ranged from 18 patients to 6853. The models with the highest numbers of patients were implemented in the US. The comorbidities mentioned with greater frequency were hypertension, asthma and obesity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 9, 2020. . https://doi.org/10.1101/2020. 10 

The articles described five main stages in remote home monitoring for COVID-19: 1) referral and triage to determine eligibility, 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, assessment of the information by the medical team), 4) escalation (if required), and 5) discharge from the pathway.

Patient information recorded at triage included:

• Patient demographics (age, sex, race/ethnicity, insurance type in the models in the US) • Clinical variables (clinical signs and symptoms, medical history and medications) • Health data for risk assessment and vital signs data (body temperature, heart rate, respiratory rate and oxygen saturation)

Three studies included some degree of detail in relation to the categorisation of patients in relation to risk 14 17 18 . The article by O'Keefe and colleagues 12 described and evaluated a risk assessment model based on age, medical history and symptom severity. This model was able to identify the need for hospitalisation in initially non-severe COVID-19 patients.

In six of the examples included in the review, monitoring was based on patient record of observations using a paper-based system and then communicating the information to a member of the medical team by telephone (see Table 1 ). Nine of the examples relied on the use of an online mechanism, either through an app or online form. One example offered patients a telephone or an app option 19 . Another example relied on the use of wearable sensors to continuously monitor temperature readings and transfer these to the medical team 6 . Ten of the models relied on the use of pulse oximetry from the beginning of implementation, four models did not use pulse oximetry, one model added pulse oximetry three weeks after implementation and two articles indicated that the use of pulse oximetry was being considered in the near future.

Escalation was actioned depending on pre-established thresholds. Not all articles have reported the thresholds for escalation and most only refer to the worsening of symptoms. Shah et al. 16 indicated that patients on their remote home monitoring pathway were flagged as deteriorating if reporting SpO2 below 92% after a double reading. Xu et al. 20 used a SpO2 reading of below 93% or BP less than 90/60 mmHg. Some of the examples included in the review established safety-netting options in cases when patients could not be reached via phone such as calling the police so they could visit the patient at home 6 .

Most patients were followed-up until their symptoms improved or the patient opted out of the pathway. Medina et al. 13 reported following up patients on the step-down pathway for 7 days post-discharge from hospital and those on the pre-admission pathway for 14 days. Shah et al. 16 followed-up patients on their pre-admission pathway for 7 days. Hutchings et al. 6 . CC-BY 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 9, 2020. . https://doi.org/10.1101/2020.10.07.20208587 doi: medRxiv preprint referred patients to their GP for follow-up after discharging them from the remote home monitoring pathway.

The outcomes recorded in each remote home monitoring model are listed in Table 1 . They can be grouped in three main categories: 1) process outcomes related to the remote home monitoring pathway, 2) process outcomes related to secondary care and 3) patient outcomes (including clinical and experience). Process outcomes related to the remote home monitoring pathway included: time from swab to assessment, time to escalation and ambulance attendance/emergency activation (i.e. calling 999 or 911). Process outcomes related to secondary care included length of stay. Outcomes considered at the patient level included: emergency department attendance/reattendance, hospital admission, ICU admission, readmission, mortality, ventilation or non-invasive ventilation needs, and patient satisfaction.

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 (Table 3) . Mortality rates were low, admission or readmission rates ranged from 0 to 29%, and ED attendance or reattendance ranged from 4 to 36%. Four of the models reported data on patient feedback, with high satisfaction rates 5 8 18 21 .

Remote home monitoring process outcomes were only included in six of the articles, with time from swab to assessment ranging from 2 to 3.7 days 12 17 19 and virtual length of stay from 3.5 days to 13 days (see Table 3 ). Only one article presented findings on reduction in length of stay, calculated at 5 days fewer per patient 21 . is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 9, 2020. 

None of the selected studies for this rapid review provided any form of economic analysis, though some of them mentioned the potential for cost savings based on the utilisation of virtual monitoring programs for other treatments in similar settings 14 21 . Some of the selected studies highlighted the fact that, during the pandemic, the intervention used existing resources and staff that were made available due to the emergency situation 7 12 14 22 . However, they also highlighted that, with the return to normal workloads in the health care system, a question of allocation of resources and sufficient staffing still remains.

In this article we have sought to make a contribution to the rapidly growing evidence-base on the use of remote home monitoring models for patients with confirmed or suspected COVID-19. The review has pointed to factors that need to be taken into consideration in relation to the design of these models. Most of the models included in the study were led by secondary care but some authors argued that coordination between primary and secondary care could facilitate the implementation of remote home monitoring pathways 5 7 13 . Primary care led models might be more adaptable to evolving patient and system needs and easier to replicate in contexts with limited secondary care access and capacity 17 . Three models integrated mental health and social care support during and after patient monitoring, highlighting a wide range of patient needs 6 13 17 .

Even though several of the examples used apps and other types of online platforms, discussions in relation to the use of health technology were limited. The use of apps for monitoring allowed the follow-up of a higher number of patients (compared to paper-based models) but some of the studies indicated that models based on telephone calls were more inclusive (i.e. including patients without internet access or technological literacy) 19 . Patient experience was captured in some of the examples we reviewed 8 21 but the analysis of patient experience was limited. An analysis of patient experience and engagement is important as the literature on the use of remote patient monitoring for other conditions has demonstrated that higher levels of patient engagement with remote patient monitoring technology are associated with better patient outcomes 23 .

Similarly to other reviews on remote patient monitoring in other conditions, another limitation was the lack of attention placed on the implementation of the models and the failure to identify the programme theories guiding their design, factors that acted as barriers and facilitators and the extent to which the pathways were implemented according to their original plans 24 . This could be due to the limited evidence on COVID-19 and the management of patients with this disease at the time of designing and implementing these models as well as the general limited use of programme theories in the design of healthcare interventions that has already been documented in the literature 25 . is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 9, 2020. . https://doi.org/10.1101/2020.10.07.20208587 doi: medRxiv preprint

Emerging international evidence has indicated that lower thresholds for oxygen saturation, are associated with worse patient outcomes 2 26 . In the case of our review, even though some authors argued that pulse oximetry identified the need for hospitalisation when using a cutoff of 92% 16 , we could not reach conclusions in relation to patient safety and the degree to which remote home monitoring models can conclusively identify cases of deterioration at an earlier stage in the disease trajectory. The main reasons were lack of standardised reporting across articles in relation to these outcome measures and how these were measured, as well as the limitation that none of the articles used comparators.

Issues with using pulse oximetry were also highlighted such as: patient physiological measures needed to be recorded several times a day to correctly identify cases of deterioration, some remote home monitoring examples used standardised home pulse oximeters to avoid variability between different brands, pulse oximetry readings were made less accurate by nail polish, severe anaemia, hyperbilirubinemia, hemoglobinopathies, or poor peripheral perfusion from severe vasoconstriction or poor cardiac output 16 27 . Some authors also argued that patient training was a key determining factor of the success of health information technology as it ensured readings and other observations were carried out accurately 6 . Remote home monitoring needed to be seen as an approach to maintain patients safely in the right setting rather than as an admission avoidance model.

Remote home monitoring for COVID-19 patients was expected to have a positive economic impact, mainly due to costs savings in staff time and PPE utilisation, avoidance of infection of frontline medical staff and reduced hospitalisations 14 21 . However, the economic evidence in relation to these was limited. None of the selected studies included any form of economic analysis. The selected studies have, however, raised the issue of resource allocation and funding, especially when it comes to the continuity of such programs after the first emergency situation. Most of the staff who worked on remote monitoring interventions for COVID-19 came from other services and the resources used were already existing. Yet, with the return to normal workloads, providing sufficient staff and enough resources may become a problem. Previous studies have indicated that remote monitoring in itself has contributed to increased efficiency in the use of resources (such as reduction in length of stay, increasing bed availability without compromising patient care safety, etc.) 15 21 . A complete economic analysis in this context could indicate if remote home monitoring for COVID-19 patients is a cost-effective intervention and could help inform accurate planning of the needed resources and staff. This economic analysis would also need to include costs and benefits beyond the actual remote home monitoring models, a reliable control group, as well as a longer follow up period. This review has a series of limitations. The last search was carried on 21 September 2020, so any articles published after this date were not included. We have included preprints as a way to address delays produced by external review and publication. Furthermore, although we employed multiple broad search terms, it is possible that we missed articles that did not use these terms. Due to the variability in study designs and the descriptive nature of the articles we did not assess these 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.

The review pointed to several future areas of research. These could include an analysis of patient experience, beyond measures of satisfaction and the exploration of potential inequalities in patients' access to remote home monitoring models or patients' difficulties interacting with technology. Technological barriers have been reported in other studies of remote home monitoring and should not be overlooked when exploring the experiences of patients with COVID-19 28 29 . Additional attention needs to be paid to the processes used to implement these models and how these might vary based on the healthcare sector, patient population, size, wave of the pandemic and approaches used for triage, monitoring and . CC-BY 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 9, 2020. . https://doi.org/10.1101/2020.10.07.20208587 doi: medRxiv preprint escalation. As mentioned earlier, primary care might need to play a more central role in the coordination of remote patient monitoring models, providing more holistic care for patients and reducing the demand on hospital services 30 . The evaluation of remote home monitoring, considering its impact on patient outcomes through the use of comparators is also required. We also need to consider the sustainability of these models during multiple epidemiological peaks, compare different approaches to remote home monitoring and assess their cost-effectiveness.

. CC-BY 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 9, 2020. 

The findings from this review will be disseminated widely, including to patient organisations.

All of the relevant data are included in the manuscript and supplementary files. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted October 9, 2020. . Pulse oximeters were connected to an app. The app sent a prompt to patients to record oxygen saturations 4 times a day.

The app triggered an alert and the medical team contacted the patient and gave instructions on next steps.

Readmissions, length of stay, ICU admission *Used a pulse oximeter **Did not use pulse oximetry in the main model described in the article but flagged the launching of a companion programme or incorporating pulse oximetry and escalation based on oxygen saturation at a later date. ***Pulse oximetry added three weeks after implementation SC=secondary care PC=primary care

Pulse Oximetry as a Biomarker for Early Identification and Hospitalization of COVID-19 Pneumonia

Oxygen and mortality in COVID-19 pneumonia: a comparative analysis of supplemental oxygen policies and health outcomes across 26 countries

Pulse oximetry to detect early deterioration of patients with COVID-19 in primary and community care settings

On the Front (Phone) Lines: Results of a COVID-19 Hotline in Northeast Ohio

The "virtual wards" supporting patients with covid-19 in the community

Virtual health care for community management of patients with

Rapid Implementation of an Outpatient Covid-19 Monitoring

Rapid implementation of a COVID-19 remote patient monitoring program

Rapid reviews to strengthen health policy and systems: A Practical Guide

Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement

Using the framework method for the analysis of qualitative data in multi-disciplinary health research

Initial Experience in Predicting the Risk of Hospitalization of 496 Outpatients with COVID-19 Using a Telemedicine Risk Assessment Tool

Home monitoring for COVID-19

Leveraging Health System Telehealth and Informatics Infrastructure to Create a Continuum of Services for COVID-19

Remote monitoring of oxygen saturation in individuals with COVID-19 pneumonia

Novel Use of Home Pulse Oximetry Monitoring in COVID-19 Patients Discharged From the Emergency Department Identifies Need for Hospitalization

Lessons from a Family Medicine Led Remote Monitoring

The Development and Implementation of a Virtual Discharge Ward for Patients with COVID-19 Pneumonia: Data on The First 300 Patients

A virtual care program for outpatients diagnosed with COVID-19: a feasibility study

Monitoring and Management of Home-Quarantined Patients With COVID-19 Using a WeChat-Based Telemedicine System: Retrospective Cohort Study

Home telemonitoring makes early hospital discharge of COVID-19 patients possible

Remote monitoring of patients with COVID-19: Design, implementation and outcomes of the first 3,000 patients in COVID Watch

Diabetes Management Through Remote Patient Monitoring: The Importance of Patient Activation and Engagement with the Technology

Home telemonitoring and remote feedback between clinic visits for asthma

Demystifying theory and its use in improvement

Oxygen Saturations Less than 92% Are Associated with Major Adverse Events in Outpatients with Pneumonia: A Population-Based Cohort Study

Pulse Oximetry for Monitoring Patients with COVID-19 at Home. Potential Pitfalls and Practical Guidance

Patient Attitudes Toward Mobile Phone-Based Health Monitoring: Questionnaire Study Among Kidney Transplant Recipients

Attitudes of Heart Failure Patients and Health care Providers towards Mobile Phone-Based Remote Monitoring

Strengthening the UK primary care response to COVID-19