Inappropriate gold standard bias in cervical cancer screening studies