This chapter draws on a manuscript by Herlitz et al.214 uploaded as a preprint to MedRXiv. This is an open access article 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 the original work is properly cited (see https://creativecommons.org/licenses/by/4.0/).
What was already known
- Studies have shown that patients and staff like remote home monitoring, but there are some barriers to delivery and engagement, and it may not be appropriate for all patient groups.
- The availability of dependable, high-quality, and clinically-useful technology systems that patients and staff can engage with influences implementation of services.
- Few studies have explored patient and staff experience of COVID-19 remote home monitoring services.
What this chapter adds
- Older patients, patients with a lower level of educational attainment and ethnic minorities were more likely to relay symptoms through phone calls with the service.
- Staff considered tech-enabled and analogue models better equipped to manage large patient numbers; but improvements were needed to improve functionality of technology systems to better fit clinical and operational needs.
- Tech-enabled and analogue models were not a substitute for human contact, as phone calls were used in both tech-enabled and analogue services and analogue-only services.
Background
COVID-19 remote home monitoring services have been implemented using a range of technology-enabled and analogue modes (see Vindrola-Padros et al.32 and Vindrola-Padros et al.36 and Chapters 1, 3 and 4).
For patients to engage with remote home monitoring services, previous research has indicated that patients need access to private space, carer support, the necessary digital infrastructures required to engage with remote home monitoring services,208 a choice of modality and access to technical support.215
For staff to deliver remote home monitoring services, previous research has highlighted the importance of the accessibility and functionality of technology for meeting the needs of work practices, interoperability with existing data systems and sufficient staffing for monitoring and responding to data.164,216–218 Furthermore, the beneficial impacts of technology on work activity can be reduced by technical problems and a lack of technological infrastructure.219
Research indicates that patients and staff have emphasised the need to maintain human contact in the clinical relationship to preserve high quality care.216,220–222
While previous research provides an indication of the types of factors that may enable staff and patient use of remote home monitoring services, the experience of patients and staff when using different modes of COVID-19 remote home monitoring have not yet been fully explored. Emerging evidence indicates that digital apps may facilitate the follow-up of a higher number of patients with COVID-19 compared with telephone calls,32 but might also limit the engagement of older patients.
Within this chapter, we report on patient and staff experiences of engaging with rapidly implemented tech-enabled and analogue models of remote home monitoring of patients with COVID-19. We hypothesise that differences in how patients relay readings in the models may impact on how patients and staff engage with and experience the service. We aimed to answer the following questions:
- Which models of tech-enabled and analogue COVID-19 remote home monitoring were adopted in the UK?
- Were there differences in the characteristics of patients using these models?
- What were patients’ experiences of relaying symptom data using tech-enabled and analogue models?
- What were staff experiences of processing symptom data for large numbers of patients using tech-enabled and analogue models?
- How did the use of tech-enabled and analogue models affect patients’ and staff’s experience of the patient–clinician relationship?
Methods
A summary of the methods is provided below; further details can be found in Chapter 2 or Report Supplementary Material 1.
We used the technologies (dependability, functionality and familiarity) and clinical relationship (patient–clinician interaction, knowledge of patient and trust) domains of the Planning and Evaluating Remote Consultation Services (PERCS) conceptual framework to examine the newly established services.208
Design
This chapter draws on staff and patient/carer surveys and staff and patient/carer interviews. Data collection took place between February and June 2021 (patient surveys retrospectively included January).
Sample
Patients/carers who received these services and staff leading and delivering services or involved in service data collection/analysis (see Chapter 2 and Chapter 4 for sample details).
Measures
We developed the survey and semistructured topic guides specifically for this study (see Chapter 2 and Project website documentation – Study instruments). Interviews with delivery staff included a ‘think aloud’ section where staff narrated the process of using the platform.
Data collection
We conducted a survey with patients/carers from 25 sites and with staff from 28 sites. We conducted interviews with staff, patients and carers from 17 sites (see Chapter 2 for details).
Data analysis
Inductive thematic analysis was used to analyse interview data (relating to patient views/experiences of relaying symptom data, staff processing data and interactions between patients and clinicians), inputted into RAP sheets.70 The PERCS domains of technologies and the clinical relationship were used as a sensitising lens.208 Transcripts were reviewed to gain further in-depth data on reasons for adoption of models and knowledge gained from models. Each code’s data were checked for consistency of interpretation and re-coded as necessary. The data were organised into lower- and higher-order themes. We constructed the typology of the models based on two key characteristics: the modality of the technology and human contact between patients and clinicians. Data relating to modality and frequency of human contact were coded deductively.
Survey data were analysed using SPSS (version 25). The following patient data were analysed: sociodemographic information, patient-reported ease of completing activities as part of their care, frequency and quality of contact with the remote monitoring team, and whether they had received any support from family or friends. Data related to experience of relaying data and the clinical relationship were also coded deductively from two open-ended survey questions. Survey responses were compared with patients who had reported they had relayed readings via tech-enabled or analogue modes. We analysed staff survey respondent characteristics. Staff reported on their experience of delivering the service, capacity/resources to deliver the service, the impact of the role on their workload and their views of patient engagement with the service. Data related to processing and responding to symptom data and the clinical relationship were coded deductively from one open-ended survey question. To make comparisons across models, staff were categorised as delivering a tech-enabled and analogue model if sites adopted a mixed approach, or an analogue-only model if a tech-enabled platform was not used.
For patient and staff survey data, we used descriptive statistics including frequencies and percentages, and univariate analyses were performed. Where data were missing for specific questions, available case analysis was performed and the denominator reported. We conducted Pearson’s chi-square tests and logistic regression modelling to compare the characteristics of patients using tech-enabled and analogue-only models and to determine the association between patient characteristics and mode of submission. We used non-parametric Mann–Whitney U-tests to compare patient-rated experiences of receiving services and staff experiences of delivering services for tech-enabled and analogue-only models.
Results
Participant characteristics
Surveys were received from 292 NHS staff and 1069 patients and carers. We conducted 58 interviews with staff and 62 interviews with patients/carers (see Chapter 4 and Report Supplementary Material 2 for demographic characteristics).
The majority of staff reported delivering tech-enabled and analogue services (81% interview participants, 77% survey participants) compared with analogue-only models (19% interview participants, 23% survey participants).
More patients reported using a tech-enabled mode for data submission (66% of interview participants and 51% of survey participants) than an analogue-only mode (31% of interview participants and 49% of survey participants). Of the patients surveyed: 29% reported providing readings via text, 25% via digital app, and 1% via e-mail.
A comparison of participant characteristics for survey participants who used tech-enabled compared with survey participants who used analogue methods of submission is shown in Report Supplementary Material 8. The logistic regression analysis found statistical evidence of differences in patient characteristics between survey participants who used tech-enabled versus analogue-only models, for age (p = 0.005), level of education (p = 0.011) and ethnicity (p = 0.043) (see Report Supplementary Material 8).
Models of COVID-19 remote home monitoring adopted
Services were categorised into four models (Figure 25). Services within model 1 (10/17 services) and model 2 (4/17 services) offered tech-enabled and analogue services submission modes. Services within model 3 (1/17 service) and model 4 (2/17 services) offered analogue-only modes of submission. Models 1 and 3 outline ‘universal’ services which offered the same submission options and frequency of human contact to all participants, whereas models 2 and 4 outline ‘risk-stratified’ services which adapted the submission method and/or frequency of contact depending on patients’ needs. For example, two services within model 2 offered self-monitoring (analogue-only options to low-risk patients) and some services within model 4 offered more frequent phone calls or face-to-face visits for high-risk patients. See Report Supplementary Material 4 for further details.
FIGURE 25 Models of COVID-19 remote home monitoring. (Figure not reproduced here; see the original article.)
Patients’ experiences of using tech-enabled and analogue methods to relay symptoms
We constructed two key themes related to patients’ experiences of relaying symptom data using tech-enabled and analogue, and analogue-only modes: (1) allocation of mode; and (2) ease of relaying readings and the functionality of tech-enabled systems.
Allocation of mode
Most patient interviewees did not recall being given a choice about the mode in which they could submit readings; however, most patients were happy with the mode they were given. Several patients mentioned that it had been unproblematic to switch modes during their time with the service. Staff interviewees across many sites reported that they were aware that older patients, in particular, might lack digital skills/infrastructure or experience cognitive/physical impairments that might make engaging with tech-enabled systems challenging. These patients were offered phone calls as appropriate. Given the differences between patient groups using tech-enabled and analogue-only modes, staff may have been attuned to which option might best suit patients or may have assumed that certain patients (e.g. older patients) would not be able to manage a tech-enabled approach.
Ease of relaying readings and the functionality of tech-enabled systems
Most patient interviewees across sites reported that it was ‘easy’, ‘simple’ or ‘straightforward’ to relay readings whether by tech-enabled or analogue modes. A number of patients thought that the simplicity of the tech-enabled systems was its strength:
I’m really un-techy, but . . . the text message reminder came through, you clicked on it, you put in . . . your name and date of birth, it took you straight through it . . . you could be as thick as you like and you would manage to do that. It was really [laughs] . . . no technical prowess required.
Similarly, survey analysis found patients rated both tech-enabled and analogue-only modes for submitting data positively. Patients using a tech-enabled mode reported that recording readings (p = 0.009) and providing readings (p = 0.001) to services was easier than patients using analogue-only methods (see Report Supplementary Material 8). However, differences might reflect differing characteristics of patient groups.
The main barriers to submitting readings (for both tech-enabled and analogue, and analogue-only models) reported by patient and staff interviewees were feeling too poorly or tired to engage or forgetting to submit readings. Many patients and staff highlighted that family members or carers had assisted patients in relaying readings, particularly for patients using a tech-enabled modality, who lacked confidence in their technical skills:
In the beginning I found it [entering the information] difficult because I was poorly but then I had a member of my family to help me and then it was just easy to do and I’m not good at things like that. I found it easier to do after a while.
The functionality of the technologies was not discussed in detail by patients, however several patients and staff reported issues particularly related to set-up because they required downloading and remembering log-in details. Two sites offered face-to-face support to patients at onboarding for this.
Staff experiences of using the models to process symptom data for large numbers of patients
Two themes emerged as key to staff experiences of processing symptom data: the scalability of the models, and the ease-of-use and functionality of tech-enabled systems.
Scalability of the models
Tech-enabled and analogue models were considered by staff to be more scalable than analogue-only models. Service leads from three sites reported in interviews that they had adopted tech-enabled platforms to ensure services could cope with a high volume of patients, and staff from six tech-enabled sites highlighted that there were efficiency gains from data being entered by patients in real time:
At the peak we were on I think hundred and ninety-six on the service . . . and each of them were submitting three or four points of data, three times a day . . . . that’s well over two thousand bits of data coming in a day. There’s just no way manually we would have been able to do that.
Staff interviewees from two analogue-only sites noted that phone calls were labour intensive and necessitated a larger workforce (relative to patient numbers), and manual data entry, was required for each patient so that services could keep abreast of patients using the service. Staff survey findings indicated that mixed tech-enabled and analogue models had more positive impact (p = 0.034) on workload compared with analogue-only services: 52% of staff using tech-enabled models reported a positive impact while only 35% of staff using analogue-only did so (see Report Supplementary Material 8). However, service lead/manager perceptions of the capacity/resources needed to deliver services (as reported in the survey) were not statistically significantly different (p = 0.554) across models (although the smaller sample size for this question should be noted, see Report Supplementary Material 8).
Ease-of-use and functionality of tech-enabled systems
Staff reported monitoring patients as statistically significantly easier (p = 0.005) for tech-enabled and analogue service models (77% of staff reported monitoring as easy/very easy) compared with analogue-only models (56% of staff reported monitoring as easy/very easy; see Report Supplementary Material 8). However, staff interviewees reported an extensive number of ways that tech-enabled platforms could be enhanced to support their clinical and operational needs (see Appendix 5, Table 28), and staff across several sites reported that they were frustrated by the length of time it took for tech providers to make changes to platforms.
Patients’ and staff’s experiences of the clinical relationship
Three themes emerged relating to patients’ and staff’s experiences of the patient-clinician relationship embodied in the remote home monitoring models: staff knowledge of patients’ condition, reassurance and providing appropriate care, and concordance in the patient-clinician relationship.
Staff knowledge of patients’ condition
Phone calls were an essential element of both models, enabling staff to gain comprehensive knowledge of their patients’ condition. Staff using tech-enabled and analogue models used the data they received to get an overview of patients’ health status and called patients if there were signs of deterioration to conduct a more thorough assessment (e.g. assessing breathlessness, retaking readings, measuring oxygen saturation after exertion, and assessing patients’ coherence and anxiety). Phone calls were also useful for gaining information about patients’ social circumstances and mental health, sometimes prompting referrals to other services:
Many of the patients when we come across . . . live on their own. So, it’s not just their health it’s the . . . wellbeing of the patients as well . . . There are many times where I have done the referrals . . . and provided more support for them, for food and other medications and things like that.
Some staff members (across both tech-enabled and analogue and analogue-only models) highlighted that on the phone (or in person) patients were more likely to mention other medical symptoms in passing, which might require attention. Staff from two analogue-only sites reported that the depth of knowledge gained from phone calls was one of the key reasons for not adopting a tech-enabled model of care.
Reassurance and providing appropriate care
Patient and staff interviewees across all sites reported that patients found services very reassuring at a time when they were feeling very unwell, isolated or anxious about the disease. There was no evidence of a difference (p = 0.559) between models in how patients rated contact with the service team (see Report Supplementary Material 8). Staff across all sites reported that the opportunity for patients to discuss symptoms with a clinician reassured them and prevented patients from unnecessarily contacting emergency services or encouraged deteriorating patients to go into hospital when they were afraid to:
So a lot of [patients], when you speak to them, and you say, ‘Actually look, your levels were 92. I’ve exerted you for a minute they’ve dropped down to 89, can you hear yourself gasping on the phone?’ They say, ‘Yes.’ And I say, ‘Well that’s really serious you do need to call an ambulance.’ And then they will come around to it.
Regardless of the mode (p = 0.158), patient survey respondents reported that contacting a health professional was relatively easy. However, seeking further help if they had concerns about their health was statistically significantly easier (p = 0.006) for patients using tech-enabled modes (see Report Supplementary Material 8). Several patients using tech-enabled models reported that despite receiving an automatic prompt to escalate their care when readings were low they preferred to first speak to the service for a more personalised/thorough assessment.
Managing the patient-clinician relationship
Patient interviewees reported that being contacted by staff if they had missed readings gave them a sense of security and reassurance. However, for staff, chasing and managing non-submitters could be time-consuming, frustrating and worrying, consuming valuable resources. Services tried to encourage patients to relay data by setting out service requirements at onboarding, and emphasising the importance of relaying readings:
What we found initially was that [expectation of submitting twice a day] wasn’t explained fully and then we were having to chase them a lot for readings. So that is really explicit at the start now that . . . almost like as daft as it sounds like a verbal contract. You know, ‘We will do all of this however, what we expect of you is you know readings twice a day to be compliant’.
Several staff delivering tech-enabled and analogue models highlighted the importance of patients understanding that while readings were reviewed regularly, communication was not synchronous and patients still needed to call emergency services if oxygen saturation levels were low. However, some patients reported preferring to speak to services before escalating their care.
Discussion
Key findings
Tech-enabled care was not a substitute for human contact in COVID-19 remote monitoring services; phone calls and human relationships were a feature of all models of remote monitoring and patients found services reassuring (across models). This indicates that technology cannot completely replace phone calls. Patients using tech-enabled methods tended to be younger, well-educated and more likely to identify as white British compared with those using analogue methods and most patients did not recall being given a choice about the mode offered. Staff considered tech-enabled and analogue services to be more scalable and able to deal with larger volumes of data and patients than analogue-only services.
How findings relate to previous research
Consistent with previous studies,164,216,218 staff noted that better functioning systems that fit with their work practices would have greatly improved efficiency, highlighting the need to involve clinicians in the design and development of technical platforms. Patient involvement was also missing from their development; patient participation groups and patient experience work were considered non-essential early on in the pandemic.223 As a number of different tech-enabled models have now been established, there is an opportunity for service leads to trial different systems and incorporate patient feedback/involvement.
Patients in all models were contacted by telephone if their symptoms deteriorated or they did not submit readings, with the exception of two sites that offered a self-monitoring service. Staff reported that pursuing non-submitting patients could unnecessarily consume resources. Consistent with previous research showing staff concerns about patient safety in chronic conditions, the acute nature of COVID-19 caused staff anxiety about rapid deterioration.164 There is a question of whether services’ resources could have been better optimised for patients’ needs, by addressing concerns about risk in staff supervision and training, by offering more lower-risk patients self-monitoring, by adapting tech-enabled platforms to offer varying levels of automated contact, and/or following more stringent protocols for managing non-submitters (e.g. discharging patients back to the care of their GP if they miss three data submission points).
Previous studies have shown that both patients and staff have responded positively to remote home monitoring for chronic health conditions,216,224 and a small number of studies have been conducted on patients’ experiences of COVID-19 remote home monitoring.211 This is the first study to examine in-depth patient and staff experiences of using tech-enabled and analogue modalities to respectively relay and process COVID-19 symptom data and compare experiences of the patient-clinician relationship. It builds on our work looking more broadly at patients’ experiences of COVID-19 remote home monitoring services (see Chapter 9).
This study supports prior work that has shown that some patients may need additional support to engage with technology or access to analogue options,208,220 and the importance of maintaining human contact.221,222 Older patients, less well-educated patients and patients identifying as non-white were more likely to relay symptoms through phone calls with the service, with staff primarily determining the allocation of mode. Co-designing digital health services with patients, improving digital health literacy in the population, training and supporting the workforce to build digital skills and confidence, and gathering data on patients’ access to technology and digital skills could widen patient participation in digital health care.206,215
Staff engagement has been explored in previous studies,164,216 but this study extends previous research by considering the scalability of models, with staff considering tech-enabled and analogue models better equipped for managing large patient numbers without jeopardising clinical relationships. However, future models of remote monitoring should tailor their frequency of contact to patients’ level of risk and include systems which create a dialogue between patients and staff as to their preferences.
Strengths and limitations
There was a high level of variation between different models (see Chapter 4). Few sites adopted analogue-only modes of data submission, and no sites used tech-enabled-only modes, making it difficult to assess whether differences in patient and staff experiences between tech-enabled and analogue models were due to the engagement with the technology, or related to other service factors. This also meant that it was not possible to use multivariate analyses to look at staff experiences (based on survey responses) across models, due to the limited sample size in analogue-only sites. The univariate analyses of staff and patient experiences could therefore have been subject to false positives due to the number of comparisons made. However, in-depth data collection from staff on the functionality of tech-enabled models at four sites enabled us to draw out specific tech-related findings.
We had originally planned to use ‘think aloud’ methodology with patients but in practice, these did not work because patients no longer had access to the tech-enabled platforms they had used and their recall of their use and functionality during ill health was poor.
The language and terminology describing technology was unfamiliar to some staff and patients, for example, understanding the difference between a web-link and a digital app.
Conclusions
Organisations implementing tech-enabled remote home monitoring at scale should consider providing flexible models which can tailor processes and resources to accommodate patients with different needs and requirements. These include risk levels, abilities to engage with and access technology, support available at home from family or carers, and levels of desire for support versus self-management. Services routinely collecting and monitoring data on patients’ access to technological infrastructure and skills could inform digital inclusion strategies for remote home monitoring.