Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
The Evidence-Based Medicine Manifesto aptly identifies massive ‘too much’ problems: too much hidden information, too much bias, too much research waste and too much overtreatment.1 2 One potential solution is to focus efforts on a small number of needed treatments while allowing all the excess to fall by the wayside. Here I explore the potential role of the essential medicines list (EML) in meeting some of the big challenges posed in the Manifesto.
What is an essential medicines list?
EMLs usually include around 300 medicines that meet the priority needs of a population.3 4 The WHO created its Model EML in 1977, and since then more than 100 countries have adapted the WHO’s Model EML to their own circumstances.3 4 Today more than five billion people live in a country with an EML.
Ultimately we need to know which treatments should be used and EMLs are ‘positive lists’ that can help direct people to effective treatments (as opposed to ‘negative lists’ of medicines to avoid). Given all the clinical research that has been done over the decades, we ought to be able to write down a list of medicines that were proven to have important benefits without excessive harms in studies that were properly done and reported. The fact that it is difficult to do so represents a huge failure of the status quo. There is also a huge opportunity: many people interested in evidence-based medicine are not yet trying to create such a list and some do not even know about EMLs.
Short EMLs are an attractive solution to parts of the ‘too much’ problems. If we took the resources currently engaged in synthesising evidence about all of the tens of thousands of medicines in existence and deployed them to deciding which few hundred medicines should be on EMLs, we could build on existing EMLs and create a very strong list in short order. It would not happen overnight because some of the relevant evidence will be hidden in clinical study reports and some old evidence may be difficult to obtain and assess. But this would be a much more manageable task than trying to improve information transparency for all medicines. Despite important data transparency advances such as the restoring invisible and abandoned trials initiative, a large amount of information remains hidden.5 6 EMLs reduce the number of needles we need to find in the haystack.
Too many medicines
The huge financial incentives for ‘me too’ medicines that are similar to existing medicines have cluttered and biased the medical literature. A recent systematic review of antidepressant comparative effectiveness included 522 trials involving 116 477 participants.7 The most impressive finding was that when a medicine was novel, it appeared to be more effective than when the very same medicine was trialled against a newer one (difference 1.18 times, 95% credible interval 1.09–1.27), that is, a medicine appears more effective when it is new. This improbable finding is likely due to flawed primary studies. What is the way out of this mess? Based on past experiences, adding to the 522 completed head-to-head studies of antidepressants with new trials of the latest medicine will only make things worse. We could choose to learn from all this and try a different approach.
An organised approach using medicines on EMLs at our lodestar could help avoid research waste. If a medicine is working well in practice, there may not be a reason to study similar medicines; instead we can move on to studies that will substantially improve health. The vague hope that a new medicine will be better than the ones we have is usually misguided: systematic reviews show that within-class differences are small in many important therapeutic areas including cardiovascular and respiratory diseases.8 9 Just like for antidepressants, the apparent differences within a class are similar in size to the artefact caused by the ‘too much’ problems with evidence. A focus on EMLs helps to expose the futility of trialling nearly identical medicines such as citalopram and escitalopram against each other in a haphazard manner. Importantly, discipline will help us focus efforts where they are needed and allow us to recognise true innovations when they appear.
For example, fluoxetine is on the WHO’s Model EML. If we start by asking whether or not fluoxetine is serving its intended purpose and, then, if not, explore what better medicines (or treatments) are available. We could use real-world clinical data to determine if people taking fluoxetine are actually benefiting. If most people with anxiety disorder are substantially benefiting (as we anticipated based on clinical trials comparing fluoxetine with a placebo), and there are no substantial harms, we could stop there and leave fluoxetine on this list as the need may be met. But if fluoxetine is not working well for a large number of people, we should certainly look for better alternatives. This would involve carefully sifting through the clinical study reports and individual participant-level data for completed head-to-head trials involving fluoxetine and potentially conducting new studies. That would be a lot of work, but in the end we would be confident that we selected the best medicine.
A better future is possible
Imagine a world where both clinicians and patients knew they could rely on a publicly available EML to determine which medicine should be taken. An online resource could be searchable by either the medicine or the condition or symptom. If you typed in ‘overactive bladder’, you may get no results because there is no highly effective treatment; non-medical treatments should be used.10 The challenge today is that it is difficult for anyone to know which medicines work, and the vast array of medicines makes it seem like there is a pill for every problem—there may literally be more pills than medical problems today. Creating such a database seems impossible now because we are accustomed to being overwhelmed by a large number of medicines and a deluge of information. A focus on EMLs could mean less hidden important information, less bias, less waste and less overtreatment.
I thank Marcy McCall MacBain and Peter Doshi for comments on the drafts.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent Not required.
Provenance and peer review Not commissioned; internally peer reviewed.