Evaluating the use of a temperature sensor to monitor spectacle compliance in warm versus cold climates.

BACKGROUND: This study investigates the utility of a temperature sensor data logger to monitor spectacle compliance for future application in research and clinical settings. Specifically, the question of whether warm versus cold climates negatively impact accuracy of the sensor to monitor spectacle wear is investigated. METHODS: Fifty adults from Houston, Texas (summer) and 40 adults from Columbus, Ohio (winter) wore a thermosensor on their spectacles for one week while keeping wear-time logs. Temperatures during reported spectacle wear (ON) were compared to temperatures during non-wear (OFF) between sites. Two methods to approximate wear time were evaluated by percent error with respect to subject-reported wear time. Method 1 filtered temperatures, classifying the range of 28.4 to 35.2°C as wear. Method 2 utilised examiners interpreting temperature versus time plots. Separate analysis of periods of reported outdoor wear was performed to identify the percentage of time examiners correctly identified wear. RESULTS: Group mean ON temperatures did not differ between sites (p = 0.72), but group mean OFF temperatures were significantly warmer in Houston (Houston: 24.7 ± 2.0°C, Columbus: 20.3 ± 2.1°C; p < 0.0001). Median percent error of the filtering technique to approximate subject reported wear time was 4 per cent for Houston and -8 per cent for Columbus. Median percent error for examiner 1: Houston 1 per cent, Columbus 0 per cent; median percent error for examiner 2: Houston 3 per cent, Columbus 0 per cent. Houston outdoor wear was correctly identified 88 and 97 per cent of the time by the examiners versus 79 and 81 per cent for Columbus. CONCLUSION: Despite environmental temperature differences, measured temperatures during spectacle wear were similar across subjects and median percent error was less than 10 per cent for both wear time approximation methods. The device studied was effective for objectively monitoring spectacle wear in both warm and cold climates with the caveat that subjects spent the majority of time indoors.

Objective measurement of spectacle wear with a temperature sensor data logger.

PURPOSE: This study seeks to establish the utility of the SmartButton Data Logger (www.acrsystems.com) to monitor spectacle wear for research and clinical applications. METHODS: Fifty adults wore a thermosensor on their spectacles for 2 weeks for each of two mount types while keeping wear-time logs. Temperatures during reported spectacle wear (ON) were compared to temperatures during non-wear (OFF) with repeated measures analysis of variance (ANOVA). In addition, two strategies to approximate spectacle wear from temperature data were evaluated: (1) Filtering data based on temperature ranges to identify spectacle wear (either group mean ON temperature, or an individual's mean ON temperature), and (2) Separate examiners inspecting temperature against time plots to identify spectacle wear. The success of these methods to approximate wear time was evaluated by per cent error with respect to subject reported wear time. RESULTS: Group mean ON (31.8 [0.6]°Celsius [°C]) and OFF (24.7 [1.5]°C) temperatures differed significantly (F1,47  = 471.2, p < 0.001), but there was no difference in temperature between mounts (F1,47  = 1.9, p = 0.18). Median per cent error and first and third quartiles (Q1, Q3) of each technique used to approximate wear time were: group mean filtering = 8% (Q1 3%, Q3 18%), individual mean filtering = 7% (Q1 4%, Q3 19%), Examiner 1 = 6% (Q1 2%, Q3 14%), Examiner 2 = 7% (Q1 3%, Q3 12%). CONCLUSIONS: The SmartButton can monitor spectacle compliance in patients with all approximation methods evaluated providing less than 10% median per cent error in wear time.