Mastering the creation of an accurate and standardized master formulation record (MFR) for compounded medications is essential for PharmD students. Elizabeth Hageman, PharmD, faced challenges in providing her students the chance to repeatedly practice MFR creation, obtain immediate feedback, and build on their learning.
“Facilitating this with faculty would demand countless hours that we simply don’t have,” stated Dr. Hageman, a clinical assistant professor of pharmacy practice and assistant director of skills education at Binghamton University School of Pharmacy and Pharmaceutical Sciences in New York. “I started to wonder if it was feasible to train an AI program to evaluate and review MFRs.”
This idea became an ideal project for Dayanjan (Shanaka) Wijesinghe, PhD, an associate professor in the Department of Pharmacotherapy and Outcomes Sciences at Virginia Commonwealth University School of Pharmacy in Richmond. An expert in health precision methods focusing on digital technologies, Dr. Wijesinghe guides pharmacy students in developing and implementing digital health applications in real-world settings.
With Dr. Wijesinghe’s support, pharmacy students Dorjan Leka and Talal AlMuzaini partnered with Dr. Hageman to utilize OpenAI’s GPT-4 foundational model in creating a customizable AI tool. This tool evaluates students’ MFR creation in alignment with USP General Chapter <797>. It features a 10-section grading rubric that includes ingredients and calculations, compounding steps, and beyond-use dates and storage conditions, with the AI “teaching assistant” programmed with detailed instructions for awarding or deducting points in each section.
For instance, in the container closure system section, students are penalized for mentioning the container and size without specifying the type. Along with their grade, students receive in-depth feedback from the chatbot, such as, “Overall, your MFR is well-structured with comprehensive details in most sections. Improvements could be made in specifying dosage forms and clearer unit conversions in the name and ingredients section. Additionally, providing specific quantities in the compounding steps could enhance clarity and precision. Great job on the references and quality control procedures, ensuring the safety and efficacy of the compounded medication.”
The tool was finalized in April, and Dr. Hageman intends to implement it with one of her student groups in the upcoming academic year. This project is just one of several AI-driven educational tools that Dr. Wijesinghe has developed with his digital health APPE students over recent years. “We noticed that pharmacy students were lacking sufficient experiential education opportunities before advancing to their P4 rotations,” he remarked. “Our attempts to fill this gap have led to the creation of various AI tools to enhance students’ clinical skills.”
Developing Pharmacy Counseling Skills
One of the initial applications they created, launched in June 2023, is a conversational AI that enables students to refine their pharmacy counseling skills. According to Dr. Wijesinghe, the AI convincingly portrays a patient with a unique persona and health concerns, making students believe they are interacting with a real person. These exchanges allow students to integrate various areas of knowledge into patient counseling—a skill that requires practice.
The AI has been programmed to display behaviors typical of real patients, such as confusion if the student is unclear or overly complex in their explanations, as well as misconceptions regarding their medications. After the conversation concludes, the AI evaluates the student’s performance, assessing skills such as empathy, question quality, response accuracy, and use of jargon. Following this feedback, students can request further guidance on how to improve their performance.
“Limited opportunities exist for us to offer students to practice these skills in a classroom environment,” Dr. Wijesinghe noted. “This approach allows independent practice outside the classroom, enabling students to improve at their own pace.”
Dr. Wijesinghe and his students have developed a range of AI tools for pharmacy education, including applications for practicing SOAP (subjective, objective, assessment, and plan) notes, pharmacogenetic counseling, vaccine recommendations, medication reconciliation, and geriatrics pharmacy counseling. They have also designed an innovative Socratic tutoring tool for case-based learning, presenting students with patient scenarios and prompting them to suggest diagnoses and appropriate tests.
“The AI doesn’t give answers directly; rather, it encourages critical thinking, allowing students to formulate their own responses based on provided information and test results,” Dr. Wijesinghe explained. “Each interaction is unique, offering diverse learning experiences.”
Similar to many of their projects, this was developed collaboratively with Ashim Malhotra, BPharm, PhD, a professor of pharmacology and assistant dean of academic affairs and accreditation at California Northstate University College of Pharmacy in Elk Grove. The success of these initial tools paved the way for their latest initiative, the MFR assessment tool created with Dr. Hageman.
AI Tutors Rated on a Scale
Dr. Hageman emphasized that AI pharmacy tutors do have their limitations. In pharmacy compounding, she noted, “there are many nuances” that AI may not capture. For instance, when applying Chapter <797> in practice, “there are some gray areas regarding acceptable data, such as including additional ‘miscellaneous’ information to ensure a compound is dispensed with a filter. However, other elements, such as the name of a compound, determining the concentration of base drugs, and the actual compounding steps, are straightforward.”
She also pointed out that students and faculty should avoid becoming overly dependent on AI. “Drug information is constantly evolving, and there may be delays in how quickly new information is incorporated into the AI’s algorithm,” she cautioned. “While it’s a fantastic tool, it doesn’t replace clinical decision-making or diminish the role of faculty in teaching. We must utilize these tools in a manner that supports our instructional efforts.”
‘Future-Proof’ Design
The AI tools created by Dr. Wijesinghe and his team are designed to be flexible and “future-proof,” enabling them to function across various platforms and integrate different AI models, including open-source options. This ensures adaptability as the field rapidly advances.
“It would be detrimental to become locked into a single vendor, especially since there’s no definitive leader in the market yet. This design allows our programs to easily switch between different AI models on the back end,” Dr. Wijesinghe noted. Such flexibility keeps the tools updated and reduces costs while enhancing performance over time as more advanced AI models emerge.
These AI tools are freely available to pharmacy schools and students on a web-based platform called PharmtutorAI.com, with more applications to be launched soon. They can also operate in multiple languages, including Spanish, Arabic, and Mandarin. “Anyone around the world can access these applications to practice their skills,” he said. “Since we’re not monetizing this, we’re providing it as a service to improve pharmacy education globally.”
Located in the Bronx, New York, Mediserv Pharmacy is committed to supporting the community’s health and wellness. As a local pharmacy, we prioritize personalized care and provide access to the latest advancements in pharmaceutical education and technology. We embrace innovative tools that enhance pharmacy practice and ensure that our pharmacists are equipped with the best resources to serve our customers effectively.
Explore how we combine traditional pharmacy services with cutting-edge technology to better meet your healthcare needs at Mediserv Pharmacy.
Call MediServ Pharmacy for prescription support, medication guidance, compounding questions, and local Bronx pharmacy care.
Call (718) 395-4000