Today, evidence-based medicine is facing multiple serious challenges. Some claim that EBM has been hijacked by vested interests. We see an emphasis on surrogate outcomes with uncertain clinical value, on statistically significant results with marginal clinical benefit, and even outright manipulation of data. Publication bias and selective outcome reporting presents huge challenges and there is a clear tendency to focus on benefits, and disregard harms of interventions. Apart from economical incentives, the problems related to intellectual conflicts of interest are considerable, and the overwhelming amount of studies produced is another very difficult challenge.
It has been suggested that, after its initial success, EBM has become a movement in crisis. Measures to deal with publication bias and selective outcome reporting have been taken; pre-registration of protocols and a US law saying that all trials have to publish their results within 1 year after completion. Unfortunately, it is doubtful whether these initiatives have had the intended effect; studies reveal that unpublished trials and selective outcome reporting is still a problem. But perhaps more importantly, even if these initiatives will eventually work, publications are very brief summaries of trials. They do not contain enough information to allow re-analysis by independent researchers. For this, access to raw data is needed.
Cochrane researchers are among the leading forces in the struggle to make raw data from drug agencies available to independent researchers and used. The importance of this can be exemplified by the Cochrane review on neuraminidase inhibitors to reduce complications from influenza. When an update was extended to include re-analysis of raw data, the conclusion changed from suggesting a positive effect to showing no evidence of an effect on hospitalisation, serious complications or pneumonia. As another example, access to clinical study reports from the European Medicines Agency made it possible to find out that antidepressant drugs double the occurrence of aggression in children and adolescents. These examples raise doubts about the validity of much of the evidence base today. And they provide an insight that should revolutionise our methods for conducting systematic reviews.
Apart from economical incentives, the problems related to intellectual conflicts of interest in research, as well as guideline development, are considerable. Harms from interventions, opportunity costs and patient preferences are often not given due consideration. To maintain credibility, EBM as a movement needs to deal with this. What constitutes intellectual conflicts of interests is of course not clear-cut, but more or less based on judgement. An appealing way forward is to have methodologically qualified researchers from other, unrelated, research fields evaluate the evidence base in guideline development. Internationally, this approach is becoming more common, and Sweden should follow.
To counteract the many threats towards EBM today, it is vital to constantly challenge ourselves within the EBM movement. A constructive debate usually leads to progress. Recently, a Cochrane review on methylphenidate for adults with ADHD was withdrawn due to criticism of the methodological quality and too uncritical evaluation of the risk of bias in the included studies. This approach is essential; when flaws are revealed, we must drop our prestige, and we must constantly work towards making the changes needed.
In conclusion, EBM needs progression. And it will take a lot of effort and courage from the movement itself to ensure that EBM continues to improve healthcare and will become even more useful and relevant to healthcare providers and, of course, most importantly, to patients.
Minna Johansson, MD, Gothenburg, Sweden