How could a high loss to follow-up of those in the HIGH FAT DIET group potentially affect the estimate of the measure of association for high dietary fat intake and hypertension?