If you are not familiar with the term Antimicrobial Resistance (or AMR, for short), you should be – but that is not on you as I demonstrate below. You should know this term very well because you are either a contributor to the problem, or will potentially be affected by it, or both–with the likelihood increasing if we do not address it swiftly and comprehensively. 

Antimicrobial resistance (AMR) is when microbes like bacteria, viruses, fungi or parasites develop resistance to known medications from antibiotics, antiviruses, antifungals, and antiparasitics (like antimalarials). This would make common infections hard to treat, and also increase the risk of spreading diseases, leading to severe illness and death. 

Developing resistance happens as a natural process, but the problem arises when resistance happens at a much quicker pace than scientists’ ability to develop new antimicrobials to replace the ones that no longer work. This process  of developing, testing, and getting approval for public use takes 10-15 years on average. 

This acceleration in resistance happens as a result of the overuse and misuse of antimicrobials in humans, animals and plants. 

For more context, it is estimated that in 2019, 4.95 million deaths were associated with bacterial AMR, and  approximately 1.27 million of those deaths were directly attributed to bacterial AMR. It is forecasted that from 2025 to 2050, there will be 169 million deaths associated with AMR, out of which 39 million will be directly attributed to it. When it comes to economic costs, a study by the Center for Global Development modelled the global economic impact of AMR in humans showing that global GDP may decline by USD 1.67 trillion by 2050 with accelerated resistance. The highest burden of AMR is borne by low and middle Income countries.

Because of all of the above, I am leading an emerging research agenda at Busara that is focused on AMR in humans, animals, plants and the environment from a behavioral, social and economic angle. To do so, we are joining forces with researchers and policymakers in this space to produce and share knowledge and insights, test, and embed interventions addressing this issue.

I was, hence, keen and had the privilege of being present at the 2026 AMR summit held in February in Sydney, Australia. The summit was so powerfully organized by the Fleming Initiative and Australia’s National Science Agency, the Commonwealth Scientific and Industrial Research Organisation (CSIRO) and attended by representatives of life and social sciences, high, middle and low-income countries, public and private sectors, civil society, researchers and research organizations, as well as politicians and policymakers. I share below a summary of my key insights from the summit, each of which could easily be adopted and turned into a project for research to inform policy and guide action and practice.

“Deadly infections [that used to be treatable, but are now] with no cure”

How many of us or of the people we meet day-to-day actually ever heard of the term antimicrobial resistance? How many heard of it and knew they had an active role to play in combatting that challenge? From my interactions, I would say very few know what AMR is outside of health-related fields, and sometimes not even the clinically trained get it completely right–which is something we learned from our AMR field work with pharmacies in Kenya. This first reflection from the summit is about framing, communication, and the frequency of communication.

As Professor Mark Davis from Monash University put it, when communicating about AMR, we should avoid the ‘discovery narrative’, which means avoiding talking about this as a purely scientific problem, which basically leads to taking people’s agency away. The public, hearing about AMR under very technical scientific framing, is likely to think: ‘this is probably a problem for science to figure out a solution for’, which completely erodes any sense of ownership they could have towards the problem. As Jane Halton, Chair of the Coalition for Epidemic Preparedness Innovations (CEPI), mentioned, we should find a language that is inclusive of everybody. The language should be simple, accessible and intend to inspire ownership and accountability. We can simply define AMR, as the panel suggested, ‘Deadly infections [that used to be treatable, but are now] with no cure.’

One example of a failure in communicating the extent and prevalence of AMR is that we are never explicitly told in hospitals when resistance occurs and is the cause of death. For instance, it is very common to say that a cancer patient died because of cancer, but do we know that cancer patients are one of the most vulnerable populations to resistance due to their compromised immunity and that some of the deaths that happen are actually caused by failing to treat an infection they acquired? To that end, it was encouraging to hear from Dame Sally Davis, UK Special Envoy on AMR, that in England and Wales, death certificates are now being issued to reflect resistance as the cause of death, when it is.Finally, embedding the discussion of AMR within broader already-established frameworks is key such as: Health security, universal and primacy health care, food security. For example, It is estimated that by 2050 livestock will experience a loss between 8 to 11% because of AMR, with low-income countries bearing the highest loss. Biosecurity threats in low and middle-income countries where animals and humans interact with little biosecurity measures in place, can experience a rapid spread of infections by resistant bugs from livestock to humans. Doing this would allow us to talk about the issue more, while reinforcing how it is relevant and connected to existing efforts.

Start from where the context you are operating in is

I believe that one of the successes of the AMR2026 summit was its ability to bring strategy and policy voices together, along with voices from on-the-ground research and practice. It was also able to bring together voices from both; the global north and global south together in the same room. Nevertheless, one thing I would love to see in future summits and conferences discussing AMR is a more balanced representation between the different categories, with a special focus on a usually underrepresented category, especially in the global south, that uses on-the-ground insights to develop programs and interventions that help bridge the gap between and make progress towards strategy and policy aspirations.

There are multiple reasons for this recommendation. First and foremost, in the global sense of this public health challenge, the highest burden falls on LMICs (low and middle income countries). Global convenings are one of the few available spaces for LMICs to discuss their–sometimes unique–challenges, and exchange ideas with a diverse audience who cares about the global nature of AMR and seeks to coordinate efforts. On an even more focused level, involving organizations–from research and practice–who work on AMR on the ground in LMICs and have an understanding of the social, behavioral and economic dynamics and incentives at play in the different contexts, will bring practicality and realism to the discourse and showcase where the gaps lie. And in some cases, this is where the starting point should be, in order to be able to achieve the overarching National Action Plans (NAPs).

Erta Kalanxhi, Director of Partnerships at the One Health Trust, rightfully noted that while stewardship programs that aim to provide guidance regarding the correct use of antimicrobials are key, the reality in much of Sub-Saharan Africa, where the burden is among the highest, is that there is a  lack of access to, and availability of the required antimicrobials for a large portion of the population.This usually leads to further spreading of infections or using whatever is available, which creates a different problem. So the starting point here is completely different; the basics are not there. 

Relatedly, and therefore, National Action Plans (NAPs) and stewardship programs must reflect, and should not be separate from, the existing cultures, norms, beliefs, and contexts of the countries, if we ever hope to see any progress. In countries where there is availability of decent public healthcare services, for example, regulations can be enforced and can be effective. However, when there are many obstacles to accessing public healthcare services, such as long queues, unaffordable consultation fees or stock-outs, people resort to other channels to satisfy their healthcare needs. And so regulations become much less enforced in practice and much less effective as an intervention on its own. It, hence, becomes necessary and integral to understand the existing incentives, social norms and economic contributors in the communities in order to design the right and effective interventions. One great example of how deeply contextual stewardship should be, took place in Bangladesh and was shared by Senjuti Saha, Deputy Executive Director at Child Health Research Foundation. In Bangladesh, because of resistance, a second-line antibiotic had been in use and prescribed to patients. Because of stewardship measures, there was a noticeable reduction in use and the microbe started responding to first-line antibiotics. However, going back to using first-line antibiotics was not straightforward. First, doctors needed convincing, because we should remember that doctors are sometimes under immense pressure from patients’ treatment expectations. Then, patients needed convincing, because they got used to a certain medication being prescribed in certain conditions and now we are changing it. After surpassing those two layers, and in thinking that success was achieved, researchers discovered that when patients took the prescription to the pharmacies to get their first-line antibiotic, pharmacists would tell them ‘oh, that is not a good or effective antibiotic’, and they would give them the second-line antibiotic again. A case in point showing how important it is to deeply understand the context and how systems feed into each other.

Yes, AMR can be reversed or reduced–but we need data to track and report this

You read the example in Bangladesh correctly, AMR was reversed. But to know and track that, data on dispensing, use, and resistance–or what is called surveillance data–is needed.

Discussions on data in the summit covered many important dimensions. To start with, thoughtful and intentional effort needs to go into designing data collection  to become an automatic process that requires minimum effort, especially if it is being done by Community Health Workers (CHWs) who are already overwhelmed, over-burdened, and working in unpleasant conditions. If we add a demanding layer of tiresome tasks, we are just creating conditions for potential data fraud or, at best, bad non-reflective data quality.

In addition, data privacy and security must be preserved. In her discussion on the importance of thinking beyond data de-identification when addressing privacy and confidentiality for indigenous populations in Australia, Dr. Kalinda Griffiths, Director at Flinders University, shed light on an important perspective that bias can arise not just from identifiable individual research participants, but also against entire subpopulations or against minorities. This could potentially happen during the analysis or the interpretation of results. She also discussed, what I believe is, a critical framework to consider for data management, even beyond indigenous populations–and especially useful in the Global South. The framework introduces a set of principles that recognizes power differentials and historical context as we think about and push for data sharing. They are called the CARE Principles for Indigenous Data Governance.

Relatedly, it is important to ask: who gets to interpret and make sense of the data? Who makes decisions? What is the pathway from data collection, to analysis, to interpretation and to turning into policy and action? And how can we enable governments to actually make sense of and use surveillance data they collect and turn data into empowering decisions and policy-making tools? How can we reinforce messages that prevent surveillance data from being used for ‘moral policing’? The latter is especially important and happens, whether intentionally or unintentionally, when there is a sharp turn from no enforcement of regulations to sudden enforcement, disregarding the social and behavioral context, and forgetting that everyone is in fact doing the best they can given the constraints, pressures and the environment they are operating in.

Finally, AMR is a behavioral and social sciences problem as much as it is a life sciences one

One key reflection that has been confirmed and reinforced through the summit is that AMR cannot be tackled by life sciences alone–or by social (which includes economic) and behavioral sciences alone for that matter, but the former tends to be less obvious. Without being informed by social and behavioral insights, newly developed antimicrobials are still going to be misused and overused by humans, for human health as well as for animal and plant health. The uptake of diagnostics when they become available will still be low – and we know that from the uptake of existing diagnostics, whether offered at the pharmacy as the first point of care in low-income settings, or even at hospitals. Without using insights from behavioral science, people will continue being hesitant about vaccines, exhibit unhealthy waste disposal behavior, and engage in activities that make them more susceptible to infections. At the end of the day, understanding context means genuinely letting people take the lead in informing us researchers about their environments and the types of solutions and interventions they need and not the other way around. One of the points mentioned in the summit that has really stuck with me was that genuine co-design is not going with a ‘half-baked cake’ and seeking validation from the community to say ‘look at this cake that we baked together’. Thank you Brett Sutton AO, Director of Health and Biosecurity at CSIRO for framing this so nicely.