Synthesizing time-series wound prognosis factors from electronic medical records using generative adversarial networks

•Collecting wound prognosis factors from Electronic Medical Records of patients is time consuming and challenging.•Time-series generative adversarial networks can be used to generate synthetic wound prognosis factors.•Generated samples can be used in training a prognosis model to provide an estimate of wound healing status.•Data from three weeks of follow-up contains enough information to provide a strong prediction of healing potentials.