Federica Santoro 0:15

More and more pathogens are becoming resistant to antimicrobials these days, and that places healthcare systems all around the world under lots of strain. This is not the beginning of a science fiction movie, but a very real and urgent public health threat. Antimicrobial resistance, or AMR, has been recognized for decades, but the scientific community is still struggling to find adequate solutions. Specialists often cite the need for multidisciplinary approaches to tackle this ever-growing problem. And so the obvious question is: can pharmacovigilance help? My name is Federica Santoro, and this is Drug Safety Matters, a podcast by Uppsala Monitoring Center, where we explore current issues in pharmacovigilance and patient safety. Joining me today are Jean-Marie Vianney Habarugira, Senior Project Officer at EDCTP, the European and Developing Countries' Clinical Trials Partnership, and Albert Figueras, Senior Consultant in Pharmacovigilance and Antimicrobial Consumption. Under Albert's supervision, Jeanne-Marie recently completed a doctoral project on the role of pharmacovigilance networks in AMR surveillance. And so we talked about what the drug safety community has to offer and how it can help inform better use of antimicrobials. I hope you enjoy our conversation. Hi Jeanne-Marie, hi Albert, and welcome to Drug Safety Matters. I love it when listeners approach us with ideas for the podcast, and that's exactly what happened here. Because you, Jeanne-Marie, contacted me earlier this year and said, Can I come on the pod and talk about pharmacovigilance and AMR? Which I thought was a fantastic idea. So obviously, we invited your PhD supervisor as well, Albert. And so here we are. But before we dive into the details of your research, what does AMR have to do with pharmacovigilance? Because the way I see it, AMR is about ineffective medicines. Whereas pharmacovigilance is about harmful medicines. Or is AMR a safety issue?

SPEAKER_00 2:45

Right. Maybe I should answer from the end. It's a yes. AMR is a safety issue. People insist on ineffectiveness, but maybe you are going to hear us talk a lot about inappropriate use. And under inappropriate use are a lot of other terminologies that will come: misuse, overuse, medication errors, but of course also patients not responding. So there's that bad outcome. So it's a safety issue. And pharmacovigilance collects data on safety issues.

Federica Santoro 3:16

So you both spent the last few years investigating the potential synergy between these two disciplines. So tell me, Albert, why do you think the pharmacovigilance network could help AMR surveillance?

SPEAKER_02 3:30

Well, really, I think that the pharmacovigilance network is perhaps the biggest and strongest network involving health professionals. Because, well, in addition to the 170 something countries that are part of the pharmacovigilance network, the WHO program, each country has a network with the health professionals at least working in this country. So if you look at the whole network, it's a very big and strong network. So when we thought that antimicrobial resistance is a real problem and that it's important to try to know where in the planet there are some potentially resistant bacteria, if you we have so a spectacular network, why not try to use that network, try to strengthen that network, and so try to detect suspicions of possible resistance. And this was the initial idea.

SPEAKER_00 4:54

If you look at uh developed countries, there are other methods which are sufficiently funded or where skills are enough to detect suspicions of resistance. But when you look at the the now fast developing pharmacovigilance capacity in developing countries, for them, this is an opportunity to use these new infrastructures, these new skills to document suspicions of resistance, of inappropriate use. If there is off-level use of antibiotics, if they can use that newborn uh kind of institution, the the pharmacovigilance centers to quantify that, it is maybe a quick or easy method for them to collect that information on suspicion while countries are still waiting for that other capacity which will take years maybe to be at a level where developed countries have it. For me, that's important to look at uh what we are proposing, this idea as something maybe meaningful for low and middle-income countries.

Federica Santoro 5:56

And let's talk about the practicalities of your proposal then. As in any kind of research, the way we collect data is crucial and will ultimately determine what we can do with it afterwards. So I was wondering how is AMR data captured in pharmacovigilance databases nowadays? And how would you like it to be structured instead?

SPEAKER_00 6:21

Right. So maybe I um I have to go back to uh 2017, I think when I read the UPSA report number 74. Someone decided to draft an article in that report suggesting that AMR is an overlooked adverse drug reaction. So I thought, why overlooked? Why is it overlooked and why is it an adverse drug reaction? Then the idea was let me look at how PV specialists are reporting ADRs and see where it is hidden, that overlooked adverse drug reaction that we call antimicrobial resistance. And I had to go back to what is known as the MEDRA codes, the medical dictionary for regulatory affairs calls, which actually are used by those reporting safety data across the pharmaceutical industry or those running clinical trials or those in routine uh clinical practice. But then those codes, if you go into those databases, it's a lot, it's thousands of them. I think the latest version is at 23,000 codes. So if if you think about an overlooked adverse drug reaction, I thought, okay, if it's overlooked, maybe it's hidden in those thousands of codes. And it took a while. Of course, I had Albert as my thesis director, so we worked a lot on refining this idea. But my first instinct was to dive into that 20-something codes, see if there is anything which means anything related to AMA. And what we have found, we managed to shortlist 17 Metro codes. And these are classified in four categories: there's resistance, there is ineffectiveness, there's off-level use, and there's then errors, medication errors. So each of these calls has a combination of words, whether it's English or Swedish or French or Spanish, they mean something if if you are using it to report an adverse drug reaction. But what is important in the world of pharmacovigilance is that these codes, these combination of numbers, they are universal. Whatever language you speak, if you want to report off-label use of an antibiotic, the same codes that is used in Brazil is the same code which will be used in China or in Sweden or in Spain or in the Netherlands. So that is what is overlooked, and what we'd like to see then not overlooked is to make sure everyone working in antimicrobial resistance or AMR surveillance is actually aware of these 17 codes. Then we can talk later about how they can be used.

Federica Santoro 8:59

So we've talked about data structuring and now moving on to the next step. Once the data has been collected, it has to be analysed. So what will AMR data analysis look like compared to so to say standard signal detection approaches in pharmacovigilance?

SPEAKER_02 9:18

Well, in my opinion, there is a crucial point here, and it's the fact that uh in pharmacovigilance we collect suspicions, suspicions of adverse reactions. And I said that this is crucial because in our proposal it's exactly the same as Jean-Marie has just explained. There are some codes hidden in the dictionaries which allow to report something related with resistance, suspicions of resistance. So the idea is quite simple, it's just to promote using these codes when health professionals are reporting in different countries. And well, we know when we receive just one report with one suspicion. Well, we consider that single report as a single report, and the value of pharmacovigilance from my point of view is when you start receiving a cluster of reports with similar medicines, similar active ingredients, and similar defects. So if, for example, in one province in Kenya you start receiving five, uh, ten reports of a suspected resistance to X antimicrobial, then perhaps you as a pharmacovigilance center could start thinking that there is something there. So this is the general idea that we have. Not different from standard pharmacovigilance. On the contrary, the idea is to use pharmacovigilance because we have the tools, we have the codes, and we have the network.

Federica Santoro 11:14

Do you see then AMR specialists accessing pharmacovigilance databases or pharmacovigilance specialists being trained and perhaps specializing in antimicrobial resistance?

SPEAKER_02 11:26

No, I think that it's good that specialists in microbiology know that it is possible to report suspicions of resistance through the pharmacobigilance network. And on the contrary, it's important that training people that are already working in Pharmacobigilance Center are aware that some of the information collected can be of interest to all professionals working with antimicrobial resistance, etc. etc. Just from my point of view, at least, it's just a matter of promoting pharmacobigilance in this case and the potentialities of pharmacobigilance.

Federica Santoro 12:13

So we've talked now quite a bit about detecting inappropriate prescribing or inappropriate use of antibiotics. But another major driver of antimicrobial resistance are substandard and falsified medicines. Would your system capture these issues?

SPEAKER_00 12:32

Yeah, I think to a certain extent our system would capture these issues because in countries where you have falsified medicine, there is no one system. Of course, there are policies put in place to track down illegal commercialization of such medical products. But if you have a strong pharmacovigilance network or agency in the country where basically you inform those who see first to report what they see, a lot of pharmacists, a lot of clinicians or those involved in the whole supply chain on a country level, they are better equipped maybe to detect falsified medications. And then what we see with our 17 codes, some of these are related to resistance, others ineffectiveness, others off label use, and then we have a group related to errors, medication errors. But if you focus on that second category of ineffectiveness, if you are treating a patient, or if a patient comes back to the drugstore and you actually start suspecting that there is ineffectiveness, therapeutic non-response, maybe there is a resistance, maybe the medication is not working, or it is not working because it's not the intended medication, maybe it's not correct. So it's from that suspicion that from the moment you have that one report at least from pharmacist X or clinician X, and if you are able to say, okay, in this period, we got similar reports of suspicion of ineffectiveness of non-response, and more than one pharmacist or clinician suspect that there is a falsification issue, then if, as Albert was talking about clusters, if you have these kinds of clusters and then you are able to place them in time and you have systems in place to link them to batches that came into a country or to imports that came into a certain region, if you are then able to track down that supply chain and see what went in the end to this cluster of pharmacies or hospitals that are reporting the suspected falsified medicines, then we can say that using that particular set of codes of ADRs, you can actually track down fake antibiotics, which are actually a big problem for many of the developing countries where, first of all, at the ports of entry, there is no robust ways of ensuring that what goes in the end in the country is quality assured. You have to rely on the end users to tell you that it's not working, and you start then suspecting why it's not working, noting that falsified products would be a factor in that cluster of their ineffectiveness.

Federica Santoro 15:13

So it's great to hear about the strengths and the potential of the system. But let's talk about problems now. What do you see as the limitations of your proposal?

SPEAKER_02 15:26

Well, I think that uh in part the same limitations that pharmacovigilance or the deployment of pharmacovigilance in countries in general, and especially in low and middle-income countries, which has some budgetary problems and turnover of health professionals in pharmacovigilance centers. So uh this can be a bottleneck for reporting in general, and of course for reporting these specific adverse effects. I would say that the involvement of health professionals in the countries is also very heterogeneous, and it's also one of the main reasons leading to underreporting, which is the Achilles Hill of Pharmacobigilance, as we all know well. And the last point is as I said before, to try to understand that we report suspicions and we collect suspicions. So we have to investigate these as suspicions, not as certainties. So these are uh the drawbacks of uh using pharmacobigilance in this case.

Federica Santoro 16:36

How about you, Jean-Marie, as lead author on the research? What are the sore points that you see going forward?

SPEAKER_00 16:43

Yeah, I think some of the weaknesses, of course, as I was mentioning, it is that issue of capacity. I've traveled a lot in countries where you look at uh the whole healthcare system from the capacity point of view, from the skills, the infrastructure, the manpower. One of the bottlenecks will be to be sure that what you are collecting in a country A can be easily compared to what you are collecting in country B. If you look at the history of the WHO program for international drug monitoring, which tasked the Observer Monitoring Center to collect adverse drug reactions reports, you will see that there are countries that have been members of that network since the 60s, the 70s of the last century. But you also have countries which joined maybe two years ago, and you still have countries that are still having this uh associate members status. They've not yet managed to collect the minimum of 20 ADA reports. So the weakness is in that gap of capacity. So if you if you want to use our system, you'll have to trust those who will annotate a report and give it a specific code. You'll have to trust that it is a well-informed coding. Because in a country with 20 year experience, you'll have maybe a department that has been coding these reports for the last 20 years. They are sure if they say we are going to code this as off-level use or we are going to code this as pathogen resistance, for them the certainty level is so high that when you come to the centers that have been doing this for the last 12 months, it will be difficult then to know that the off-level use coding, if it was correct. So we still have a lot of work to do in training, in making sure the pharmacovigilance capacity gap is narrowed. The narrower you have this gap, the better your data you'll have. So it is for me, it is still that capacity issue. But then again, since now everyone, from politicians to everyday clinician, we're talking about AMR as being not the next but the already present biggest global health problem. If we are able to sell this instrument, these PV instruments as one of the tools we have to curb AMR, then probably also that capacity of pharmacovigilance will be given some attention to make sure what we are collecting can be trusted as a source of information to revise policies.

Federica Santoro 19:05

Albert, in 2015, the World Health Organization launched GLASS, the global antimicrobial resistance and use surveillance system. And the idea with that was to collect and analyze and share AMR data in a more standardized fashion. What would your approach add to WHO's system?

SPEAKER_02 19:31

Well, in my opinion, both ideas are complementary. What GLAS is doing is they involve countries and they try to collect from national laboratories local data from resistance. And also GLAS is collecting data on antimicrobial consumption in the participant countries. In my opinion, this is a good approach. This is an approach that since 2015 is growing and is uh having good results. But in some countries, the number of bilobal laboratories and local data on resistance are lacking. So this is why in this approach we try to, let's say, go to the border of pharmacovigilance and look beyond the fence. And usually when you approach the border of some discipline and try to look beyond the fence, you see other ways to do things, you see other people that are approaching the same problem and different way. And this is why we are proposing that sort of symbiosis between both networks. We or the idea that is beyond the thesis of Jean-Marie is not to try to do the work that is already doing glass, it's just to try to identify gaps and try to bridge these gaps, collaborating with their methodology and their ideas. Just try to provide some data that we can collect from an alternative way. And as we said several times, our approach is just to have suspicions. GLAS obtain direct data from laboratory. We just suggest wow, perhaps in that place, in that province, in that country, there can be a problem of resistance, equal resistance to this and that. And that's it. And we try to report to alert on this. And then they are the specialists in this area, so if they have the opportunity or or they judge that this is important, they will go to the place and try to confirm this.

Federica Santoro 22:00

And if you'll allow me to dig a little deeper in this issue of laboratory resources and how the pharmacovigilance tools can complement that. Because I can certainly understand your argument that pharmacovigilance networks can help track resistant pathogens in communities with limited lab coverage. But in the long run, shouldn't we anyway be boosting diagnostic capacity in resource-poor settings? Because I imagine that is the way to identify and handle AMR cases quite quickly, isn't it?

SPEAKER_00 22:38

Yeah, I think on the long run you will have to rely on strong diagnostic capacity. But as we know, the diagnostic market is also developing fast, but it's still in favor of the places where pre-existing capacity can can house, can host the new technologies. If you look at the ongoing pandemic, the PCR capacity, the technology was there, but in the end, countries that were able to use PCR to track down the virus were those countries that had already, you know, laboratory capacity already implemented. In places where you didn't have already good solid lab capacity, some were trying to um to use the little they have, but also some were trying to see if there is something which is maybe point of care kind of diagnostic that can be used. You do have places, for example, where the number of people who were able to basically to do the routine PCR testing are a few individuals if I'm talking about a country, let alone a hospital or a clinic. That's why I was talking about the um the quick versus the slow evolving field if you talk about pharmacoputers. So it goes, it is the same with the diagnostics. We will have at least in the coming five, ten years, twenty years, countries that will lag behind in terms of what they can do to analyze. You will have these countries that will always, at least in the coming decade, rely on one well-equipped, privately funded laboratory or hospital in the capital city. But then if you think about the the remote corners of certain countries, internet is already there. These regions, though they don't have skilled lab technicians to do PCL or other diagnostics, they will have capabilities with the mobile phones to maybe to report X suspicions with accurate geographical tagging. If we are able to localize that suspicion, at least you can then rely on that scarce resource of diagnostic capacity. Let's say if a cluster is suspected to be localized in a few provinces far away from that uh well-developed, well-equipped city, at least you can use the few resources to then confirm. If you can rely on pharmacovidence data to say, okay, the cluster we should be tracking down is the one in the north, not the one in the west, then you can be sure that your small lab capacity can focus on that uh suspicion where you are likely to find something instead of misusing that little capacity you have.

Federica Santoro 25:17

Absolutely. And uh, following on on this theme of uh many disciplines interacting together, what are your thoughts on other disciplines and other approaches that you think will help?

SPEAKER_02 25:32

Well, in my view, the near future engineers will take an important role, for example, and big data managers. There are some devices that are under development. They are a sort of cell phones. They are connected to the satellites, but they have the capability to test droplets of, for example, a sputum or urine. Then just with one droplet, the machine can start analyzing basic genomic questions. I don't know the technicalities of the bacteria that is producing the infection in a patient. So that uh small tool connects automatically to a big database where they have sequences and then can identify possible resistance to one antibiotic, another antibiotic, etc. etc. But we are uh talking from let's say our view today in uh laboratories, and it would be excellent that I don't know each uh province, each department in every country had a complete uh microbiology laboratory, but perhaps in the near future this will not be necessary. We will have these devices that are portable, etc. etc. I don't know, it's the future, uh, but uh it's it's a possibility.

Federica Santoro 27:09

And we will see what the future holds. I actually trained as a molecular biologist, so what you just described makes me really excited. Um finally, Jeanne-Marie, what lies ahead? Where are you hoping to take the project next?

SPEAKER_00 27:25

Yes, that's a question which uh it doesn't keep me awake, but it keeps me thinking a lot almost every day. It is actually good to see when you have an idea. It feels good to see that you are not alone. So from the moment the few papers were published on this topic, I got a few people contacting me, pretty much from every continent. The latest is uh someone who's in Haiti, someone who's really working in the area of antimicrobial resistance, but someone who has also an interest in pharmacovigilance. He's now reading the papers. I hope to get in touch with him again. But the idea is that we then discuss on how to confirm this failure. Because for me, shortlisting those 17 med records, so those 17 adverse drug reaction terms, it is a start. Now, what I'm looking for is a group of us thinking in the same way or others not thinking the same way, coming together as a network of individuals who actually see the potential in this. Maybe I'm not going to be the one to implement the whole thing, but in the first step, as I was saying, it is to make everyone aware of these overlooked 17 codes. They are hidden in a huge database of more than 20,000 codes. But if I can at least get a few other thousands of people who work in pharmacovigilance on antimicrobial resistance surveillance systems, if these thousands of people can be aware of the existence of these 17 codes, and then we can, in the next step, inform them about why we think this can help in generating suspected resistance or inappropriate use. At least that would be a good step where we can say maybe in a year or two, pharmacovigilance is already contributing to tracking AMR. So that's a good start. So I'm looking at basically establishing a small international network. I just mentioned someone from Haiti, I have a contact from Uganda, contact in Rwanda, someone who is in India, a lot of people in Europe, someone who's in Canada. I do not belong to a pharmacovigilance agency of a country or to an AMR surveillance agency of a country. But the idea would be to have this network of individuals who can then carry the idea further. And then if you ask me, what then would you like to see, let's say, in five, ten years, maybe something that I didn't mention earlier is the aware classification from the WHO, the access, the watch, the reserve classification of antibiotics. That is something which is in its infancy, if you'll ask me. But I'm looking at then a step where our methodology, which uses these 17 codes, if we can say, okay, look at antibiotics in those three categories, and then we can use these 17 codes to focus on limiting consumption of, let's say, reserve antibiotics. That could be a goal which I'd be proud to say, WHO, thank you for offering that classification of the away groups. And thank you also for using these sets of codes that we came across when we were looking at this potential use of PV to curb AMR. I can go on for hours, but in a way that that's what I want to see. I don't want to see this as a standalone strategy which will evolve or die its own slow death. I want to see it integrated in what is existing, integrated in what is to come. I want to see it picked up by the engineers that we're talking about and the big data mining specialists who will be joining the field. So it's a tool which, in the first place, I want the world to know about.

Federica Santoro 31:06

And I hope this episode will help you spread the word even further. So, people, the word is out there. Jean-Marie is looking for collaborators to expand upon his research. So if you're interested in a potential link between pharmacovigilance and AMR, do reach out to him. Alberte, do you want to add something on that?

SPEAKER_02 31:25

Yeah, I think that it's a big opportunity to try to push more and more pharmacovigilance in countries. We know that there are many countries with pharmacovigilance program, a pharmacovigilance center. Unfortunately, not all countries are pharmacovigilance is working at the same speed. So well, this can be an additional uh back, an additional effort to try to help these countries to strengthen their network. So it would be a good opportunity.

Federica Santoro 31:59

Well, with that, I'll thank you both for your time. It was a really interesting discussion. Thanks again for taking the time to talk to me.

SPEAKER_02 32:08

Thank you.

Federica Santoro 32:13

That's all for now. But we'll be back soon with more conversations on medicine safety. If you'd like to know more about Jean-Marie and Albert's research, check out the episode show notes for useful links. While if you're interested in AMR from a social and behavioral perspective, then don't miss episode number eight in the Drug Safety Matters archive. If you like our podcast, subscribe to it in your favorite player so you won't miss an episode. And spread the word on social media so other listeners can find us. Apart from these in-depth conversations with experts, we host a series called Uppsala Report's Long Reads, a selection of audio stories from UMC's Pharmacovigilance magazine. So do check that out too. Uppsala Monitoring Center is on Facebook, LinkedIn, and Twitter, and we'd love to hear from you. Send us comments or suggestions for the show, or send in questions for our guests next time we open up for that. For Drug Safety Matters, I'm Federica Santoro. I'd like to thank Jeanne-Marie Vianneja Barughira and Albert Figueiras for their time, Matthew Barwick for post production support, and of course you for tuning in. Till next time!