Article Text

Download PDFPDF
Case-controlled study
Early warning tools may assist in the identification of clinical deterioration in the new-born nursery
  1. Christopher Parshuram1,2,
  2. Karen Dryden-Palmer2
  1. 1Department of Paediatrics, University of Toronto, Toronto, Ontario, Canada
  2. 2Center for Safety Research, Child Health Evaluative Sciences, The Research Institute, The Hospital for Sick Children, Toronto, Ontario, Canada
  1. Correspondence to: Dr Christopher Parshuram, Department of Paediatrics, University of Toronto, 555 University Avenue, Toronto, ON M5G 1X8, Canada; christopher.parshuram{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Commentary on: OpenUrl

Implications for practice and research

  • Prospective identification of new-born infants in maternity wards who are clinically deteriorating may be assisted with the use of early identification tools.

  • Differences in physiological norms between late preterm, early-term and post-term new-born infants may limit the performance ‘one size fits all’ neonatal early warning tools.

  • Additional research about ‘early warning’ tools for new-born infants in maternity wards is needed. Specifically, refinement and re-evaluation of identification criteria are required to optimise detection characteristics as there is limited consensus on ‘critical’ and ‘subcritical’ physiological indicators for new-born populations.

  • Further consideration of how provider expertise is integrated within the tools to support timely identification is also required.


Neonates in maternity nurseries are undergoing complex transitions from the intrauterine environment to physiological independence. During this transition, the presence of congenital and, or acquired disease provides new challenges that may lead to clinical deterioration. Early signs of deterioration may be subtle and missed by busy clinicians. These clinical signs may be observed, should be reflected in valid early identification tools, and should result in timely escalation of care. The performance of available trigger criteria, visual prompts and early warning scores for neonates is largely unknown. Other than face validity, there are limited data about potential clinically utility.


In this study, the investigators used retrospective data collected during the application of a locally developed severity of illness score, the ‘Standardized Observational Tool’ (SOT). This single-centre study evaluated the performance of SOT observations when applied to three early warning tools; the Neonatal Early Warning Observation (NEWO) chart, the Australian Capital Territory (ACT) score and the New South Wales (NSW) method. Evaluated patients received care in a maternity ward. Case patients were transferred to the neonatal nursery. A 3:1 case to control study was performed using matching within each of three strata—34–36.6 weeks, early term (37–38 weeks), and post-term (>42weeks).

All vital signs were abstracted from medical records from admission to transfer to the neonatal nursery (case patients) or to hospital discharge (control patients). Thresholds for ‘escalation’ were used for the scores and trigger criteria, and were compared with the actual status of the patient (case or control) to determine the true-positive rate (sensitivity) and the true-negative rate (specificity) and related measures.


The three groups evaluated included 8, 16 and 2 case patients and their matched controls. Correct identification of escalation in the 26 case and non-escalation in the 78 control patients (sensitivity, specificity) was observed in SOT (92%, 59%), ACT (12%, 99%), MEWO (58%, 69%) and NSW (81%, 31%). Ideal would be (100%, 100%). The multiple comparisons are reported within each of the three age groups show variable results.


Timely detection of clinical deterioration in neonates is a patient safety imperative. Early warning tools have the potential to add to clinical expertise, enabling timely detection and treatment. Evaluation of the clinical utility of the three ‘hypothetical’ tools and the single ‘actual’ method applied to guide escalation of care in this modest-sized, single-centre study shows opportunity to improve all methods. For example, the local method—which was in use at the time of the study—suggested escalation in two in five ‘control’ patients who remained in the maternity ward. Further, the high ‘appropriate’ escalation rate (92%) may reflect the clinical use of the SOT rather than an unbiased assessment of performance. The late, early preterm, term and post-term neonate groups demonstrated uneven results across all tools, and suggest potential advantages from age-matched identification methods.

The ideal screening tool is able to meaningfully separate patients at risk from those not at risk, and help guide clinical decisions about care escalation. Trade-offs between detecting all cases (sensitivity) and the number of false-positive events have important implications for day-to-day use by front-line clinicians. Administrators may wish to consider workload, rates of false reassurance and the upstream impact of screening, beyond the apparent favourable ‘optics’ of having an Early Warning System. Inclusion of gestational age is one area where future versions of these ‘age’ agnostic neonatal scores could be improved.

These early findings are important and directive, and will permit greater consideration of workload, alarm fatigue and the broader clinical utility of these clinical tools before larger scale implementations. Inclusion of term neonates may aid in our understanding of the typical population in maternity nurseries, as will inclusion of greater numbers of patients in each category. This is important initial work that will guide future research, and help inform clinical care and administrative decisions about risks and potential benefits of implementing techniques to help the early identification of neonatal risk.


  • Competing interests CP has shares in a clinical decision support company, and is a named inventor of the Bedside Paediatric Early Warning System—a documentation-based system of care for hospitalised children.

  • Provenance and peer review Commissioned; internally peer reviewed.