AI: Friend or Foe?

AI Friend or Foe?

One University of Georgia professor looks to algorithms to predict and prevent adverse drug events, while academia begins to explore ways to teach with ChatGPT while preserving learning.

By Athena Ponushis

Artificial intelligence has the potential to be lifesaving, helping physicians detect signs of disease and helping pharmacists offer more personalized care. Such technology holds the promise of improving health outcomes while reducing healthcare costs. It sounds like a dream, albeit a little bit daunting. 

AI has become the center of conversation as it continues to permeate our everyday lives. Chatbots, like ChatGPT (a large language model released by OpenAI), have seeped into the classroom, posing a potentially Goliath-sized threat to education—enabling cheating and plagiarism, while inhibiting comprehension and aspiration. 

ChatGPT image illustration

But just as Goliath was defeated, the technology might not be as big a threat as perceived, suggested contributing writer Ian Bogost in an essay in The Atlantic entitled, “ChatGPT is Dumber Than You Think.” As Bogost writes in the article, large language models that mine the text they have been fed to predict the next pleasing word “are surely not going to replace college or magazines or middle managers. But they do offer those and other domains a new instrument—that’s really the right word for it—with which to play with an unfathomable quantity of textual material.” By embracing AI as an “instrument,” or rather, as an assistant, pharmacists and professors are finding ways to help patients and students, all for the benefit of public health.


“With every decision in the ICU, there’s a reason we are doing it and there’s a risk associated with it. You are weighing risks and benefits. We hope that AI-based tools like this can give you more nuanced information…and that might sway your decision as a clinician because you have more patientspecific data.”

—Dr. Andrea Sikora

A Tool to Guide Clinical Decisions 

Each year in the United States, 1.5 million patients admitted to the intensive care unit (ICU) experience a serious adverse drug event (ADE). The related healthcare costs to treat such ADEs exceeds $1 billion. Dr. Andrea Sikora, clinical associate professor at the University of Georgia College of Pharmacy and critical care pharmacist at Augusta University Medical Center, has found that patients in the ICU suffer three times more ADEs than other patients because of their complex, high-risk drug regimens. When critical care pharmacists are part of the ICU team, pharmacists make preventative medication interventions, reducing ADEs by 70 percent. 

The problem remains that critical care pharmacists are busy, often responsible for too many patients, making it hard to perform such timely interventions. Wanting to help patients and pharmacists, Sikora turned to AI. In 2016, she developed a clinician-friendly IT tool, called the Medication Regimen Complexity-Intensive Care Unit (MRC-ICU) Scoring Tool. Essentially, the tool adds up the medications on a patient’s chart, giving the patient a score (similar to a Body Mass Index (BMI) score) and that AI-informed score makes a prediction, alerting a pharmacist to prevent an ADE before it happens. 

“With every decision in the ICU, there’s a reason we are doing it and there’s a risk associated with it,” Sikora said. “You are weighing risks and benefits. We hope that AI-based tools like this can give you more nuanced information…and that might sway your decision as a clinician because you have more patient-specific data.”

When Sikora first started playing with the MRC-ICU scoring tool, she used traditional statistical methodology. She saw some good things, but nothing great. Then someone approached her at a meeting and asked, ‘Have you ever thought about using machine learning for this?’ She hadn’t, and she wasn’t even sure what machine learning meant. “That’s the beauty of attending national conferences,” she quipped.

What struck her about machine learning was not just its ability to deal with so many data points, but its capacity to deal with nonlinear relationships and nonmonotone patterns that are more difficult to model using traditional logistic regression to predict disease outcomes. “That sounded very much like the problem we were facing,” Sikora said. “ICU data is very heterogeneous, very complex, and there are lots of different factors put into any situation. Looking at the preliminary data, my team and I thought, ‘Wow. We think this is going to work,’ and that was really exciting.”

Sikora defines AI as the science and engineering of creating intelligent machines that can achieve goals like a human, whereas machine learning is a computer’s ability to learn without being explicitly programmed. Using the analogy of a traffic light, Sikora said, “You tell a computer, ‘Every time the light is green, cars go,’ but machine learning watches the traffic light and after a while, it determines, ‘it seems like cars stop on red and go on green,’ so it’s able to infer the pattern on its own.”

Sikora and her team have been trying to predict which patients are most at risk for fluid overload so pharmacists can give them diuretics, concentrate their medications or offer other interventions. When they did traditional logistic modeling and ran logistic regression, the team found a decent way to predict fluid overload but it had nothing to do with medications. “It was weird to us that drugs were not significant predictors of fluid overload, because fluids come from drugs,” she noted. But when they ran their machine learning analysis, it created better prediction algorithms. When they looked at the feature importance graph, which tells you what was important to the algorithm, medications were at the top. 

A Tool to Guide Clinical Decisions

The idea behind the MRC-ICU score was that it did not have to deal with severity of illness but rather the medication regimen complexity of the patient, which Sikora thought was a more meaningful metric for what a pharmacist actually does. In order to apply that metric to predict and prevent ADEs, however, she needed machine learning because it can handle nonmonotone patterns. “The hypothesis that we are working with is that medications are independent risk factors for disease but they are often not included in prediction algorithms, and that doesn’t necessarily make sense,” she explained. “We are basically ignoring this entire wealth of relevant information when we are trying to make decisions, but the problem today has been that medications are messy. It’s not just the drug, it’s the dose, route, frequency and so many other elements, and that’s really difficult for traditional modeling to be able to handle, whereas machine learning is far better able to handle that. So that’s the thing we are exploring.”

Sikora received a $1.86 million grant from the Agency for Healthcare Research and Quality (AHRQ) last August, complementing an earlier AHRQ grant of $296,000 to pursue her research on how AI may guide pharmacists to provide better care. Her vision would be for AI tools to give clinicians more detailed information on the risks and benefits of the decisions they are making. “I see it as clinical decision support, but it’s more like it’s giving you relevant, real-time predictions,” Sikora said. ‘That is what is most exciting about the work that we are doing.”

Sikora can imagine AI helping pharmacists with workflow, improving inventory management, maybe even filling prescriptions. She does not see these tasks as replacing the pharmacist, but freeing up the pharmacist to spend less time on physical labor and more on cognitive skills: ‘Do I agree with this medication?’ ‘Have I talked with the patient?’ ‘Have I helped them strategize how to take their medications?’

“ChatGPT is not going to replace the learning that I want them to experience—they still have to write their reflections. ChatGPT will just help them refine it and that might actually help them become better writers.” 

—Dr. Sara Trovinger

Leveraging ChatGPT to Enhance Learning 

AI is coming under greater scrutiny thanks to the recent proliferation of chatbots. Dr. Sara Trovinger, associate professor of pharmacy practice and director of the Distance Pathway program at Manchester University College of Pharmacy, facilitated an AACP webinar in May focused on how ChatGPT might enhance pharmacy education and student engagement. She compares the release of ChatGPT to that of the calculator: When calculators became ubiquitous in education, math teachers thought, ‘Oh, no. I’m going to lose my job.’ Instead, teachers just learned to teach math in a different way. They let students use calculators. And when computers and the Internet first appeared, teachers worried, ‘All my students are going to cheat.’ Again, teachers figured out a way to integrate computers. Now that professors are coming around to the idea that students are going to use ChatGPT, they are wondering, ‘If we can’t stop it, how can we teach with it?’ 

“I think we are going to be better off if we embrace it and use it effectively, appropriately, and teach our students how to do that because they are going to have it in the real world and they need to know how to operate with it in pharmacy,” said Melissa Bray, director of instructional design at Manchester University, who served as one of the webinar’s panelists. 

Noting ChatGPT’s limitations, Bray said it does not search the Internet, it does not give up-to-date information, and for that reason, it’s not advantageous for health science education because it’s not timely. ChatGPT has been fed thousands and thousands of pages of text through September 2021 and it does its best to predict words that should come next. Bray compares this to grand predictive texting on smartphones, but because of its propensity to spit out misleading information, sometimes it does a good job and sometimes it makes things up. 

But the attraction of ChatGPT is so strong, professors are thinking of productive ways to use it. “One potential use is it can help students when they are staring at a blank page,” Bray said. “It can help them brainstorm information. So, if a student needs to do a presentation or write a research paper, they can plug into ChatGPT and say, ‘Give me an outline. Where should I start my research? What are some common questions related to this topic?’ It can be that jumping off point. Then the student can fill in the gaps, do the actual research and find up-to-date information.” 

ChatGPT can also help students translate information. If they do not understand the explanations professors are giving in class, they can ask ChatGPT to rephrase the material into a language they can understand. “I remember when I was a pharmacy student, I was really struggling because I did not know my Greek alphabet,” recalled Trovinger, who hired a tutor to teach her. ChatGPT gives today’s student pharmacists another place they can go to grasp that base knowledge to help them ask subsequent questions. But again, because ChatGPT has been designed for predictive text, students must make sure to confirm that what they have learned squares with what the professor said.

“To piggyback on that, students can take the actual text from a textbook, just like they can take the words a professor has said about a drug or a process, and ask ChatGPT, ‘Can you rewrite this for a freshman-in-college level?’ And it does a pretty decent job. Because you are providing the text to begin with, it’s not going to go off and make things up, it’s just going to convert words into layman’s terms,” Bray said. 

A professor can have ChatGPT generate text that he or she knows is incorrect, maybe using outdated pharmacy guidelines and present it to the class, Trovinger and Bray suggested, to have students find and fix the errors to illustrate the platform’s pitfalls. This kind of reverse engineering, Bray explained, creates an assignment for students to use ChatGPT that does not erase, but may enhance, learning. 

Trovinger has even found a way to use ChatGPT in reflective writing assignments. She asks students to write about their rotations. They reflect on their experiences and go through all the revelations of the writing process. Then they can use ChatGPT to finesse their writing. “ChatGPT is not going to replace the learning that I want them to experience—they still have to write their reflections,” Trovinger said. “ChatGPT will just help them refine it and that might actually help them become better writers.” 

Even with the sudden integration of AI tools in pharmacy education, Trovinger and Bray believe the adaptation will move slower in pharmacy practice because pharmacists tend to be risk averse. They are trained to trust but verify, knowing that when they go into practice, they have people’s lives in their hands. 

Similar to students taking a professor’s words and putting them into ChatGPT to reword the lesson into more relatable language, pharmacists can use ChatGPT when it comes to complex instructions for a particular therapeutic or disease state, asking, ‘How can I explain this to a patient in the hopes that they will understand?’ In that scenario, one can see the beauty of human judgment combined with AI assistance. And from that view, it looks more helpful than harmful. P 

Athena Ponushis is a freelance writer based in Ft. Lauderdale, Florida. 


To those who may feel trepidatious about AI, Dr. Andrea Sikora recommends the book “Thinking, Fast and Slow,” by Daniel Kahneman, psychologist and winner of the Nobel Prize in economic science. She said the book points out that human intuition is flawed, “we have a lot of biases that we bring into prediction in particular,” but intuition can be aided by providing structure, and oftentimes, that structure comes from using a computer or algorithm. When Kahneman and his longtime collaborator, psychologist Amos Tversky, pitted a person against an algorithm, they found that the algorithm outperformed the person. When the person worked in conjunction with the algorithm, using the algorithm up to a point but still able to apply intuition, the pair outperformed the machine. “I think that makes sense,” Sikora said. “Being thoughtful and structured in your thought process, having more relevant data available to you, sounds like a better way of going about things.” 

The AACP webinar “What is ChatGPT and How Can You Leverage It in Pharmacy Education” referenced in this article can be viewed here: