The complexities of individualized medication dosing for critically ill patients, particularly those in intensive care units (ICUs), present significant challenges for clinicians. Sharmeen Roy, PharmD, BCPS, chief strategy and science officer at DoseMe, highlights the need for real-time adaptive dosing models that integrate patient data to optimize care.
Critically ill patients experience rapid physiological changes, which can significantly impact drug metabolism and clearance,“You and I can be given the same medication, such as a beta-lactam antibiotic, and our bodies would process it differently due to physiological differences,” Roy explained. “This variability is even more pronounced in ICU patients, where factors such as kidney function and drug absorption can change from hour to hour.”
To address this, DoseMe is developing models that continuously adapt to a patient’s clinical status by integrating real-time patient data. This approach aims to ensure that dosing decisions are tailored to the individual rather than relying on population-based estimates, which may not account for diverse patient conditions.
Limitations of Clinical Trials and the Importance of Real-World Evidence
One of the key barriers to individualized dosing is the reliance on clinical trial data. Roy notes that while FDA-approved medications undergo extensive testing, clinical trials often exclude diverse patient populations, limiting the generalizability of their findings.
“In clinical trials, patients are carefully selected based on inclusion and exclusion criteria, which means real-world patients, such as pediatric populations, elderly individuals, and those on extracorporeal membrane oxygenation (ECMO) or continuous renal replacement therapy (CRRT) may not be represented,” Roy says.
Without robust data on how medications behave in these populations, clinicians are left with limited guidance for individualized dosing. Real-time adaptive models aim to bridge this gap by providing personalized dosing recommendations based on real-world patient data rather than solely relying on controlled trial results.
AI and the Future of Individualized Dosing
Artificial intelligence (AI) is playing an increasing role in healthcare, particularly in clinical decision support systems. Roy acknowledges both the potential and the challenges of integrating AI into real-time dosing models.
“AI can help support clinicians by analyzing large datasets and identifying patterns that would be difficult to discern manually,” Roy explained. “However, in healthcare, patient data privacy and compliance are paramount. The integration of AI must be transparent and controlled, ensuring that clinicians understand the recommendations being made and that clinical judgment remains central to the decision-making process.”
The role of AI in antimicrobial stewardship is another important aspect, particularly in the fight against AMR. By optimizing dosing strategies, AI-driven models can help ensure that antibiotics are used effectively, reducing the risk of resistance development. Roy emphasized that AI should not function as an autonomous system but rather as a tool that enhances clinical decision-making while maintaining necessary safety parameters.
“We need to avoid the ‘black box’ problem, where clinicians are unsure how AI-generated recommendations are made,” Roy said. “AI in dosing should be implemented in a way that provides transparency and allows for continuous monitoring and clinician oversight.”
