Skip to content
View commentaries on this research

Please note that this summary was posted more than 5 years ago. More recent research findings may have been published.

This is a plain English summary of an original research article. The views expressed are those of the author(s) and reviewer(s) at the time of publication.

This systematic review and expert consensus process found 15 non-surgical interventions affecting adult mortality in critical care. Seven interventions, such as the use of tranexamic acid after severe blood loss, reduced deaths. Eight interventions, such as improving oxygen supply by using a drug, dobutamine, increased them. 2015 UK guidelines only partly reflect these findings. They may need amending to more explicitly identify interventions considered beneficial and harmful in critical care. This interesting study looked at a number of interventions in a particular clinical setting and asked the views of over 500 experts.

The review also found clinicians around the world were not necessarily acting in accordance with the evidence, highlighting an important gap between the evidence and practice. The NIHR Critical Care Clinical Research Network is actively promoting development and use of high quality research.

Why was this study needed?

Critical care is resource intensive and critically ill patients have high mortality rates or high risk of poor long-term outcomes. Research into interventions affecting mortality could save many lives and reduce the high costs of critical care if implemented effectively. However, the extent to which evidence was informing the care of critically ill people was not clear. This research aimed to see whether clinical experts were using evidence on effective and ineffective interventions to inform what they did in clinical practice.

Before this study there was a lack of research on whether critical care trials showing benefits from interventions are designed and implemented differently from trials showing harm. This knowledge would help show how research uptake should change.

What did this study do?

This mixed-methods study started with a systematic literature review of randomised controlled trials (RCTs) on mortality in critical care. Next, the researchers hosted a meeting of experts to prioritise the best quality studies on the most relevant interventions. Finally, an online survey of 555 experts from 61 countries examined their agreement with the systematic review findings. The survey also asked how experts would use the research findings on beneficial and harmful interventions in practice.

The study design contained many aspects that made it reliable, such as including only multi-centre RCTs. However, there may be more recent trials and the online survey of experts may not be representative of all critical care health professionals. Also, the number needed to treat and number needed to harm statistics should be treated with caution as they were pooled across trials.

What did it find?

Based on 24 multi-centre RCTs:

  • Seven interventions decreased mortality. These included: non-invasive positive pressure ventilation, compared with an oxygen mask, for people with chronic obstructive pulmonary disease in acute respiratory failure; lying chest down for people with severe acute respiratory distress syndrome; active cooling of a person (mild hypothermia) after a heart attack; and treating severe blood loss with the drug tranexamic acid. The average (median) number needed to treat in these trials was 7 (interquartile range, 5-8).
  • Eight interventions increased mortality. These included using a haemoglobin-based blood substitute for people with severe, trauma related blood loss, and for critically ill people administering growth hormone or raising oxygen supply to organs by using dobutamine. The average (median) number needed to harm in these trials was 9 (interquartile range, 5-16).
  • The average size of the RCTs was small, involving few centres. RCTs reporting harm were larger, involved more centres, and were more likely to contain blinding compared to those reporting benefits.
  • Experts' views varied on findings from the review and consensus meeting: 81% agreed on interventions that lowered mortality and 82% agreed on interventions that increased it. Of those agreeing with the findings, only 71% followed them in practice.

What does current guidance say on this issue?

The Intensive Care Society has published several guidelines, including the detailed 2015 Guidelines for the Provision of Intensive Care Services. Three of the beneficial interventions are recommended in these guidelines: non-invasive ventilation, lying prone, and ventilation using small volumes of air for acute respiratory distress syndrome. Of the harmful interventions, the guidelines say that two should be avoided or cause harm. These are: hydroxyethyl starch for severe infection, and rapid frequency ventilation. The other six are not mentioned in the recommendations.

Another of the beneficial interventions, mild hypothermia after heart attack, has been recommended by NICE since 2011. However, the authors of this study say there is new evidence that questions the benefit of this intervention.

What are the implications?

This study's findings suggest that 2015 UK guidelines may need amending to reflect this clinical evidence and expert consensus specifying interventions considered beneficial and harmful in critical care.

The finding that clinicians around the world are not necessarily acting in accordance with the evidence highlights the important gap between the evidence and practice. It suggests that critical care practitioners need to engage with the evidence base, and incorporate that knowledge into their practice. In the UK, the NIHR Critical Care Clinical Research Network is actively promoting the development and use of high quality research.


Landoni G, Comis M, Conte M, et al. Mortality in multicenter critical care trials: an analysis of interventions with a significant effect. Critical Care Medicine. 2015.


Clinical Research Network. Critical Care [homepage on the internet]. London: National Institute for Health Reseach; 2015

Institute for Healthcare Improvement. Evidence based care bundles. Cambridge (MA): Institute for Healthcare Improvement; 2015.

NICE. Therapeutic hypothermia following cardiac arrest. Interventional procedure guidance 386. London: National Institute for Health and Care Excellence; 2011.

Tidy C. Numbers needed to treat. PatientPlus [internet]. London: EMIS Group; 2012.

Produced by the University of Southampton and Bazian on behalf of NIHR through the NIHR Dissemination Centre


NIHR Evidence is covered by the creative commons, CC-BY licence. Written content may be freely reproduced provided that suitable acknowledgement is made. Note, this license excludes comments made by third parties, audiovisual content, and linked content on other websites.

  • Share via:
  • Print article


Number needed to treat (NNT) is the number of patients who need to be treated to prevent one additional bad outcome, relative to a control group.

Number needed to harm (NNH) is the number of people who are exposed to a harmful intervention before one person suffers harm, relative to a control group.

Pooled NNTs (or NNHs) derived from meta-analyses can be misleading because the baseline risk of an event often varies appreciably between the different trials. This explains why the study authors have only reported the average (median) results across the included trials.


Back to top