See show notes for this episode: S 02 | Ep 44 How to Turn "The Good, the Bad, and the Ugly" of Pharma Data into a Trusted Asset .
Alex Shevelenko: Today, we're joined by Michelle Bridenbaker, Global Head of Medical Excellence and Communications at Recordati, which is a publicly listed pharma company. She's also—just to keep herself busy—VP of MILE (the Organization of Medical Information Leaders) and CEO and owner of Unbiased Science. Welcome to the pod, Michelle.
Michelle Bridenbaker: Hi, Alex. Thank you so much for having me. Really appreciate it.
Alex Shevelenko: Well, the reason I'm so fired up to have you—and the three roles, the three hats that we just described—will signal to people why this is a super important episode for us. We know that there is a ton of scientific, evidence-backed information out there that somehow doesn't get through to the average healthcare professional, clinician, and certainly not to patients and the audiences that care about them. That includes governments and the whole ecosystem.
You are at the forefront of trying to make this complex scientific information come to life and separate the hype from reality around medical information. I couldn't be more excited to learn what works in this world, and where you think we could be learning as an industry from other parts of the communication spectrum.
Michelle, maybe you could set the scene with some background on how you got into medical information and communication, because your journey is very unusual. It's probably one of the reasons why you're such an authority on this topic.
Michelle Bridenbaker: Yeah, thank you. I think for me, my foray into science and communication definitely started after graduating with a background in biology, but then really becoming a healthcare professional as a nurse. Particularly in a critical care setting—and even sometimes the emergency setting—I had to spend a lot of time communicating.
I realized that that was the most important thing I would do. Even when other healthcare professionals, like physicians or pharmacists, would speak with the patients, the patients still had a lot of questions. I remember at that time realizing how critical my role was in that communication pathway, and that they needed a trusted messenger. That really became obvious to me.
When I progressed into the industry, it initially came down to working in drug safety. That gave me a really strong background in understanding drug development, how you monitor the safety of medicines, how to make sure we communicate side effects effectively, and also, of course, updating the materials used to dispense medications.
As that developed over time, I moved into medical information. In the medical information space, I knew this was where I really wanted to be. I had also done some freelance medical writing with Medscape when I was in nursing school because, as you can appreciate, nursing school and universities in the United States are very expensive. It was actually just gig work, but it was a great experience because I was creating resource centers with experts in cardiology to break down information for clinicians to digest quickly.
I carried on with the medical information work—obviously in the beginning just creating content itself—but then progressing to leadership in that space. That leadership really came from a lot of trial and error.
I handled some complex product portfolios and realized that, a lot of times, people just didn't really connect with the prescribing information. Even in our own companies, we weren't seeing patterns that we needed to see to realize there was a misunderstanding—maybe with a device or a mode of administration. That continued to fire up my interest in that space, even with technology. I was like, "Okay, how do we connect? We have all this data, and we're constantly looking at it, but it's not really hitting the spot." That's actually how I started many years ago with IBM Watson, creating a chatbot at the time. It was an epic fail, but I was super, super happy for that because I learned a lot. I learned what works, what doesn't work, and how hard it actually is to go into that space. That was okay because it just kept me on this path to keep exploring ways to make things more accessible, and to use technology to augment our capabilities.
Now that I've been in Europe for 20 years, that has developed into working with the European market, which is multilingual and multicultural. At that point, again, we were working with technology more in the translation space. We use English as our core language as scientists, but if you're going to reach people, you obviously have to speak their language.
It was an interesting journey. I kept digging into different things, realizing that technology helps achieve certain aspects, but not in totality. You're getting little bits and pieces here and there. As I kept developing my career in medical affairs, I did every role. I've been a medical director, a medical advisor, med info... really, I've done everything.
For me, that's important. I think it's crucial that I've had that deep experience in different aspects, even quality assurance and pharmacovigilance (PV), as I said before. The information ecosystem is large at our companies, and we don't leverage a lot of it. Once I started to get more senior in the medical information space, I realized that to be able to augment our teams, this ecosystem is really, really important, or we're never going to get our data out there. Truly, that is what also led me to look at a new space, which was Unbiased Science.
Michelle Bridenbaker: It was really about communicating during COVID and realizing the power of social media. With social media, I absolutely realized—and was humbled by—the complexity of that space in the attention economy. It really put into perspective where we, as an industry, fit.
You know this—we've been talking about multi-channel, omni-channel, and all these different things. They're just the modes for getting things out there. But the biggest challenge is actually catching somebody's attention. This is where social media gave me another toolkit to be able to say, "Okay, now you've got the good communication, you've got the technology, and you've got the channels, but you have to figure out how to make that information sink in and ensure it gets in front of the people who actually need to see it." So, you can see it's a very logical progression. Even if you looked at it on paper and wondered, "How did she get to that point?"—that’s how it progressed for me.
Alex Shevelenko: Well, let's take a quick step back for people who are not in the life sciences and biopharma universe. Take a moment to explain: what exactly is medical information, and why is it important? I think we take it for granted or consume it sometimes—probably not enough—but people might underestimate its importance and the sheer number of stakeholders involved.
Michelle Bridenbaker: No, that's actually a good shout-out because I remember speaking at a conference for call centers, and I was alongside banking, insurance, and all these other companies that had no idea pharmaceutical companies are legally required to have this function. It supplies information about medicines not only to the healthcare professionals who prescribe or support their usage, but to the patients and their caregivers as well.
What we really do is take all of our complex internal information and blend it with what has been published in the external world. Our job is to provide completely fair, balanced information. I always tell people, "We give the good, the bad, and the ugly." If you're doing a good job, you are making sure someone understands the right patient to give this medicine to, the right way to administer it, and the evidence that supports its use. In some cases, you are saying, "No, this is not the right patient. They shouldn't be receiving this medicine." Maybe they are taking a different medication that could interact badly, or they have a condition that won't see a good result from this therapy.
We help bring together the precise information that time-poor clinicians need to competently prescribe the medicines our companies make. For the patients, our job is to ensure they take it correctly, use the devices properly, and understand why it's being administered a certain way in the hospital. It's a great extension of the healthcare environment, proving that we are also a part of your healthcare team. It's a fair balance. There are people in this function who are solely out there to protect the patient so they don't receive the wrong medicine or make a wrong decision by accident.
Alex Shevelenko: One of the challenges we see as specialists in regulated communications is that to provide this fair, balanced, and detailed approach that integrates everything, you end up creating a lot of information. A great example for folks outside of medicine would be insurance policies—they can run incredibly long with all their provisions and details. That's a regulated world. Or if you're an investor looking at SEC-mandated 10-K filings or annual reports, there is a massive list of risks filling up a huge chunk of that report.
It sounds like a very similar scenario here. The challenge is: how do you get to the essence without getting lost in a sea of information? It's there with good intentions and it's important to have, but it can lead to people being overwhelmed and just not reading it. What have you found that works—and what doesn't—in medical information?
Michelle Bridenbaker: I think it is still a process of trial and error. It’s just not yet on the level that I've seen in social media, which is much quicker and more dynamic. But over the years, I have seen an evolution in transparency. You can get information from a pharma company, and you should read your product insert.
Honestly, I don't think it was until COVID that people actually knew they could report side effects. Before that, no one really understood that there was a mechanism to report anything they experienced, or that instead of throwing out the leaflet inside their medication box, they should probably have a good look through it and ask their doctor questions. That was not necessarily a behavior we saw as much as was needed before COVID.
Alex Shevelenko: So what you're saying—and I'll pause on this, but we'll come back to it—is that in psychology, there is this notion of authority. Someone in a medical uniform of one shape or another holds authority, and people used to rely entirely on that authority when taking medication or trusting what happens. COVID shook all of that up because authority was challenged, and different people were claiming authority in different ways. Is that a fair way to put it?
Michelle Bridenbaker: Yeah, I think that was part of it. At the time, we also had a system that was heavily time-constrained. A clinician can only go through so much. You also have the challenge of trying to get people to change their behaviors. Are you really going to successfully talk about someone's smoking habits, exercise, or diet in a ten-minute visit? Probably not.
The healthcare system has evolved in a way that makes it challenging to have the deep conversations that clinicians need to have. At the same time, patients are finding information online and doing their own research to be more informed. That momentum was already building, but it was put under a microscope during COVID. It accelerated significantly at that time.
What we see now, though, is a new voice—a new attention-grabbing force in our lives: social media. That was ramping up simultaneously, and so was AI. It was a perfect storm, in my opinion. You had COVID, you had social media really taking off, and you had health influencers emerging—not necessarily health communicators.
Alex Shevelenko: And, you know, worst of all, AI is being used to proliferate information across some of these social media algorithms. Unfortunately, it's usually not great information because the good, robust stuff doesn't perform as well on an algorithm. So it really has been a perfect storm. Basically, the crappy influencer content—which may have a non-scientific agenda, at a minimum wanting more clicks, or at a maximum sponsoring some kind of product—is polluting the AI ecosystem.
If people are using generic sources to get their information, how is that influencing the medical information that sometimes goes to payers and government entities who allow certain products into specific markets? What have you found with that audience regarding the shortcomings and successes of the medical information field in engaging that range of stakeholders?
Michelle Bridenbaker: What I always find with social media, particularly content that is very successful in gaining virality, is that it often has a kernel of truth, but it's presented without any sort of nuance. Some people don't do it maliciously; there's a difference between misinformation and disinformation. Disinformation is when there's always a hook, like, "Oh, I have a code for a product that can actually cure your disease instead." That becomes a completely different territory.
What I see is that information has gotten out there, and people are now trying to pull back the curtain to understand very complex systems—systems that we know are complex because we maneuver through them every day. Whether it's a healthcare system, a payer system, or even just regulatory approval, there's a lot of behind-the-scenes work that no one sees. No one sees all the ingredients and everything that goes into the final approval of a medication.
To give a concrete example: there's this constant belief that there was never placebo-controlled testing for the COVID vaccines. This stems from a strong anti-vax movement, of course, but it spreads because people don't understand vaccine development or how clinical studies for vaccines are actually designed and conducted. It's a complicated area, which is why people specialize in it. There are experts who focus entirely on clinical trials and clinical trial communication because they are out there trying to help recruit patients and ensure they have true informed consent when they participate. That's where the first layer of communication happens.
Another aspect—and I think Europe is a good example—is that they wanted transparency on things like side effects, so they opened up a public database where you could go in and see what adverse events were happening for different medicines. The problem, again, is that the nuance is missed, and the data literacy isn't always there for people to understand what they are looking at. They don't realize there are a lot of reasons why something might be listed in a giant database of adverse events. They don't understand causality, and they don't understand the complicated mechanisms we use to determine, "Is this a real safety signal? Do we need to put this on the product label?" There's a lot of complexity to our healthcare system and within pharma's framework for getting medicines approved around the world.
Because of that, people get reductive. Influencers successfully capture people's attention because they deliver a simple message. You and I know there's no simplicity to the process; a massive amount of work goes on behind the scenes to create that final product. That's the difficulty we face. We're starting from a different baseline than before because trust has been eroded. We have highly complex systems that are hard to maneuver, and people have others telling them, "It's simple to understand." Except, it's really not. It's never that simple.
Alex Shevelenko: Right. We're all becoming clinical experts with a little ChatGPT and a lot of naivete. What you're saying is there was a reason you couldn't just casually crack a molecular biology course—because you needed to go deep and really understand the fundamentals before you could draw accurate interpretations.
The average person will go online and challenge it, asking, "How did the experts get it so wrong? How did all those experts mess up?" I'm sure you've heard that kind of remark. What's your thesis? First of all, was there a big mess-up by the experts, and where did they get things wrong, especially during COVID? Feel free to share your own journey in fighting that. Generally speaking, whether it's in health or other areas, the human mind is always looking for shortcuts. Complexity is not a sexy answer. Saying, "Trust us because it's complex" doesn't land well anymore, especially when there's a perception that AI knows more than a professional. In some areas, it might, but guide us a little bit on how we help answer that.
Michelle Bridenbaker: I think what people fail to realize is that COVID was a very novel situation. There's a reason it was called a novel virus; it was a new situation that we had never faced on such a global level. At the same time, unfortunately, many of our public health organizations were struggling—perhaps underfunded—and were not great at communicating. It's funny because now I actually work with some public health offices in the United States through my work at Unbiased Science.
Humans need answers, and that's completely fair. You want to know why something is happening to you. "Why did I have a side effect from this medication? Why am I not getting better?" People want to understand these things and receive a straightforward answer.
As I said before, the amount of time our clinicians have to really talk through these things is incredibly limited. I know the moral injury that our clinicians suffered during COVID—both nurses and physicians—was because they saw devastating effects of the virus in the hospitals, but they were also losing that trust with their patients. Patients were just so lost because of what they were inundated with; they were drinking out of a firehose on social media.
Out of that COVID period, a few of us realized we needed to do something differently. I had a healthy-sized Facebook following at the time, and people knew I worked in this space, so they would constantly ask me questions. We realized we were science communicators, and we needed to find a way to help people understand highly complex ideas. For us, it became about putting it in very human terms.
Michelle Bridenbaker: But it was also about not insulting them and saying, "God, that's such a stupid question," or "Why do you think that?" in a judgmental way. It was about saying, "Help me understand where your perspective comes from. Where did you hear that information?" Connecting with people on a human level is really critical.
The time constraints on our clinicians are reducing their time to be human. This is why I say that if you use AI in that space, it needs to allow them to have more time to be human, not just to document more. I mean, documentation is an insurance requirement, but AI should free up time for them to spend with their patients.
It's the same for us with what we do on social media. We try to give people a way to critically appraise the information they receive. When you're getting that much information, how do you show people that if a post gives you a very strong gut reaction, you should probably double-check it? That reaction already tells you there's something that is likely not entirely truthful or fully correct. It's about giving people the tools to say, "I need to consume information differently." But it also pushes us to say, "We need to find a way to communicate better, more clearly, and in a more engaging way." We need to go where the people are, but also be human and truly connect with them—which is the hardest thing to do, to be honest, especially in this industry.
Alex Shevelenko: So when you're saying something hits you emotionally... typically, in the social sphere, fear works really well.
Michelle Bridenbaker: Yes.
Alex Shevelenko: And you could imagine a lot of fear around health-related topics. So, you amplify the fear, and then you come up with a simple recipe to address the pain you've just activated. That's essentially the playbook of an average influencer campaign to drive engagement. You're saying that nothing is actually that simple—the fear is overstated, and the solution is also overstated because it oversimplifies something that is a lot more elaborate. Am I looking at this the right way?
Michelle Bridenbaker: Yeah, you're framing it correctly. The problem is that social media, especially now, is heavily geared toward clicks, virality, and engagement because it's monetized. It gets people onto their pages.
I actually learned a lot about this from understanding sentiment analysis when I was working with AI. When people are experiencing confusion, anger, frustration—all the negative emotions, or even the super extreme positive ones—that's when engagement spikes. Whenever you feel an extreme reaction, that’s when I tell people to just take a minute. I’m the same as a consumer. Sometimes I read something—whether it’s political, scientific, or medical—and when I get that intense reaction, I know I need to take a step back. I need to validate it across a few different platforms or check a few different global news sites to see how they are talking about it. I have to ask, "Okay, is this even plausible?" That’s the hardest part for people when it comes to understanding science. Now, when you search in ChatGPT for an answer to your question—and you know this very well because you work with AI—if you frame the question a specific way, you have already potentially biased the prompt. If you're just trying to validate an existing belief, it can lead to this extreme division between people. Whether it's politics or health, you are appealing to what people want to believe. If new data doesn't fit their narrative, it becomes highly discordant and very challenging for them to accept.
So, first, we have to get good information out there so people can see different perspectives. But we also have to encourage people to look at the other side of the coin for any given scientific topic so they can realize the truth is usually somewhere in the middle. Unfortunately, the algorithms—and even some of the AI tools we have—can still create confirmation bias. The good news is, as I’ve tested these tools, we are starting to see data where they actually present more fair, balanced answers. The technology is improving to help get better information into the hands of end-users.
But it's hard. Health literacy is not particularly high, and general literacy can be a challenge too. This is something we have to be very conscientious of, especially if you have a person face-to-face. You need to connect with them right there, because people still learn best from other humans. We need to consider this whenever we have that opportunity—particularly as healthcare professionals, scientific communicators, and definitely within the industry. When you have that person's attention, you need to make sure you do it right.
Alex Shevelenko: Coming back to the AI topic, what I heard you say is that even when writing a prompt today, you need to set the context. For example, you should type, "Imagine you are a trained professional clinician and an expert on X, Y, and Z. Now, give me your opinion," instead of just asking a generally biased question. That sounds like a pretty obvious best practice for prompting, as well as instructing the AI to only look at published evidence. Most people probably wouldn't think to do that because they aren't trained in it.
The second piece—and this is something we work on quite a bit on the AI side—is restricting the environment so we only allow verified, correct information for the AI to actually digest. That gives you the confidence that you aren't getting influencer content mixed in. From there, you can still ask the AI to generate a version of the explanation suitable for a 10-year-old or a 15-year-old, but you are much more confident that the output is compliant and accurate.
You can also supplement that published information with a medical dictionary or other trusted terminology feeding into your system. So, it's not just a matter of uploading five PDFs from published studies; you can supplement it with incremental, verified sources within a contained ecosystem. That feels like a much safer approach.