Purpose
Quality assurance study to determine the statistical consistency of repeat masked IOP measurements in the context of a prospective study of patients under treatment for mild to moderate glaucoma.
Methods
Case control study. Subjects: 20 unique pairs of 4th year optometric interns (26 total interns, 9 participating in more than one pairing) performed applanation tonometry with the masked dial randomly pre-set between 5 and 40 mmHg on the first 10 adults enrolled in an IRB-approved SLT study. Each pair of interns performed 3 pre-randomized right eye IOP measurements at 20 total study visits. Intra- and inter-observer variability were assessed by repeat measures 1-way ANOVA.
Results
IOP measurements (n=120) ranged from12-34mmHg (mean 20.0 ± sem 0.6mmHg). The mean range of intra-observer measurements was 4.6±0.7mmHg. The mean difference between paired observers for the 3 averaged measurements was 1.6 ±0.7mmHg. This was essentially equivalent to the mean difference between the first-measured IOP value of the 3 (1.9 ±1.1mmHg, p=0.7). Furthermore, the mean of all 6 IOP measurements was not significantly different from the mean of the first measured of each examiner (p = 0.3). Inter-observer differences were independent of the IOP level, but presumed tonographic decrease in IOP from the first measurement to the sixth was significant (-3.3 ±1.0mmHg; P=0.01; see figure). Dial preset position showed no association with the measured IOP (R2 = 0.0001).
Conclusions
Attempts at improving objectivity and accuracy in IOP measurement by masked and randomized applanation tonometry is not improved by multiple readings. Tonographic artifact would likely be reduced by limiting each masked reader to a single IOP measurement.