Glaucoma Screening is Affected by Length Bias

Abstract

Purpose:: Patients detected by a periodic screening programme tend to have less severe and less progressive disease as compared to patients detected in another way. This phenomenon has been named "length bias". Aim of the present study was to investigate whether there exists a length bias in glaucoma screening.

Methods:: Participants aged >=55 years from the prospective population-based Rotterdam Study underwent at baseline (1991-1993) and at follow-up (1997-1999) the same ophthalmic examination, including visual field testing and simultaneous stereo optic disc photography. 87 out of 3842 participants developed glaucoma during a mean follow-up time of 6.5 years. Six out of these 87 incident glaucoma cases had to be excluded. We compared glaucoma cases diagnosed before the follow-up examination to those who remained undetected until the follow-up examination.

Results:: Of the 81 incident glaucoma cases, 23 were detected before the follow-up examination, whereas 58 patients remained undetected until the follow-up examination. The undetected cases showed more often glaucomatous optic neuropathy without glaucomatous visual field loss (31 of 58) as compared to the detected cases (4 of 23; P=0.001). Of the 27 (58-31) undetected cases with visual field loss, only 4 had developed significant loss in their better eye.

Conclusions:: Glaucoma screening is affected by a length bias. Expected benefits of a screening programme in addition to case finding as performed currently should be tempered accordingly.

Keywords: clinical (human) or epidemiologic studies: biostatistics/epidemiology methodology