ABSTRACT
Simulation provides a risk free and controllable environment for training of healthcare providers. The limited realism of available simulators and training programs impedes immersive training in obstetric emergencies. In developed countries, intrapartum monitoring in high-risk cases involves continuous evaluation of foetal heart rate and uterine contractions signals. We present an essential component of a high-fidelity simulator for normal and critical situations in labour and delivery, namely an intrauterine pressure generator. The signal model behind the generator consists of a truncated Gaussian curve with the programmable features: amplitude, frequency, duration, and resting tone. Through analysis of 44h of physiological data, we demonstrate that the natural variability of these features and of the baseline pressure can be approximated by deterministic trends and stationary stochastic processes. Signal parameters can be controlled by simulation instructors, scripts, or other models to reflect different patients, pathologies, and evolving clinical situations. Twelve 40-min tracings reflecting three different patients in labour were presented to three clinical experts, who attributed similar realism scores to simulated and to real tracings.
