Patient data that was populated as follows: The statistical distribution of curated data from the MIMIC-III dataset. Data based on these distributions was generated for 19 parameters. The data generated for these parameters was fed into a virtual patient model to predict a further 4 parameters: PaO2, PaCO2, pH, and HCO3. The outputs from the virtual patient was appended to the original generated data and can be used to train neural networks as inputs and expected outputs.