Evaluation of the pneuma-tonometer

Intraocular pressure in 125 randomly selected patients visiting the eye clinic at the Massachusetts Institute of Technology for routine care was measured with a Goldmann applanation tonometer and a model 30R Pneuma-Tonometer. Pairs of reading were taken with each instrument when possible, and each trace of the Pneuma-Tonometer measurement was subsequently read blind and independently by two observers. The standard deviation of the differences between the two readers' values obtained on 416 traces was 0.44 mm Hg. The standard deviation of the differences between paired Pneuma-Tonometer readings on 186 eyes was 1.75 mm Hg. With data selected for cause the mean Pneuma-Tonometer reading was 0.40 mm Hg higher, the standard deviation of the differences between the instruments was 1.50 mm Hg, and the correlation coefficient was 0.88. Alternatives to the customary use of linear regression methods of assessing calibration are proposed.