Evaluation of Bias Towards Medical Professionals in Large Language Models