Nursing staff perspectives of continuous remote vital signs monitoring on surgical wards: Theory elicitation for a realist evaluation

Abstract Rationale, Aims and Objectives Continuous remote monitoring (CRM) provides a novel solution to the challenges of monitoring patients' vital signs in hospital, but the results of quantitative studies have been mixed. Acceptance by staff is a crucial determinant of the success of healthcare technologies and may explain these discrepancies. Drawing on the approach of realist evaluation, this paper aims to identify theories about how, why and in what conditions nursing staff perceptions vary regarding the CRM of patients' vital signs. Methods Multiple methods were used to elicit theories about factors likely to facilitate or impede the successful implementation of continuous remote vital signs monitoring. This included a literature review, consultation with patients and observational work conducted during a randomized controlled trial (RCT) of CRM. In addition, a priori theories developed through informal interactions with patients and ward staff during the day‐to‐day set‐up of the trial were included. Results The findings suggest that the perceptions of nursing staff regarding remote monitoring can be influenced by the type of patients under their care and their previous experience of telemetry. Factors which may undermine the engagement of staff are perceived staff burden, which can be dependent on contextual factors such as staffing levels, time of day and senior staff attitudes. Staff attitudes are also likely to be influenced by patient perspectives and the utility of the devices associated with remote monitoring. The successful implementation of CRM may be dependent on staff training, research staff input and hospital culture. Conclusions Theories regarding nursing staff engagement with remote monitoring are numerous, varied and contradictory. The theories elicited in this initial phase will be refined during interviews with the nursing staff involved with the RCT.


| INTRODUCTION
The successful implementation of new healthcare technologies into routine clinical practice is predicated on engaging both staff and patients. 1 It is crucial to assess the experiences of the people using the technology to identify contextual factors that support or constrain optimal utilization, which could influence the effectiveness of the device. 2 The remote monitoring of patients' vital signs is an area of increasing interest due to the innate limitations of manual vital signs monitoring in hospital. 3 The SensiumVitals ® remote monitoring system consists of a patch worn on the patient's chest and continuously measures heart rate, respiratory rate and temperature.
These data are transmitted wirelessly to a mobile device which alerts the nurse if the vital signs stray outside of normal parameters, potentially allowing earlier detection and treatment of patient deterioration in hospital. This technology has been evaluated in two feasibility studies in the surgical population. 4,5 The TRaCINg feasibility randomized controlled trial (RCT) compared usual intermittent vital signs monitoring, in the form of the National Early Warning Score (NEWS), and continuous remote monitoring (CRM) in addition to NEWS.
The small number of quantitative studies in the field of continuous monitoring have shown mixed results. 6 The success of these technologies is context-dependent and reliant on both patient and practitioner engaging effectively with the technology. We have previously studied the perceptions of patients regarding continuous vital signs monitoring in hospital 7 ; in this study, we undertook semistructured interviews with surgical inpatients as part of a study testing a remote continuous monitoring device and analysed the results using thematic analysis.
Realist methods were chosen to determine the perceptions of staff members, given that acceptance by staff may be the single most important determinant of the success of healthcare technologies at a local level. 1 Realist evaluation is a theory-driven approach to the evaluation of complex interventions in healthcare. 8 It is based on the idea that interventions (such as a new monitoring system) offer resources to people, but it is how people choose to respond to the resources that determine their impact, and such choices are highly dependent on context. 9 Realist evaluation aims to explain why the intervention works in some circumstances, but not in others. It involves eliciting stakeholders' theories and then gathering empirical evidence to test and refine those theories. In realist evaluation, the term 'theory' refers to participants' ideas and thoughts about how an intervention works, based on their everyday experience. This type of 'informal theory' is always at work in improvement work, although practitioners are often not aware of it or do not make it explicit. 10 Staff are the users of the monitoring system from a realist perspective, and we were interested in their response to the system as this will determine the impact of the intervention on patients.
This paper presents the theory elicitation phase of the realist evaluation that was undertaken alongside the TRaCINg feasibility RCT.
The theory elicitation phase aimed to identify stakeholders' theories concerning how, why and in what conditions continuous remote vital signs monitoring is optimally used on the surgical wards of a large teaching hospital. Elucidation of these contextual factors and their effects will inform potential wider implementations of this technology and may reveal strategies to support staff in the future.

| METHODS
The first phase of a realist evaluation is that of theory elicitation. 11 Realist theories are presented in context-mechanism-outcome (CMO) configurations, with the mechanisms divided into resources and responses. In the case of CRM, the technology provides a fixed resource; it is the response to the resource that determines if the desired outcomes are achieved. This response is determined by the context in which the resource is implemented; for instance, the clinical area itself, or the experience levels of the staff employed there. As an example, in the context of engaged senior colleagues, staff nurses may respond by carrying the devices and acknowledging alerts appropriately, leading to recognition of the deteriorating patient (the desired outcome).
Multiple methods were used to elicit theories about factors likely to facilitate or impede the successful implementation of continuous remote vital signs monitoring. This included a literature review, consultation with patients and observational work conducted during the TRaCINg study. In addition, a priori theories developed by CD through informal interactions with patients and ward staff during the day-to-day set-up of the study were included. In brief, the selection and appraisal of identified papers were based on relevance to the review question, as is the case in the theory elicitation phase of a realist review. 12  of the selected papers was not undertaken as the purpose was solely to identify potential theories to be refined in later stages of the research, rather than evaluate the truth of the theories at this stage. Theories and theory fragments were extracted from the literature and then grouped together and refined as the review progressed. Conflicting theories were also included, with care being taken to note the context in which these contradictory ideas were founded.

| Patient consultation
Patients' ideas about nursing perceptions of CRM were gleaned from face-to-face interviews at the hospital beside, 7 informal interactions during the day-to-day management of the TRaCINg study and two patient focus groups conducted as part of the Patient and Public Involvement work ahead of the feasibility trial. The full methodology of the patient interviews has been published elsewhere. 7 The topic guide for the focus groups was developed for this study and is provided as Supplementary Material. Data from the transcripts of the interviews and focus groups were coded to identify themes in the participants' responses. These codes were then refined to identify patient theories, which were added to those identified in the literature review. These field notes were reviewed after the end of the TRaCINg study and coded to identify common themes, which were further refined to draw out new theories concerning the perceptions of nursing staff with regard to the CRM devices. These theories were added to those identified through the literature review and patient consultation alongside a priori theories developed by CD through informal interactions with patients and ward staff during the day-to-day set-up of the study.

| Literature review
The search retrieved over 1000 references. After the selection process, a total of 84 sources were identified. Three papers were systematic reviews of studies of continuous vital signs monitoring; one article was a nonsystematic review. There were 25 individual studies of CRM, including both quantitative and qualitative data.
These were evaluated together with 16 editorials and 39 websites.
There was considerable repetition of theories across the sources identified. These theories largely fell within three larger themes: nursing perceptions of CRM, the development of CRM technologies and the implementation of CRM technologies.

| Theories regarding nursing perceptions of CRM
Five studies specifically reported nursing perceptions of CRM systems [13][14][15][16][17] and all identified similar themes. In general, nursing staff could see the potential for continuous monitoring to enhance patient safety. Nurses perceived that greater 'availability and accessibility' of vital signs information would support their decision-making and provide reassurance to patients. 15 Context did appear to have a role in determining the perceptions of nursing staff. Jeskey et al. 16 found that nurses with prior telemetry experience were more likely to perceive the monitoring device as beneficial and more clearly understood the device. It was also suggested that the devices were perceived to be more beneficial by night staff rather than during day shifts, potentially due to reduced staffing levels and more frequent monitoring of high-acuity patients 'in the immediate postsurgical period'. 16 An alternative theory was that in the context of night shifts, the increase in patient: staff ratios may lead to the devices being perceived as an addition burden (response), causing failure to engage (outcome).
Two papers reported that nurses were worried that visibility of information and alarms would cause patient anxiety, leading to increased time spent to reassure them. 15,18 Both of these studies were conducted on respiratory wards, which may have highacuity and therefore high-anxiety patients. However, the visibility of information on CRM devices was also considered to provide opportunities for increased engagement of patients in their own care.

Prgomet et al. reported concerns from both doctors and
nurses about over-reliance on CRM leading to decreased bedside interactions. 15 A conflicting yet recurring theme across the literature was that of staff burden. Van Loon et al. highlighted the fact that CRM devices typically collect large amounts of information, which has the potential to overwhelm users and dilute important indicators of deterioration. 18 Other studies reported concerns that CRM overburdens busy ward staff or takes nursing staff away from other tasks. 19 This was particularly evident during day shifts, when staff are typically busy with a wider variety of duties than during night-time hours. 16 The underlying theory appears to be that In the context of a busy ward environment, the nurses will be too busy for an extra task and will fail to engage with the devices (their response), leading to clinical deterioration going unrecognized (the outcome).
Eight studies reported concerns about alert burden. These studies shared a common context of high acuity patent populations and higher patient: nurse ratios. Banks et al. found such a problem with nuisance alarms that monitoring had to be abandoned because of nursing complacency towards the alarms. 17 Alarm fatigue and data inaccuracy were also reported by Jeskey et al., who found that excessive false-positive alerts interrupted nurses and distracted them from other responsibilities. 16 There was also concern that doctors might become overburdened and desensitized to calls. 15 This suggests that in the context of very sensitive devices, there will be a high number of false alerts, leading to alert fatigue, desensitization and failure to respond to alerts (the response), with the outcome of unrecognized deterioration.

| Theories regarding development of CRM technologies
Three articles commented on the limitations of current CRM devices, outside of concerns about false alerts. Patient comfort was a priority 15,18 ; the underlying theory appears to be that in the context of patients finding the devices uncomfortable, or feeling anxious wearing them, the nursing staff may consider the devices to offer more harm than good, with the outcome that they fail to engage with the CRM technology.
It is also suggested that nursing staff should also feel comfortable with the devices 20 to avoid losing confidence in the technology as a whole. In the context of nurses lacking confidence when using new technology, their response will be to fail to engage with the devices leading to the outcome of unrecognized clinical deterioration.
Other theories suggested that merely notifying caregivers of abnormal readings are inadequate and that usability of devices would be improved by incorporating a suggested action in response to notifications, 18 especially where devices collect a large amount of data for interpretation. In the context of the devices gathering large amounts of information, the nursing staff may feel overwhelmed (response) leading to a lack of confidence when interpreting and acting on alerts (outcome). In addition, in the context of the devices not suggesting an action after an alert, the nursing staff may not know how to respond to the alert and may subsequently fail to act on notifications. Basing suggested actions on local policy could enhance perception of CRM as integrated into the usual care pathway. 20 If the CRM system is not incorporated into local protocols and policies (context), nursing staff may be ambivalent towards the technology (response) and fail to engage (outcome).
Another potential way to improve integration is to remove notification devices from individuals and instead promote a wardbased responsibility for CRM, by incorporating big screens at the nurses' station. 21,22 In the context of wards being divided into sections, each of which is the responsibility of a single nurse, the nurses may perceive the device as an individual burden, rather than as a collective responsibility, leading to disengagement from the system and decreased responsiveness to alerts. Allowing ward-based responsibility could help overcome another limitation of individual nurse responders: in the context of nursing staff only seeing the benefit of the CRM system on a patient-by-patient basis, or only in patients who have deteriorated, they may underestimate the global impact of the devices (response) with the outcome that their engagement with the CRM devices is impaired. Another suggested tactic to improve engagement of early adopters was incentivizing staff to use the devices appropriately 23 ; suggested incentives ranged from updating staff about recent patient success stories, ranking wards against each other or providing 'gifts'

| Theories regarding implementation of CRM technologies
to highly engaged teams. In the context of staff not having incentives to respond to alerts, they may not be motivated to engage with the devices (response) with the outcome that they do not respond to alerts.
Other theories concerned the context of initial implementation.
One broad idea was the need to ensure that innovation is supported in the local hospital culture. 20  it is incorporated into local monitoring protocols alongside explicit escalation guidance. 18,20 To this end, it might be helpful to extend staff training in the new technology to non-ward-based staff such as doctors and outreach teams. 15 If vital signs monitoring is considered to be an exclusively nursing task (context), the nurses will not be empowered to use the CRM alerts when escalating a patient to doctors (response), and will therefore no longer consider the monitoring systems to be worthwhile within their care protocols (outcome). Incorporating CRM alongside traditional observations could increase perceptions of its utility, encourage nursing staff engagement and incite wider institutional acceptance.

| Patient consultation
The full analysis of the patient interviews and focus groups has been reported previously. 7

| Nonparticipant observation
This was a particularly rich source of theories which incorporated informal, 'throwaway' comments from ward staff and close observation of interactions between and within staff members and patients.

| A priori theories
Theories were developed by CD through informal interactions with patients and ward staff during the day-to-day set-up of the study.