Simulator development using natural language: clinician-led innovation through artificial intelligence

Healthcare simulation training faces significant barriers due to the "clinician-developer gap," where educators lack programming expertise to create customized digital simulators. Natural Language-Driven Development (NLDD) is an emerging paradigm that enables clinicians to develop educational technology through conversational artificial intelligence interfaces.

We implemented NLDD methodology to develop Open Vent Sim, a comprehensive mechanical ventilation simulator designed to replace anesthesia machines and ventilators in educational contexts lacking dedicated equipment. A multidisciplinary team comprising anesthesiologists, residents, a research nurse, IT, and biomedical engineers collaborated using Google AI Studio to iteratively create a web-based application through natural language prompts. Development proceeded through conversational cycles in which clinical requirements were translated into functional code via large language model assistance.

Open Vent Sim was successfully developed in about 40 h over two weeks, featuring three simulation environments: anesthesia workstation, ICU ventilator, and high-flow oxygenation systems. The simulator incorporates physiological patient profiles (normal, ARDS, COPD) with dynamic compliance calculations and realistic waveform generation. Clinical validation was achieved through the integration of continuous resident feedback during iterative development. The application was successfully implemented in SimZone 1 as an interactive skill trainer and in SimZone 2 for team-based clinical scenarios during formal anesthesia and critical care education. Significant technical adaptation was required to transform the AI-generated prototype into a production-ready application.

NLDD demonstrates the potential to democratize the creation of educational technology by empowering clinical domain experts to develop sophisticated simulation tools without traditional programming expertise. This approach addresses resource limitations while maintaining clinical authenticity, though professional technical oversight remains essential for production-ready implementations.