Using ChatGPT Prompts to Enhance Clinical Data Analysis Through Coding

By Jen A. Miller (ADLM)
Posted on 29 Jun 2025

Clinical laboratorians are good at many things: communication, critical thinking, problem solving, and selecting the best possible testing solutions for patients and the healthcare professionals who treat them.

But what they may not be good at — and most people aren’t — is coding. Why would a clinical laboratorian need to know coding languages like Python or R?

Image: Generative AI allows clinical laboratorians without coding experience to perform complex analyses on large volumes of data (photo courtesy of ADLM)

“People are good with collecting the data but then not working with the data,” said Isaiah Mensah, PhD, clinical chemistry fellow in the department of pathology and immunology at Washington University School of Medicine in St. Louis.

In his roundtable session at ADLM 2025 (formerly the AACC Annual Scientific Meeting & Clinical Lab Expo), Mensah will show attendees how to leverage generative artificial intelligence (AI) to jump over the coding hurdle, allowing them to access data-driven insights that were previously inaccessible to many in the field.

“We’re beginning to tackle questions that we couldn’t answer before,” said Mensah. Generative AI has made answering those questions easier for those who have a lot of data but don’t have a coding background — like clinical laboratorians.

Mensah will perform a live demonstration using real, anonymized patient data sets. He will prompt ChatGPT to obtain Python and R code and then apply that code to analyze complex clinical data. He will also lead a group discussion to guide attendees who want to conduct similar analyses. In fact, he encourages participants to bring their computers to work with the group on their own data (without any personally identifiable information).

“My goal is to share data sets with them and hopefully we can go through using ChatGPT prompts to ask specific questions,” Mensah said. “I’ll have different ways of asking for codes using different prompts, then give [attendees] some guidelines on how to maximize prompts.” He will also discuss how to format data optimally. Otherwise, it’s “garbage in, garbage out.”

Learning to perform complex analyses is critical for laboratorians because clinical labs collect a high volume of data. “At my institution, we have over one million specimens per year,” Mensah said. “We take data on age, sex, analytes, and results. We produce the most data, so it only makes sense that we will be able to ask questions about the data that we collect, and we’ll be able to answer them.”

The session will also be of interest to healthcare professionals who don’t regularly handle large data sets, including physicians, Mensah said. “We all have access to data one way or the other,” he said. “It’s so helpful for people to know how to use the power of AI to improve their data analysis.”