Abstract
Purpose:
This study aims to determine if temperature sensors mounted on spectacles can accurately indicate wear time.
Methods:
Fifty adults wore a temperature sensor on their spectacle temple for at least 10 hours for one week. Sensors were mounted with adhesive or custom silicone sleeves (randomized) and recorded temperatures every five minutes. Subjects documented wear time on log sheets. Mean temperatures indicating spectacle wear were determined by comparing subject logs to temperature recordings. Wear time was then approximated based on temperature data alone by two strategies: 1) filtering for each individual’s mean (± 1.5 sd) ON temperatures and 2) filtering for the group’s mean (± 3.4°C) ON temperature. Wear time was also approximated by two masked examiners independently analyzing temperature vs time plots with data viewing software. Approximated wear time was compared to subject reported wear time and percent error calculated for each approximation method.
Results:
ON/OFF temperatures differed significantly (p < .0001) and varied with gender (p=0.04), but not mount type (p = 0.18) (repeated measures ANOVA). Wear time approximations were within 80% agreement of subject reported wear time in 82-86% of subjects for the two examiners and in 76-82% of subjects for temperature filtering strategies. For subjects with <80% agreement, the magnitude of errors were smaller for examiners than temperature filtering strategies.
Conclusions:
Wear time was better approximated by examiners than current temperature filtering strategies and demonstrated good ability of sensors to quantify spectacle wear time. Analytical software will be developed, incorporating examiner criteria, to automate calculation of wear time from temperature plots.