FCW Wu - The Journal of Clinical Endocrinology & Metabolism, 2012 - academic.oup.com
Study design The retrospective observational design carries a substantial risk of primary
selection bias due to the nonrandom assignment of exposure (in this case, T treatment). The …