Purchase this article with an account.
C.C. Teng, J.A. Young, S. Reddy, J. Huang, E. Nissan; Determining the Minimum Number of IOP Measurements Necessary to Establish a Baseline IOP: A Statistical Approach . Invest. Ophthalmol. Vis. Sci. 2006;47(13):4467.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The ability to set a therapeutic goal intraocular pressure (IOP) for a patient with glaucoma or ocular hypertension is predicated on an understanding of the patient’s baseline IOP. Thus, the appropriate determination of baseline IOP is essential. Although an accurate baseline IOP can be determined by obtaining a very large number of IOP measurements, this is impractical because a large number of examinations would represent a significant burden on both the clinician and patient and would result in a significant delay in therapy. We propose a statistical method to determine the minimum number of IOP measurements necessary to obtain an accurate baseline IOP.
IOP’s of normal eyes were measured on consecutive visits using a single Goldmann applanation tonometer by a single examiner. Individual IOP measurements from a single eye were sequentially evaluated until a 95% confidence interval could be established about the sample mean within two levels of sampling error: +/– 1 mm Hg and +/– 2 mm Hg. The process was repeated for each of the 20 eyes in the study group. The number of IOP measurements required to satisfy the conditions of the 95% confidence interval was recorded for both levels of sampling error.
20 eyes of 10 patients were enrolled. The larger sampling error of +/– 2 mm Hg yielded estimates in all 20 eyes examined. The number of measurements required to satisfy the conditions of the 95% confidence interval with a sampling error of +/– 2 mm Hg ranged between 2 and 7 with a mean of 2.8±1.1. When the mean baseline IOP estimates were compared to the means of the larger sample, the larger sample mean was correctly predicted in all 20 cases (100%). The narrower sampling error of +/– 1 mm Hg yielded estimates in 12 eyes examined. The number of measurements required to satisfy the conditions of the 95% confidence interval with a sampling error of +/– 1 mm Hg ranged between 2 and 7 with a mean of 3.2. When the mean baseline IOP estimates were compared to the means of the larger sample, the larger sample mean was correctly predicted in 7 cases (58%). When predictions in the narrower sampling error group were constrained to at least 3 measurements, 10 of 10 were predicted correctly (100%).
The number of IOP measurements required to establish a baseline depends upon the variance of IOP readings of any particular eye, an important consideration for the clinician. Our method yields a minimum number of necessary measurements to establish a baseline IOP within a sampling error of +/– 2 mm Hg achievable within the constraints of clinical practice.
This PDF is available to Subscribers Only