Even-number measurement bias with Goldmann applanation tonometry in patients with glaucoma and glaucoma suspects.

Goldmann applanation tonometry is widely used for the diagnosis and management of glaucoma and its use is considered standard of care. However, the precision of this method may be reduced by a clinician tendency to round to even numbers. Studies have previously demonstrated an even-number measurement bias with Goldmann applanation tonometry during examination of a general patient population. Since it has not been determined whether this bias persists among glaucoma suspects and patients with glaucoma, further investigation was conducted. A retrospective analysis was conducted on a random sample from a large dataset of >69,000 patients seen during a six-year period at an urban, academic primary eye care service. Patients without suspicion of glaucoma, patients with a suspicion of glaucoma, and patients with glaucoma were selected who had Goldmann tonometry performed. Chart reviews were performed to confirm status, and even/odd-numbered Goldmann tonometry measurement frequencies were compared. The analysis included 961 controls, 506 glaucoma suspects, 159 ocular hypertensives not taking medication, and 314 patients taking medications who carried a diagnosis of glaucoma or ocular hypertension. Among controls the Goldmann tonometry even/odd digit proportions were 62.8%/37.2% (N = 961, p < 0.0001), and the even-number bias persisted among the other groups with specific even/odd distributions being 61.9%/38.1% (N = 506, p < 0.0001) for glaucoma suspects not taking medications, 66.0%/34.0% (N = 159, p < 0.0001) for ocular hypertensives not taking medications, and 64.3%/35.7% (N = 314, p < 0.0001) for glaucoma/ocular hypertension patients taking medications. An even-number measurement bias with Goldmann tonometry may be prevalent even when the examiner is aware of there being greater importance for intraocular pressure measurement accuracy. [ABSTRACT FROM AUTHOR]