Elsevier

Resuscitation

Volume 71, Issue 3, December 2006, Pages 327-334
Resuscitation

Clinical paper
Combinations of early signs of critical illness predict in-hospital death—The SOCCER Study (signs of critical conditions and emergency responses)

https://doi.org/10.1016/j.resuscitation.2006.05.008Get rights and content

Summary

Background

Medical emergency team (MET) call criteria are late signs of a deteriorating clinical condition. Some early signs predict in-hospital death but have a high prevalence so their use as single sign call criteria could be wasteful of resources. This study searched a large database to explore the association of combinations of recordings of early signs (ES), or early with late signs (LS) with in-hospital death.

Methods

A cross-sectional survey was undertaken of 3046 non-do not attempt resuscitation adult admissions in 5 hospitals without MET over 14 days. The medical records were reviewed for recordings of 26 ES and 21 LS and in-hospital death. Combinations of ES with or without LS were examined as predictors of death. Global modified early warning scores (GMEWS) were calculated.

Findings

ES with LS, plus LS only, had higher odd ratios than ES alone. Four combinations of ES were strongly associated with death: cardiovascular plus respiratory with decrease in urinary output, cardiovascular plus respiratory with a decrease in consciousness, respiratory with decrease in urinary output, and cardiovascular plus respiratory. In other combinations, recordings of SpO2 90–95%, systolic blood pressure 80–100 mmHg or decrease in urinary output in turn occurring with one or more disturbed blood gas variable were associated with death. Compared with admissions whose GMEWS were 0–2, admissions with GMEWS 5–15 were 27.1 times more likely to die while those with GMEWS 3–4 were 6.5 times more likely.

Conclusions

The results support the inclusion of early signs of a deteriorating clinical condition in sets of call criteria.

Introduction

Widespread interest continues in developing the optimal model of a clinical emergency response system (CERS) in hospitals to provide assistance to general ward staff in the recognition and management of patient with a deteriorating clinical condition.1, 2 Explicit call criteria empower ward staff to call more experienced personnel to contribute to the identification and management of patients at risk of unexpected death or cardiac arrest. The characteristics of the responding team or teams vary in the various models so far described in the literature. Also there is as yet no agreement on the most appropriate call criteria.3 The commonest are the medical emergency team (MET) criteria first described in Australia4 and the early warning score (EWS)5 or modified early warning score (MEWS)6 first described in the United Kingdom. The MET call criteria were recently validated as predictors of serious adverse events by analysis of the SOCCER database.7 The SOCCER database was derived from a cross-sectional survey examining the case notes of 3160 adult admissions in five Sydney metropolitan hospitals for a selected range of abnormal physiology or potential calling signs over a 2-week period. The aim was to establish the prevalence and the predictive value for serious adverse events. The MEWS criteria have been validated also.8 Either MET or MEWS is likely to be a part of most protocol driven CERS for some time.9 However, in a recent multicentre prospective randomised cluster trial no difference was found in death rates between hospitals with a MET using the MET criteria and those with no MET but with a cardiac arrest team.10 In another study,11 the introduction of MEWS into a CERS did not change outcomes. There are many variables in CERS that influence outcomes but the results from these two studies suggest that it is reasonable to assume that the optimal set of call criteria has yet to be established and validated in prospective studies.

The MET call criteria can be looked upon as signs occurring late in a patient's deteriorating clinical condition but the SOCCER data suggest that some early signs of critical illness are also predictive of serious adverse events.7 The possibility exists that early, rather than late, intervention may avoid or at least reduce the rate of such events. However, the SOCCER Study has also shown a high prevalence of such early signs.12 So in developing the optimal set of call criteria for a practical CERS, a balance must be struck between prognostic value for unexpected death or cardiac arrest, potential reversibility and prevalence of the signs. To investigate sets of potential call criteria, the SOCCER database was interrogated for the association of recordings of signs from groups of early signs or early plus late signs with unexpected in-hospital death. It was not possible to study the association with cardiac arrest because of the low number of unexpected arrests (five) in the study period.

Section snippets

Methods

Methods used in the SOCCER Study have been described in detail previously.12 A cross-sectional survey of case notes of general ward admissions in five Sydney hospitals over a 14-day period was conducted for recordings of 26 early signs (or symptoms) (ES) and 21 late signs (LS) (or symptoms) of critical conditions. For brevity, the signs and symptoms are called “signs” throughout the rest of this paper. The signs had been selected previously by a panel of senior emergency and intensive care

Results

From the total of 3160 admissions (4 additional admissions were excluded because of incomplete notes), there were 3046 non-DNAR admissions (96.4%) with 27 (0.9%) deaths in the non-DNAR group. Of the 3046, 1407 (46.2%) had no recordings of the selected ES or LS. Of the remaining 1639 non-DNAR admissions, 1216 (74.2%) had ES only, 409 (25.0%) had ES and LS, and only 14 (0.9%) had LS without ES.

When ES were grouped with equivalent LS (Table 1), overall late signs with or without early signs had

Discussion

Our study has a number of limitations that are characteristic of retrospective searches of case notes for recordings of clinical signs. Nursing and medical staff are unreliable in the recording of signs and are more likely to record mandatory observations than the non-mandatory observations. However as members of the staff are unlikely to fabricate observations, those recordings in the notes can be taken to represent the minimum prevalence of signs. We attempted to further reduce the impact of

Conflict of interest statement

None declared. There was no conflict of interest for any of the authors.

Acknowledgements

We acknowledge and thank the intensive care and emergency medical and nursing clinicians from the five participating hospitals who made an important contribution in the study development phase. This study was funded in part by The South East Health Service of Sydney, NSW. This provided part payment for the data collectors. The health service administration played no part in the study design, data collection, analysis or interpretation or the writing of this report.

References (14)

There are more references available in the full text version of this article.

Cited by (0)

A Spanish translated version of the summary of this article appears as Appendix in the online version at 10.1016/j.resuscitation.2006.05.008.

View full text