Expert comments

Comments on this technology were invited from clinical specialists working in the field. The comments received are individual opinions and do not represent NICE's view.

All 8 specialists were familiar with or had used the technology before.

Level of innovation

Three experts observed that the novel aspects of the technology relate to the automatic capture of physiological measurements instead of manual data entry, and how this can be extended to all aspects of identifying a deteriorating patient. This includes recognising frailty and developing acute kidney injury or sepsis. The technology also enables an overview of the hospital's acuity of illness and intervention to counter deterioration in patients who would benefit from more aggressive support. It would also minimise harmful or futile intervention in patients who are dying. Three other experts broadly concurred that early warning score (EWS) systems are a minor variation to current standard care, removing the risks to patients caused by human error in chart-based calculations by recording observations electronically and integrating with electronic health record (EHR) systems to produce alerts. Another commented that extensive resource impact studies have been carried out in their NHS organisation. Because they own and build these systems, they can react to findings and implement service change as needed. However, they also said that EWS systems are not yet well-validated and there are patient groups for whom exceptions must be considered, such as patients in burns, neurosurgery and paediatric settings.

Potential patient impact

There was a general consensus from all experts that the technology has resulted in more reliable monitoring of the National Early Warning Score (NEWS2), including:

  • less failure to record observations and escalate incidents

  • fewer calculation errors

  • the added opportunity for non-verbal communication through text alerts.

In one case, local data suggest a trend of reduced moderate harm and increased low and no-harm incidents after roll-out, and potential reductions in unplanned admissions to critical care.

Two experts also observed that a system that prompts staff about a required action improves patient safety and the timeliness of interventions. In addition, electronic systems offer clarification over who should be contacted, according to the NEWS2 score. However, another expert identified a potential risk to patients through the automatic upload of electronic observations, because nursing contact time with the patient is vital to make an overall assessment of their condition. Local experience is that a nursing concern can precede physiological changes by a median of 3 to 4 hours. Therefore their trust has opted for bedside observations and manual data entry into the electronic system instead of automatic data upload. The trust plans to re-enable automated data upload in the future so it can check for manual transcription errors.

Another expert observed that all hospital patients are likely to benefit from these technologies, because patients who become unexpectedly unwell are often the most at risk of harm from delayed escalation of care, rather than critically ill patients, who are already under close observation in NHS hospitals. Earlier recognition of deterioration and timely intervention is likely to improve patient outcomes.

Potential system impact

According to one expert, the technology saves approximately 1 minute per patient, per set of observations, releasing up to 140 hours per day of staff time across their NHS trust (90 wards). There are also financial benefits from reduced critical care admissions. Reduced length of stay releases bed occupancy.

Another NHS trust said that it has significantly reduced its set-up costs and recurrent revenue by building an in-house EWS system, requiring 1 band 6 programmer and using the existing infrastructure of the IT department to maintain and support after-project delivery. This large acute trust is part of NHS England's Acute Global Digital Exemplars programme, which will share its experiences with the rest of the NHS. Experience to date has shown that hospital patient flow can be more intelligently managed. Ward rounds can be prioritised by seeing the sicker patients (higher NEWS2 scores) first. Admissions can be distributed to less dependent wards; that is, the total dependency of a ward by cumulative observation frequency can inform how much nursing resource is required to monitor patients on that ward. More widely, the technology may enable a regional network review of hospital dependency in managing any major or mass casualty incident, diverting patients who do not need major trauma care to units with lower dependency, away from the trauma centres at times of peak demand, where appropriate.

A expert at another NHS trust advised that they will be able to study the technology in a hospital that currently does not have an electronic health record system. Outcomes of interest include the impact of rapid early response to the deteriorating patient on critical care admission. The installation costs of adopting the technology are recognised and staffing resource requirements may also increase as early intervention teams are needed. However, such costs may be offset by decreased costs of overall admissions to critical care, or reduced length of stay in critical care and hospital wards.

General comments

One expert said that there is not enough published evidence to give a reasonable opinion on the systems in the scope of this briefing and that further research in NHS hospitals is needed. Most wards in the NHS are reliant on old, slow computers and generally poor-quality operating systems, which sometimes prevent clinicians accessing vital data. They said that how easy it is to log on and input data is crucial and should be a key outcome measure in any future studies. If the NHS adopts the technology there will be the opportunity for large-scale research on patient deterioration. The commentator also said that the commercial system providers should be mandated to automatically keep up with proposed NHS changes without large costs to organisations, and to ensure interoperability with other computer systems across the NHS. The future aim should be for physiological observations captured in the community, from GP and ambulance systems, and other secondary care settings, to seamlessly follow the patient. Furthermore, historical records should be maintained, so that baseline physiological data can be easily compared.

Another expert observed that, for widespread adoption, the technology must be timely, applicable and have a good user interface, and ideally be clinician-designed and led. Wi-Fi and communication networks must be high-speed and readily accessible in NHS hospitals. Hardware must be up to date and widely available to allow individual data to be inputted from every bed space in the hospital. They said that all inpatients in their organisation are monitored with the technology (1,200 beds and more than 25,000 admissions per year). However, this is not yet the situation for large numbers of secondary and tertiary care providers in England. Another expert expressed concern about what happens if Wi-Fi is lost and devices cannot be used; then back-up paper charts are needed. Other concerns are potential faults and errors in calculation or alerting (however systems are required to mitigate for this) and the potential for over-reliance on systems, so staff do not react to a clinical concern if the system fails to alert even when there may be other signs of deterioration. One expert noted the potential for over-alerting and alert fatigue, which may divert or interrupt staff resources allocated to deal with deteriorating patients. One expert suggested that an additional core requirement is for 'worry' to be built into systems, which would generate an action and potential escalation based on the clinical judgement of frontline staff, but would not add to the NEWS2 score.

One expert advised that their trust has approximately 7,000 devices and undertakes around 50,000 observations per week, on around 5,000 to 8,000 patients. They experience around 3,000 automated escalations per week. Another trust admits an average of 90 adult patients per month in an unplanned manner to critical care, with an average predicted mortality of 20% to 30%. Small improvements through earlier identification, leading to a reduction in severity of the patient's condition at the time of critical care admission will be associated with a mortality reduction (because physiology accounts for around 80% of the Intensive Care National Audit and Research Centre (ICNARC) mortality prediction model). Therefore the technology will save a significant number of lives. Ward staff at this trust like the system and prefer it to paper charts.

Another expert concluded that it is often difficult to prove the effectiveness of a single intervention (such as the introduction of an electronic EWS technology) when it comes as part of a complex healthcare intervention. For example, an electronic solution will not be successful unless accompanied by education, feedback, and re-evaluation as part of a trust executive-led change management culture that encourages quality improvement.