Single instead of triplicate intraocular pressure measurements in dogs do not substantially lower accuracy and precision but do slightly reduce statistical power

Abstract

Abstract OBJECTIVE To compare single and triplicate applanation tonometry values across previous intraocular pressure (IOP) studies in dogs. ANIMALS 116 ophthalmologically normal dogs. PROCEDURES Triplicate IOP readings (n = 1432) from studies evaluating effect of anesthetic protocols were analyzed to estimate a range of probable differences between averaged triplicate and first, averaged and lowest, and first and lowest IOPs. The decrease in variability with triplicate measurements and the magnitude of effects on statistical power were quantified. RESULTS The 2.5th to 97.5th interpercentile range for differences of averaged triplicate values minus first IOP readings was –3 to 2.7 mm Hg; for averaged minus lowest: 0 to 3.7 mm Hg; for first minus lowest: 0 to 5 mm Hg. The 95% prediction interval for differences in study group means (n = 160 groups, n = 5 to 11 eyes per group) based on averaged minus first measurements was –1.0 to 0.9 mm Hg with associated SDs reduced by 4% on average. Analysis of previous studies using averaged instead of first IOP values resulted in minimal decreases in SEs of 3–9% (0.03 to 0.09 mm Hg). Of 11 comparisons found significant with averaged data, 2 (18%) were found nonsignificant with first measurements. Of 96 comparisons found nonsignificant with averaged data, 3 (3%) were found significant with first measurements. CLINICAL RELEVANCE With applanation tonometry in ophthalmologically normal dogs, no clinically meaningful difference was found between the first, lowest, or averaged triplicate IOP measurements, but the first reading has a larger variance and hence will result in lower statistical power.