In a groundbreaking study published on March 6 in the Journal of Medical Internet Research, researchers at NYU Langone Health have unveiled the potential of generative artificial intelligence (GenAI) in transforming the healthcare software design process. The study focuses on ChatGPT, a GenAI-powered “chatbot” known for its ability to generate human-like text responses and summarize complex materials effectively. This technology’s application to healthcare aims to develop a text messaging-based software program designed to combat diabetes by promoting healthy eating and exercise among patients.
The research team embarked on a project to test the efficacy of AI-facilitated dialogue between doctors and software engineers in accelerating the creation of a Personalized Automatic Messaging System (PAMS) for diabetes management. The results were striking, with ChatGPT enabling eleven evaluators from diverse fields to produce the tool in just 40 hours—a task that previously required over 200 hours of programming by a non-AI-assisted team.
Danissa Rodriguez, PhD, an assistant professor in the Department of Population Health at NYU Langone and a member of the Healthcare Innovation Bridging Research, Informatics, and Design (HiBRID) Lab, led the study. She highlighted the pivotal role of ChatGPT in enhancing communication between technical and non-technical team members, thereby streamlining the software development lifecycle from conceptualization to coding.
The study underscores the sensitivity of generative AI tools and the importance of prompt engineering—the art of crafting questions to elicit specific responses from AI. This skill, blending intuition with experimentation, is crucial in healthcare contexts, where precise communication can dictate the success of software solutions.
Interestingly, the research revealed that physicians and nurses, equipped with their in-depth understanding of medical nuances, are exceptionally positioned to guide the development process without needing to learn coding themselves. By inputting their requirements in plain English into ChatGPT, they could effectively communicate with engineers, bridging the gap between medical needs and technical execution.
Despite the necessity for human developers in the final stages of code generation, the integration of AI into the design process marks a significant acceleration in developing healthcare software. Devin Mann, MD, senior study author, director of the HiBRID Lab, and strategic director of Digital Health Innovation at NYU Langone Medical Center’s Information Technology, emphasized the democratizing effect of chatGPT on healthcare software design. This AI-assisted approach is poised to produce tools that are not only user-friendly and reliable but also adhere to the highest standards of coding quality.
This study signals a new era in healthcare innovation, where the collaboration between medical professionals and AI can lead to the rapid development of advanced, effective, and personalized healthcare solutions.
