Methodological variations in estimating apparent progressive visual field loss in clinical trials of glaucoma treatment