Evaluating the Feasibility and Accuracy of Large Language Models for Medical History-Taking in Obstetrics and Gynecology