Web site designed and maintained by Chris Cairns  © SICS EBM Group 2004                                  

The effect of MET systems on outcome in hospitalised patients

Citation/s: Merit Study Investigators. Introduction of the medical emergency team (MET) system: a cluster randomised controlled trial. Lancet. 2005. 365: 2091-7.

Lead author's name and fax: K Hillman K.Hillman@unsw.edu.au

Three-part Clinical Question: Does the introduction of the MET system improve the composite outcome of death, cardiac arrest and unplanned ICU admission in hospitalised patients.
 

Search Terms: Medical emergency teams, Outreach critical care, Outcome, Mortality

The Study: Single-blinded, cluster-randomised, controlled trial with intention-to-treat.

The Study Patients: All comers admitted to 23 Australian hospitals over a 6 months period being care for in ward environment (excluding ICU, theatres, ICU controlled HDU areas and A&E). Hospitals were excluded if a METs system was already in place. There was baseline data collection for 6 months prior to MET implementation.

Control group (N = 56,756; 56,756 analysed): Continuation of standard emergency team set up, i.e. cardiac arrest team.

Experimental group (N = 68,376; 68,376 analysed): Education programme (MET related) for 4 months (with the emphasis on recognition of the at-risk patient rather than their management) followed by the introduction of a MET for the 6 months of the trial. MET team consisted of a nurse and ICU or A+E doctor responding to a specific MET calling system.

The Evidence:

There was a significant reduction in adverse outcomes in the control hospitals over the study period. This was not seen in the intervention group.

Comments:

This is a well-designed study with adequate power to determine a difference in the power calculations primary outcome measure although the actual incidence of the primary outcome did turn out to be lower than expected. . Thus in retrospect, the study may have been underpowered for the primary outcome measure. The reduction in incidence of the primary outcome measure may relate to the Hawthorne effect. The study shows that despite an educational programme and introduction of MET system, mortality, cardiac arrest rates and unplanned admission to ICU were not reduced. It is only safe to assume that MET does not improve outcome in hospitalised patients.

EBM summary:

1) Do the methods allow accurate testing of the hypothesis? Yes they do. A cluster randomised trial methods is the highest method available for testing such a hypothesis. The power calculations and statistical tests seem appropriate. Effort has been made to randomise centres for baseline characteristics equally between groups.

2) Do the statistical tests correctly test the results to allow differentiation of statistically significant results? Yes they do. There is discussion about the low incidence of the primary outcome in both groups. This means that they powered on a far more common endpoint; this affects the power. They also comment on the reduction in cardiac arrests  in the control group after the implementation period. Interestingly this effect was not seen in the MET group.

3) Are conclusions valid in light of the results? They  concluded a number of reasons why this trial was negative. They  appropriately stated that the first (and most likely reason) was lack of effect of the intervention. They have also questioned (a) the power and (b) the presence of the Hawthorne effect in the control group. It was also very interesting to note that only around 34% of ICU admissions were seen by the team before admission and of these only 30% triggered the MET system for longer than 15 minutes before the event. This suggests that there may have been failure to recognise the sick patient rather than purely a lack of effect of the MET teams intervention per se. The MET relies on early recognition and calling. 

4) Did results get omitted, and why? They did not.

5) Did they suggest areas of further research? They state “In view of the overwhelming evidence that many seriously ill patients receive inadequate care in hospitals worldwide, further research should be undertaken and our results can assist the design and management of other such studies”. This seems appropriate. They do not suggest further trials of MET.

6) Did they make any recommendations based on the results and were they appropriate? They suggest that improved monitoring, documentation and response to change as well as education in critical illness is required due to the large number of missed patients and poor documentation. They state that we require “frequent and rigorous documentation of patients’ condition; and increased attention to education to ensure a timely response by appropriately trained clinicians”. They do not comment on whether MET teams should be dismantled in the hospitals that have already implemented this system. Perhaps any educational package should include ‘initial management of the at-risk patient’ as well as ‘recognition of the at risk patient’.

7) Is the study relevant to my clinical practice? Yes it is. Although MET is not identical to the Outreach system widely adopted in the UK, it is more developed and has a medical lead. It may be reasonable to assume then  that Outreach will not improve outcome. We should re-evaluate the role of Outreach in the management of the deteriorating hospital patient. The challenge is to establish a robust system which both identifies the sick patient early and leads to prompt, appropriate treatment.

8) What level of evidence does this study represent? Level 1⁺⁺

9) What grade of recommendation can I make on this result alone? Grade A.

10) What grade of recommendation can I make when this study is considered along with other available evidence?- Grade A

11) Should I change my practice because of these results? If you have an Outreach team or early warming system (EWS), then you should re-evaluate their role.

12) Should I audit my current practice because of these results? If you have an Outreach team or EWS system you should at least rigorously audit their activity and outcomes.

Appraised by: Brian Cuthbertson Intensive Care Unit Aberdeen Royal Infirmary; 25 July 2005

Email: b.h.cuthbertson@abdn.ac.uk

Kill or Update By: July 2008

Edited by RJP and CC

Printer Friendly Version

©SICS EBMG 2005