Data Analysis

Data were exported from iCare LINK software for analysis in Microsoft Excel (Microsoft Corp., Redmond, WA), IBM SPSS Statistics for Windows version 24.0 (IBM Corp., Armonk, NY), and GraphPad Prism version 7.03 for Windows (GraphPad Software, La Jolla, CA). One eye of each patient was included for the group analysis to avoid between-eye correlation bias.²⁹ For patients undergoing monocular topical treatment, the treated eye was included; otherwise, an eye was randomly selected. For patients who completed training and certification for both eyes and had data that met the inclusion criteria, a post hoc subanalysis was also performed to examine the correlation between eyes. If multiple measurements were recorded at one time (n = 161 [4.25%]), only the last consecutive measurement was included for analysis because of specific instructions to patients to repeat measurements only, if the original measurement was not performed correctly. Measurements outside the iCare HOME's range were recorded by the instrument as “<5” and “>50” mmHg and excluded from analysis (n = 92 [2.43%]), as well as outliers (n = 6 [0.16%]) identified as outside 2.2× the interquartile range for individual patients.³⁰ Patients with an insufficient number of measurements were excluded from the analysis (n 39 over the study period and n 14 in any week).

Descriptive statistics were used to report the demographic characteristics of the study population and patient-reported instrument usability. Data were presented as mean and SD unless otherwise indicated. Intraocular pressure data were binned into time intervals using (1) local clock time and (2) time of the patient's first measurement (waking time) as reference. Binning was optimized to provide even distribution of intraocular pressure measurements across time intervals and minimize variability within individual bins. To calculate diurnal intraocular pressure fluctuation, each patient's intraocular pressure data were individually analyzed to obtain mean intraocular pressure for each binned time interval. The difference between mean values was defined as the diurnal IOP fluctuation. Patient compliance was calculated by dividing the number of measurements successfully recorded (and included for analysis) by the number of measurements possible if protocol was followed (i.e., four times per day).

Shapiro-Wilk normality test was used to confirm the distribution of continuous data. For normally distributed data including age, baseline, and mean intraocular pressure, the independent t test was used to compare glaucoma and glaucoma suspect groups. Paired and unpaired t tests were applied to compare intraocular pressure variance between the two methods of binning stated above (clock time vs. patient waking time), for five-hourly and two-hourly intervals, respectively. To examine changes in intraocular pressure on consecutive days following treatment, analysis of variance and Tukey multiple-comparisons test were applied. For non–normally distributed data such as the extent of diurnal intraocular pressure fluctuations, Mann-Whitney U test was used to compare differences between groups. Comparison of categorical data including sex and ethnicity was performed with the χ² test. Pearson correlation was calculated between diurnal intraocular pressure (within binned time intervals) measured in the first day/s compared with overall mean diurnal intraocular pressure (of all measurements obtained over the study period). It was also used to examine the relationship between changes in intraocular pressure measured by applanation tonometry and iCare HOME following glaucoma treatment and the effect of central corneal thickness on these measurements. Analysis of variance was applied to test the effect of laterality on the variability of the IOP measurements. In a subanalysis of patients with intraocular pressure data from both eyes, we examined intraclass correlation of intraocular pressure measurements between eyes and also performed paired t test to compare the variability of right and left eye measurements within patients.