National early warning scores in care homes: do policy imperatives reflect a genuine need?

NEWS is increasingly forming part of policy to implement track and trigger systems in care homes. Care home residents are more frail than the hospital populations that NEWS was developed in. NEWS is likely to have diminished predictive properties in care home residents and comes with associated opportunity costs. Implementation of NEWS may not therefore be logical in a resource constrained health and social care system. In this issue of Age and Ageing, Barker et al. [1] report on the implementation of the national early warning score (NEWS) in care home residents. NEWS was initially developed in hospitals to trigger prompt response to physiological deterioration by appropriately qualified staff. It can predict adverse outcomes in hospital patients and has also been validated for pre-hospital use by ambulance staff [2, 3]. It is currently used in 76% of acute trusts and 100% of ambulance trusts across England [4]. The project reported by Barker et al. was funded by a Clinical Commissioning Group in one part of England to train care home staff to record vital signs and use these to complete NEWS scores on a digital tablet. Their data incorporated 20,643 readings over 30 months from 46 care homes. They concluded that NEWS was feasible in a care home setting but that further work was required to establish positive and negative predictive values for adverse outcomes and to consider the responses NEWS is, and should be, triggering. This work reflects national policies aimed at identifying physiological deterioration in care home residents, including work by the National Health Service Patient Safety Collaborative. The underpinning logic is that responding early to care home residents who have become unwell will result in more appropriate care. There are some problems with this assertion. In the hospital setting, track and trigger systems such as NEWS have variable predictive power. In some settings they only explain about 62% of the critical events they purport to predict [5], which is barely better than flipping a coin. Track and trigger systems may be less useful in care homes, something that needs consideration and investigation. Care home residents are affected by very significant frailty [6], and early warning scores have been shown to be less effective in frail cohorts [7]. Care home residents are also more clinically stable than hospital populations, which means the probability of acute deterioration is lower. Bayesian reasoning [8]—the appropriate way of thinking of the relevant probabilities involved in working out diagnoses or predictions—suggests that the post-test probability of NEWS correctly predicting a life-threatening event is determined by the prevalence of life threatening events and the likelihood that NEWS will correctly identify deterioration. Both of these are lower in care home populations than the hospital settings in which NEWS was developed. It is not clear what behaviour NEWS is supposed to trigger in care homes. NEWS scores place emphasis on escalation, whilst medical support for individual care homes varies. In the face of raised NEWS scores, homes that are unable to access medical services urgently are likely to trigger ambulance calls and emergency department admissions. Many care home residents are approaching the end of their lives—an important consideration is whether widespread adoption of track-and-trigger tools would adversely influence the palliative ethos that correctly prevails in many care homes. NEWS comes with opportunity costs. Time and resource spent training care home staff to use track and trigger tools, regularly collecting observations and responding to high warning scores, is time and resource that could be directed into providing more responsive, personalised, resident-centred care. Acute deterioration has been problematised from the perspective of acute healthcare providers. How residents, relatives and care home staff feel about deterioration is not at all clear. Even if care home staff could generate and use a score to correctly predict morbidity and mortality in care home residents, what sort of responses would residents want? Would the tool be for escalation, or to trigger a more palliative approach? These preferences would need to be established on a resident-by-resident basis but when, where and how would residents like such preferences to be explored? How much person-centred care would they be prepared to trade for the ability to predict deterioration, or how much more would they be prepared to pay to protect existing care provision whilst adding track and trigger tools to routine care home practice? Barker et al. stated that collecting NEWS was feasible in care homes. They meant that care home staff could measure vital signs and enter these on a tablet. It is not clear from their work, however, whether using NEWS as a tool to trigger appropriate escalation within a broader framework of care is feasible. Much more work would need to be done to establish this. The desirability of NEWS, or still-to-be-developed care home-specific track-and-trigger tools, is at least as important a question. We propose this work be done first before further implementation of a tool which, when used in care homes, may lack useful predictive power, may trigger inappropriate responses and may not be worth the human and organisational costs of implementing it. The perspectives of care home residents, relatives and staff must be sought if NEWS is to be widely implemented in care homes. None. None.