Intended for healthcare professionals


Decolonising knowledge production: our experience researching tranexamic acid for trauma victims

BMJ 2022; 377 doi: (Published 27 May 2022) Cite this as: BMJ 2022;377:o1338
  1. Ian Roberts, professor of epidemiology and public health1,
  2. Haleema Shakur-Still, professor of global health trials1
  1. 1London School of Hygiene and Tropical Medicine

In 2010, the CRASH-2 collaboration, after working for five years, on four continents, completed the largest ever randomised trial in trauma and found the first lifesaving treatment—tranexamic acid (TXA).1 The research team were elated.

Eight weeks later our paper was published in the Lancet. All team members were named. Together we had achieved what none of us could have achieved alone. Our research found that TXA use could save over 100 000 lives each year.2 In 2011, the CRASH-2 trial was named Research Paper of the Year in the BMJ Group Awards. In 2012, TXA was added to the WHO list of essential medicines. We were well on our way to improving trauma care worldwide, but then something strange happened.

Some academics started to question the quality of care in the participating hospitals in our research study. Australian authors wrote: “because substantial differences are likely between advanced and less developed trauma systems, hypotheses about TXA should be reinvestigated.”3 TXA may work in poor countries but perhaps not in Australia, they suggested. The charges levelled against the CRASH-2 trial were vague: “Only 51% of patients were transfused”…”the mortality rate was 16%, which is higher than that in advanced trauma centres in Australia.”3 Baseline mortality and transfusion rates do vary by country, but the mortality reduction with TXA does not, so why does this matter?

We thought we knew the rules of evidence and could tell a good trial from a bad one. But as a dusting of disparagement settled on the CRASH-2 trial, we started to make out a second quality criterion—one based on prejudice. A 2017 randomised trial found that all other things being equal, changing the location of a study from a high to a low-income country, say from Harvard to the University of Mzuzu, led to a decrease in the perceived strength of the evidence and the chances that it would be endorsed.4 Another study using the Computer-based Implicit Association Test, a method devised by psychologists to measure attitudes that people might be unwilling to report, found that doctors strongly associated rich countries with good research, and poor countries with bad research.5 But sophisticated methods to root out subconscious bias were not needed in the case of the CRASH-2 trial. Writing in BMJ Open, Australian authors arguing for a new trial of TXA in hospitals with “modern” trauma care point out that “Almost all patients in CRASH-2 were in low-income and middle-income countries where prehospital care was limited, blood components were uncommonly used, and where injury mortality was high. Seventy four per cent of the CRASH-2 patients were enrolled in Columbia (sic), Ecuador, Georgia, Nigeria, Egypt and India; only 340 (1.7%) patients were from Australia, New Zealand, the USA, Canada, Western Europe or the UK, where trauma system improvements have greatly reduced injury mortality and improved functional recovery.”6

Why is it acceptable to criticise the health systems of non-western countries, disparage their contribution to knowledge, and deprive patients of effective treatments? Would it have been as easy to disparage the CRASH-2 trial had all of the researchers been white? Emilie Koum Besson argues that “within academic institutions, whiteness has asserted its normalcy through the systematic marking of other groups” “differences” and depictions of cultural superiority of western societies over others.”7 Decolonising global health will only be possible she writes, when “everyone—especially white people—can “free” ourselves from the idea of the innate superiority of western culture and research paradigms.”


  • Competing interests: none declared

  • Provenance and peer review: not commissioned, not peer reviewed.


View Abstract