AI in Investigational Drug Services: Promise, Pitfalls, and Practical Uses

The healthcare landscape is evolving at an incredible pace, and technology continues to redefine how we deliver patient care. From electronic health records to automated dispensing systems, pharmacists have long been at the forefront of adopting tools that enhance accuracy, efficiency, and safety.

Now, artificial intelligence (AI) is stepping into the spotlight—offering new possibilities across healthcare. But what does this mean for the highly specialized world of Investigational Drug Services (IDS)? As a pharmacy resident working within IDS and exploring AI’s role in healthcare, I’ve been reflecting on how we can responsibly integrate AI into research pharmacy practice.

IDS pharmacists operate at the intersection of clinical research, patient safety, and regulatory compliance. Every day involves managing intricate protocols, ensuring strict adherence to policies, and maintaining meticulous records. Tools like Vestigo have already transformed how we handle protocol management, inventory, and documentation—streamlining workflows while maintaining compliance.

But as AI becomes more accessible, it’s natural to wonder: Could AI further support IDS operations? The answer is yes—but with caution.

AI isn’t about replacing pharmacists—it’s about augmenting our capabilities. Here are a few areas where AI could offer meaningful support in IDS:

• • Policy Development & Maintenance
    AI could assist in drafting and updating standard operating procedures (SOPs), helping teams stay aligned with evolving regulatory requirements by analyzing guidance documents and suggesting revisions.
  • Research Protocol Review
    Imagine AI tools scanning protocols to flag inconsistencies, highlight missing elements, or suggest alignment with institutional standards—saving valuable time during initial reviews.
  • Data Analysis & Workflow Insights
    With AI-driven analytics, IDS teams could identify trends across protocols, such as common deviations or workload bottlenecks, leading to proactive process improvements.
  • Inventory Management Optimization
    Predictive models could assist in forecasting investigational product needs, reducing waste while ensuring availability—always under pharmacist oversight.

These are just some of the possibilities available to us, but they come with important caveats as well.

While AI offers promise, it also carries risks that are particularly critical in healthcare settings:

• • Bias
    AI models learn from existing data, which can inadvertently embed biases. In the context of IDS, this could lead to skewed suggestions if the AI isn’t trained on diverse, representative datasets.
  • Hallucinations
    AI tools—especially large language models (LLMs)—can sometimes generate information that sounds plausible but is incorrect or entirely fabricated. In a field where precision is non-negotiable, this is a serious concern.
  • Brittleness
    AI systems may perform well in familiar scenarios but fail unpredictably when faced with nuanced or uncommon situations—something frequently encountered in complex research protocols.

These risks underscore why AI should be viewed as an assistant, not an authority. Human expertise, especially in patient safety and regulatory compliance, remains of the utmost importance.

Prompt Engineering: A Key to Safe AI Use

One practical way to mitigate AI risks is through effective use of an emerging practice called prompt engineering—the art of crafting clear, specific instructions when interacting with AI tools.

For example:

Instead of asking an AI:
“Summarize this research protocol.”
Try:
“Summarize this research protocol by focusing only on pharmacy-relevant sections, such as drug handling procedures and storage requirements. Do not generate information beyond the provided text.”

Some general tips:

• • Be specific about what you want as well as what you don’t want.
  • Set boundaries (e.g., “Only use information from this document.”)
  • Review outputs critically, never assuming AI-generated content is flawless.

By guiding AI thoughtfully, we can reduce errors and ensure it serves as a helpful tool rather than a source of misinformation.

As AI continues to evolve, so too will its applications in pharmacy practice. For IDS, the potential to streamline processes, enhance compliance efforts, and unlock data-driven insights is within reach. However, we must ensure we integrate AI responsibly and carefully consider all risks before diving into integration into our practice settings.

Pharmacists, with our deep commitment to accuracy and patient safety, are well-positioned to lead this balanced approach. By staying informed, embracing innovation carefully, and applying best practices like prompt engineering, we can harness AI’s strengths while safeguarding the integrity of our work.

The future of IDS isn’t about choosing between human expertise and technology. It’s about combining them to advance clinical research safely and efficiently.

Evan Kelley, PharmD

PGY2 Investigational Drugs and Research Pharmacy Resident
Atrium Health Wake Forest Baptist

Dr. Kelley is a 2^nd year pharmacy resident in Investigational Drugs and Research (IDR) at Atrium Health Wake Forest Baptist in Winston-Salem, NC, where he is training to be a well-rounded research pharmacist and clinician. The IDR pharmacy residency is a combined two-year program, with the first year consisting of a variety of clinical rotations to give residents a well-rounded clinical foundation. The second year of the program focuses on providing a diverse exposure to the aspects of the research process, from areas such as regulatory and compliance, study review and implementation, and operational aspects of IDS pharmacy itself. During his time with Wake, he has worked closely with research nurses, study coordinators, study physicians, primary and co-investigators, and IDS pharmacy staff throughout his training and has learned to become a flexible yet integral part of the research team.

Kelley has interests in critical care and pediatrics, largely influenced by previous clinical experience in both the pediatric and neonatal intensive care unit settings. He also has a growing interest in hematologic malignancies based on time spent in the Phase 1 clinical trial field as well as experience with Wake’s hematopoietic stem cell transplant service. Kelley recently presented his second-year research project “Evaluation of the transition from the use of PO to IV pre-medications in an outpatient oncology setting” at the Hematology/Oncology Pharmacy Association’s (HOPA) 2025 annual conference.

Kelley earned his PharmD from the University of Georgia College of Pharmacy in 2023 and completed his PGY1 training at Atrium Health Wake Forest Baptist in 2024. He is an active member of HOPA, the American Society of Health-System Pharmacists (ASHP), and the Society of Research Clinicians.