AI as a Tool to Support Clinicians, Not Replace Them

In this final part of our conversation with Sharmeen Roy, PharmD, BCPS, chief strategy and science officer at DoseMe, she discussed AI as a tool designed to assist pharmacists and clinicians in making more informed patient care decisions. According to Roy, AI is designed to consolidate patient information, aiding clinicians in their decision-making process without replacing their judgment.

"It should be a tool for pharmacists or any clinicians, not necessarily replacing their judgment, but really helping them make the best decision with the most information," Roy explained. "You’re not spending time gathering patients’ charts are complex. I don’t know if you’ve ever seen them, but they have a lot of information. There’s a lot of data, and there’s a lot of assessment going on, and that all takes time. The tool aims to consolidate that information so that clinicians can focus on making the best decisions for their patients."

A significant area of focus for AI is antimicrobial resistance (AMR), which remains a major global health concern. "As far as AMR goes, one of the mechanisms by which we see resistance emerge is subtherapeutic exposure to medication," Roy said. "What our tool allows is ensuring that you're getting the appropriate target quicker, so there’s less time spent with subtherapeutic levels that expose the bacteria, helping prevent the cycle of resistance."

While AI plays a significant role, Roy emphasized that it is designed to be supportive rather than replace clinicians. "The AI component is going to have to be more of a guardrail, supporting the clinician rather than replacing their functions," she noted. "Reducing administrative burdens and gathering information is where it’s going to be most beneficial."

Reducing adverse drug events (ADEs) is another crucial area where the AI tool can have a positive impact, especially in preventing acute kidney injury (AKI), which is a common and costly outcome of antimicrobial treatment. Roy highlighted that March 24 marked National Adverse Drug Event Awareness Day, underscoring the fact that ADEs are among the leading causes of death in the United States.

"Adverse drug events are the fourth or some references say the third leading cause of death in the US," Roy said. "And many of them are related to medication. You can imagine why it's an important cause for us to address. Our tool allows for the reduction of adverse events, specifically with antimicrobials."

Roy explained that one of the most critical ADEs, AKI, can lead to a cascade of complications, from prolonged hospital stays to the need for renal replacement therapy. "You can imagine the healthcare costs associated with any acute kidney injury, starting from an increased length of stay to patients ultimately needing renal replacement therapy," she said.

In addition to addressing adverse events, Roy also discussed the importance of real-world evidence in patient dosing, particularly for pediatric populations. Roy, who has a background in pediatric critical care, emphasized the challenges unique to treating pediatric patients, who range from infants with immature organ function to adolescents nearing adulthood.

"We are going to be presenting a poster at ESCMID this year, and that’s looking at a special cohort: pediatrics," she said. "This is special to me because I’m a pediatric critical care-trained pharmacist, so that’s where my practice was as a clinician. I know the challenges in that patient population. You have patients ranging from birth to adolescence to 18-year-olds, so you're looking at a spectrum of patients with very immature organ functions to adulthood."

The upcoming poster presentation at ESCMID will highlight the impact that Bayesian dosing has on target attainment in the pediatric cohort, using real-world data to guide clinical practice. "What we’re going to be presenting at ESCMID is our data and real-world evidence on the pediatric cohort and the impact that Bayesian dosing has in target attainment," Roy shared. "It’s all about guiding practice using real-world data."