Finding Clarity in Complexity

Can you tell us about your background and what inspired you to pursue research in AMR? 

My background is in pharmacometrics and clinical pharmacy. Before moving fully into pharmacometrics research, I worked closely with clinical trials, including phase II-IV studies in infectious diseases/nephrology/haematology, which gave me a practical understanding of how drug development decisions are made in real clinical settings.

What inspired me to pursue AMR research is the urgency of the problem. Antibiotics are fundamental to human health, but resistance is reducing their effectiveness and put the gains of modern medicine at risk. I was drawn to AMR because pharmacometrics can make a direct contribution to antibiotics development: by helping us understand how antibiotic exposure, bacterial response, and resistance interact, we can design better dosing strategies and support more optimized antibiotic use.

What specific aspect of AMR does your research focus on? 

My current research at Uppsala University focuses on translational antibiotic PK/PD modelling. More specifically, I work on understanding how antibiotic exposure relates to bacterial killing and treatment response using preclinical infection models.

Within the COMBINE initiative, my work evaluates meropenem and levofloxacin using the standardized COMBINE pneumonia model (murine). I am particularly interested in how experimental factors or inter-laboratory variabilities can influence pharmacokinetics and exposure-response relationships, thereby affecting subsequent PK/PD target determination.

This topic is critical because preclinical PK/PD targets are often used to support antibiotic dose selection and downstream translation into clinical development. If laboratory-specific differences are not appropriately recognised and quantified, we may over- or under-estimate the exposure needed for bacterial killing. By quantifying these sources of variability through modelling, my work aims to improve the robustness, reproducibility, and translational value of preclinical antibiotic PK/PD studies.

How do you see your research contributing to the fight against AMR? 

My research should contribute to the fight against AMR by improving how we translate preclinical antibiotic data into well-informed, strategic dosing decisions. If we can better understand the exposure needed for bacterial stasis, 1-log kill, or 2-log kill, we can design antibiotic regimens that are more likely to be effective while reducing unnecessary exposure.

This is important for AMR because inappropriate dosing can contribute to treatment failure and resistance selection. By developing and applying model-informed approaches, we can support more precise antibiotic development and help to optimize the use of both existing and new antibiotics.

Pharmacometrics adds value by turning complex experimental and clinical data into quantitative models that facilitate decision-making. PK/PD models, when well-developed, allow us to quantitatively describe changes in antibiotic concentration, exposure-response relationships, and account for variabilities that are most likely to affect bacterial killing. We can also simulate untested dosing scenarios without needing to conduct more animal testing. In AMR research, this helps to connect laboratory findings to clinically relevant dosing strategies and streamlines development of antibiotics.

What are some of the biggest challenges in this kind of research? 

One major challenge is unexplained data variability. Even when experimental protocols are standardized, differences between laboratories, bacterial isolates, infection models, and measurement conditions can still influence PK/PD results.

Another challenge is translation. Preclinical models are essential in antibiotic development, but they are simplifications of human infections. The difficult question is how we can infer results obtained from these animal and associated PK/PD models into humans in a way that is scientifically rigorous and clinically meaningful.

A third challenge is more pharmacometrics-oriented, and it is related to model identifiability. In some of the semi-mechanistic PK/PD models that we use, many biological processes may be represented mathematically but not all parameters can be estimated well from the data. Balancing biological realism with model stability is always a key challenge for a pharmacometrician.

How do you tackle problems when things don’t go as planned?

Rational thinking and compartmentalization are the key. I usually start by breaking the problem into smaller parts (or compartments) and try to resolve them separately. For example, if a model does not perform well, I will try to determine whether the issue originates from the data, the assumptions, the structural model, parameter identifiability, or the way model equations are being written.

I also try to keep the biological question in focus. A more complex model is not always better. Our goal is to develop a model that is useful, interpretable, and supported by the data.

There are also many excellent colleagues within Uppsala University and the COMBINE Initiative with whom I enjoy exchanging opinions and discussion. We often see the same problem from different angles, and that helps tremendously in my work.

How do you think AMR research will change over the next decade?

I think AMR research will become increasingly quantitative in the next decade. We will likely see more use of model-informed drug development, semi-mechanistic PK/PD models and perhaps complex quantitative system pharmacology models for decision-making.

I also expect to see more focus on combination therapy and individualized antibiotic dosing, especially in vulnerable populations such as critically ill patients, children, or patients with conditions that may alter drug pharmacokinetics (e.g. renal or kidney failure, or individuals with genetic polymorphisms).

Another important direction will be better integration of preclinical, clinical, and real-world data to better guide antibiotic development and stewardship with the increasing use of pharmacometrics modelling.

What developments in AMR research excite you most right now?

I am most excited about the growing use of mechanistic and translational modelling in antibiotic development. Traditional PK/PD indices such as fAUC/MIC or %fT>MIC are useful, but they do not always capture the full complexity of bacterial killing, regrowth, or resistance effects.

Mechanistic models can help us understand not only whether an antibiotic works, but how and why it works under different conditions. I am also excited about the potential integration of pharmacometrics with machine learning and real-world AMR surveillance data. The combination of these techniques could give researchers like us a powerful leverage in combating AMR.

What are your long-term goals in AMR research?

My long-term goal is to become an independent researcher in model-informed drug development. I aspire to develop translational modelling frameworks that can optimize antibiotic dose selection, combination therapy evaluation, and treatment strategies that concurrently consider the potential for AMR development.

I am especially interested in work that bridges pharmacometrics, clinical pharmacology and microbiology. Ultimately, I hope my research can contribute to more effective and sustainable antibiotic use.

How important is collaboration in your research? Can you share an example of a successful collaboration within the AMR Accelerator?

Collaboration is essential in my research. No single discipline can address AMR alone. Pharmacometricians need good-quality data, microbiologists provide critical biological interpretations, clinicians understand the broader context of patient treatments, and industry partners help to connect research questions to drug development needs.

A good example is my current work within the COMBINE initiative. The project evaluates the standardized murine infection model across different laboratories – in both academic and industry settings, and uses pharmacometrics modelling to evaluate antibiotic PK/PD relationships. This type of collaboration allows us to study not only antibiotic effects, but also how robust and reproducible these effects are across different experimental settings.

How do you stay motivated and passionate about your work?

For me, the motivation comes from knowing my work sits closely connected to real clinical problems. I have worked in both hospital setting and clinical trials, so I have seen how difficult it can be to move from a scientific idea to something that can genuinely help patients. That experience makes translational research feel practical to me. Ultimately, I see the purpose of our work as helping people make better decisions in the face of uncertainty.

I also remain passionate about AMR research because it constantly reminds me how extraordinary microbial evolution is. Bacterial response to treatments do not always behave the way we expect. Sometimes the data are messy, the model does not work, or the results may challenge my basic assumptions about how things should behave. But this is exactly what keeps the work interesting. Each problem forces me to think more carefully and learn from people with different expertise, and grow as a researcher.  

On a personal level, I enjoy the feeling of slowly turning a complicated problem into something understandable. When a model finally explains a pattern in the data, it feels genuinely rewarding — almost like completing a puzzle after working through many uncertain pieces. Those small moments of clarity are what keep me passionate about my work.

What advice would you give other researchers or students interested in AMR research?

My advice would be to build both biological understanding and quantitative skills. AMR research needs people who can understand the interaction between bacteria, antibiotics, and patients.

I would also encourage students to work across disciplines early. AMR is not only a microbiology problem, or a pharmacology problem. It is a systemic challenge and the most useful breakthrough often happens at the interface between fields.

And finally, I believe we need to be comfortable with uncertainty. Data in AMR research can be complex and most of them are imperfect. Careful modelling and critical thinking will help us go a long way in this field.

Can you describe a typical workday for you?

A typical workday is usually a balance between focused modelling work and discussion with colleagues. I may spend the morning working in NONMEM or R, checking model diagnostics or running stimulations, or trying to understand why a model is behaving in an unexpected way. Some days are very technical, other days are more about interpreting results and thinking about what the model is really telling me biologically or statistically.

One part of working in Uppsala that I really appreciate is the rhythm of work, especially the fika culture. It may sound simple but having coffee with colleagues create space for informal scientific conversations. Sometimes a useful idea comes naturally from a relaxed discussion over coffee, where someone from a different background sees the problem in a new way.

I also spend substantial time in writing manuscripts or scientific presentations. Communication is an important part of my job, my work is therefore not only about building models, but also about translating complex results into something useful and understandable for the wider research community.

How do you balance your professional and personal life? 

I try to maintain balance by being structured with my work and making time for activities outside research. Modelling work can be intense because it often involves long debugging and analysis cycles, so I think it is important to step away regularly. Short walks in the forest behind the office always do wonder to clear up my mind.

I also try to keep a long-term perspective. Research productivity is not about pushing for “just one more run”, and I certainly appreciate the rhythm of work in Sweden where there is more respect for personal time. Little things like fika remind me that productivity is not all about sitting in front of a screen. For me, balance means doing serious work in a way that is sustainable enough to keep enjoying the science over the long term.

Outside of research, fitness and strength training are an important part of my routine. It gives me structure, help me manage stress, and provides a very different kind of progress compared with research. In modelling work, I have to handle uncertainties; in training, the feedback is more physical and immediate, which I find grounding.

I also enjoy knitting classes with my friends. It is a nice contrast to the analytical side of my work because it is slower, more tactile, and creative. In a way, both training and knitting help me stay balanced: one keeps me physically active, and the other gives me space to slow down and enjoy the company of wonderful people.

What is the most important message you would want the public to understand about AMR? 

The most important message is that antibiotics are a shared and limited resource. AMR does not only affect people who misuse antibiotics, it affects everyone who may one day depend on them. Preserving antibiotic effectiveness is therefore a responsibility shared by researchers, healthcare professionals, policymakers, and most importantly, the people who rely on antibiotics to protect and restore their health. And everyone can contribute to the fight by using antibiotic responsibly. That means taking antibiotics only when prescribed, completing treatment as instructed, and do not share leftover antibiotics. On a broader context, we can also help by supporting vaccination, infection prevention, and practice good hygiene whenever possible. Reducing unnecessary infections reduces the need for antibiotics, which in turn helps to slow the development of AMR.

Want to learn more about the COMBINE Pneumonia Model and the bacterial strain repository we have set up for download through DSMZ? Check out the Tools & Resources section of the COMBINE project.