Text this: Benchmarking Multiple Large Language Models for Automated Clinical Trial Data Extraction in Aging Research