Elsevier

Resuscitation

Volume 73, Issue 1, April 2007, Pages 62-72
Resuscitation

Clinical paper
The objective medical emergency team activation criteria: A case–control study

https://doi.org/10.1016/j.resuscitation.2006.08.020Get rights and content

Summary

Objective

To evaluate the ability of pre-defined clinical criteria to identify patients who subsequently suffer cardiac arrest, unplanned intensive care unit admission or unexpected death; to determine the ability of modified criteria to identify these patients.

Design

Nested, matched case–control study.

Setting

Seven Australian public hospitals.

Patients and participants

Four hundred and fifty cases and 520 controls matched for age, sex, hospital, and hospital ward.

Interventions

None.

Measurements and results

Highest and lowest respiratory and heart rates, lowest systolic blood pressure, presence of threatened airway, seizures or decrease in Glasgow Coma Scale score of greater than two points and incidence of the three adverse events were measured. Combining a heart rate greater than 140, respiratory rate greater than 36, a systolic blood pressure less than 90 mmHg and a greater than two point reduction in the Glasgow Coma Scale identified adverse events with a sensitivity of 49.1% (44.4–53.8%), specificity of 93.7% (91.2–95.6%), and positive predictive value of 9.8% (8.7–11.1%). Adding threatened airway, seizures, low respiratory rate and low heart rate did not substantially improve sensitivity (50.4%; 45.7–55.2%). After modifying the cut-off values for respiratory rate, heart rate and systolic blood pressure, the best achievable positive predictive value remained below 16%.

Conclusions

In combination, the respiratory rate, heart rate, systolic blood pressure, and level of consciousness identify patients at risk of cardiac arrest, unplanned intensive care admission or unexpected death with high specificity; however the sensitivity and positive predictive value are relatively low, even after modification of the activation criteria cut-off values.

Introduction

Studies from around the world have consistently shown that hospital patients suffer serious and preventable adverse events that may result in prolonged hospitalisation, permanent injury and death.1, 2, 3, 4 Of these events, cardiac arrests, unplanned intensive care unit (ICU) admissions and unexpected deaths are three of the most serious.1, 2, 5, 6, 7 Survival from in-hospital cardiac arrest remains poor despite decades of attention and research,8, 9 and patients suffering in-hospital cardiac arrest in a non-critical care area are even more likely to die.7 Patients admitted to the ICU from the wards have a higher severity of illness and mortality than similar patients admitted from the emergency department or operating theatres.10, 11, 12 Although the circumstances leading up to these events are complex, the realisation that routinely documented clinical observations may serve as predictors for up to 24 h prior to the actual event,13 has lead to the development of “early warning systems” in various countries.14, 15, 16, 17 The first of these, and the most rigorously evaluated to date, is the Medical Emergency Team (MET) system.14, 18

The MET activation criteria, used to detect patients at risk of serious adverse events and to trigger activation of the MET, have been used without formal evaluation since 1990.14 Although clinical observations such as tachypnoea, decreased level of consciousness and systolic hypotension have been shown to be associated with in-hospital cardiac arrest and death,19, 20, 21, 22, 23 different combinations of criteria are in use, and there is no consensus on the ideal set of criteria or their optimal cut-off values.24, 25, 26, 27, 28

Many health-care systems are devoting significant resources to implement early warning and MET-like systems. Given the central role of the activation criteria within these systems, it is imperative that the activation criteria are as accurate as possible.

We conducted a nested, matched case–control study in order to evaluate the ability of published objective clinical MET activation criteria (Table 1) to identify patients at risk of cardiac arrest, unplanned ICU admission or unexpected death, and to determine whether modifying their cut-off values would improve their sensitivity, specificity and predictive value.

Section snippets

Materials and methods

The eleven hospitals from the MERIT study that were not allocated a MET system were invited to take part in this study.18 Of these, seven hospitals (four university affiliated, three community) participated. Approval was obtained from the human research ethics committee of each hospital.

The adverse events examined by this study were unexpected cardiac arrests, unplanned ICU admissions and unexpected deaths. A cardiac arrest was defined as occurring when a patient lacked a palpable pulse. An

Statistics

Where observations were missing, it was assumed that these observations would not have resulted in activation of the MET system. Univariate odds ratios for the occurrence of an adverse event were obtained for each of the activation criteria using conditional logistic regression adjusted for clustering by hospital, using the Huber/White/sandwich method of adjustment of variance estimates.29

We then assessed the diagnostic accuracy of each established objective MET activation criterion

Results

A total of 503 eligible index cases were identified. Of these, 53 could not be matched and were excluded. The remaining 450 cases (89.5% of the eligible case population) and their 520 matched controls were included. The case and control numbers by hospital are presented in Table 2. After exclusion of admissions of less than 24 h, the incidence of all three serious adverse events was calculated as 13.9 per 1000 admissions per year (95% CI 13.5–14.3/1000 admissions per year). The incidence of

Discussion

We conducted a case–control study to examine the ability of a pre-defined set of objective clinical observations to identify patients at risk of cardiac arrests, unexpected deaths and unplanned ICU admissions. We found that when a respiratory rate greater than 36 breaths/min, heart rate greater than 140 beats/min, systolic blood pressure less than 90 mmHg and decrease in the GCS score of more than two points were combined, they predicted a serious event within 24 h with a high specificity (93.7%,

Conclusion

This study confirms that in combination, a high heart rate and respiratory rate, low systolic blood pressure and a decrease in the GCS score are specific predictors of cardiac arrest, unplanned ICU admission and unexpected death. The set of objective criteria used during the MERIT study had relatively low sensitivity and positive predictive values for the adverse events studied. Even after modification, the best achievable positive predictive value was 15.7%, with a corresponding sensitivity of

Conflict of Interest Statement

All authors declare that they have no personal or financial relationships that could inappropriately influence their work.

Acknowledgements

The MERIT study is a collaboration of the Simpson Centre for Health Services Research and the ANZICS Clinical Trials Group. We thank Dr. Michael Parr for reviewing the manuscript.

The study was funded by grants from the National Health and Medical Research Council of Australia, the Australian Council for Safety and Quality in Health Care and the Australian and New Zealand Intensive Care Foundation as part of the MERIT study.

References (35)

  • R.M. Wilson et al.

    The quality in Australian Health Care Study

    Med J Aust

    (1995)
  • L.L. Leape et al.

    Five years after to Err is human: what have we learned?

    JAMA

    (2005)
  • G.R. Baker et al.

    The Canadian Adverse Events Study: the incidence of adverse events among hospital patients in Canada

    CMAJ

    (2004)
  • Institute of Medicine. To Err is human: building a safer health system: Committee on Quality of Health Care in America....
  • H. McGloin et al.

    Unexpected deaths and referrals to intensive care of patients on general wards. Are some cases potentially avoidable?

    J R Coll Phys Lond

    (1999)
  • D.R. Goldhill et al.

    Outcome of intensive care patients in a group of British intensive care units

    Crit Care Med

    (1998)
  • M.D. Buist et al.

    Recognising clinical instability in hospital patients before cardiac arrest or unplanned admission to intensive care: a pilot study in a tertiary hospital

    Med J Aust

    (1999)
  • Cited by (198)

    View all citing articles on Scopus

    A Spanish translated version of the summary of this article appears as Appendix in the final online version at 10.1016/j.resuscitation.2006.08.020.

    1

    See Appendix A.

    View full text