Measuring wearing times of glasses and ocular patches using a thermosensor device from orthodontics

Purpose: Amblyopia is one of the most common visual disorders in children. The risk of severe visual impairment on the healthy eye is doubled in patients with amblyopia. If detected early enough, the chances of visual rehabilitation are good. Treatment consists of refractive correction and occlusion of the dominant eye. Patient compliance is an important factor and can be monitored using thermosensors. It was the goal of our study to give proof of the principle that the wearing times of glasses and patches can be measured using a comparatively small and commercially available microsensor. Methods: Agreement between wearing times protocols of ocular patching/refractive correction and temperature measurements of thermosensors attached to the patches or glasses of three individuals were analysed using the Bland–Altman method. It was also analysed whether blinded persons could distinguish between temperature curves of patches and glasses, or temperature curves of an incubator or while worn in a pocket. Results: The temperatures picked up by the microsensors indicate the beginning and the end wearing times of either glasses or ocular patches through steep temperature difference and a distinct temperature curve during measurements. Although blinded test persons were able to cleary distinguish between temperature profiles from incubator/pocket measurements compared to glasses/patching, glasses and patching curves could be discriminated correctly in only 50%. Differences between wearing time protocols and temperature measurements were within the limits of agreement as stated by the Bland–Altman plots. Conclusion: The TheraMon® microsensor can reliably measure wearing times of glasses and ocular patches without making the wearer uncomfortable, although the data are not unquestionable, especially in higher surrounding temperatures. Further studies on a larger number of individuals with different wearing profiles are needed. Amblyopia is one of the most common childhood visual disorders with a prevalence of 5–8% and doubles the lifetime risk of visual impairment on the healthy eye (Rahi et al. 2002). It is often associated with refraction anomalies, strabism or other anatomical circumstances, for example congenital cataract (Gusek-Schneider 2011). Because of the difference in visual input, the cortical picture of the weaker eyes is suppressed during the development of a child. This leads to visual impairment without an anatomical correlate. This has an impact not only on the visual but also on the sociological and socio-economical development possibilities of a particular child in as much as he/she is unable to participate in all activities (e.g. certain sports) or apply for certain vocations that require good binocular vision (Rahi et al. 2002). It is generally assumed that the therapy has to start early (Hubel & Wiesel 1970) although single case reports of therapeutical successes in older patients are not a rarity. This suggests a certain cortical visual plasticity in older patients (Chen et al. 2003). The treatment of amblyopia starts with an optimal refractive correction and a period of refractive adaptation. This alone can resolve 30% of cases (Stewart et al. 2011). Further treatment possibilities are intermittent patching of the dominant eye or penalization with atropine drops (Stewart et al. 2011). The rationale is to partly suppress the dominant eye so that a cortical visual impairment of the weaker eye does not occur. New options of reducing vision of the dominant eye are constantly being developed (Spierer et al. 2010). All the therapy options depend on compliance with both the patients and his/her parents. However, an objective monitoring of refraction or occlusion therapy was not possible until recently. With the introduction of the occlusion dose monitor (ODM) in 1994, a tool was available to monitor wearing times of patches (Fielder et al. 1994). This ocular dose–response monitor did not record temperature differences. Since then, we have learned that compliance is a problem even if parents and children know that their wearing times are being objectively monitored (Fielder et al. 1994, 1995; Simonsz et al. 1999; Stewart et al. 2004, 2007). The ocular ODM is under constant development. A more recent version measures temperature differences thus indicating wearing times using two temperature sensors (Fronius et al. 2006). Its disadvantages are size, battery, data storage, moisture in the controlled area of experiment, and the overall evaluation process is toilsome for the researchers. It also displayed a lack of accuracy in temperatures between 33 and 37°C (Fronius et al. 2006, 2009). This led to the development of an advanced ODM that has a longer battery life and a data storage with more capacity (Chopovska et al. 2005). It is an interesting fact that objective measurement of the compliance to refraction therapy in the treatment of amblyopia has only recently been addressed and has not been published, yet in a peer reviewed journal (Maconachie et al. 2012). Therefore, we found it interesting to look for a monitoring device that is small enough to fit the earpiece of glasses, which would better enable researchers to monitor the wearing time of glasses. At the same time, this device would have strong advantages when being used on patches as well because it would be smaller than the ODM. The TheraMon® microsensor is 9 × 13 mm in size and can easily be attached to the earpiece of glasses as well as on the inside of eye patches without causing discomfort (Fig. 1). It is a well evaluated and safe tool for measuring wearing times of orthodontic devices (Schott & Göz 2011; Schott et al. 2011) as it is encapsulated with polyurethane making it water proof and causing no skin irritation. A software was specifically developed to evaluate wearing times. Photography of the TheraMon® microsensor (A) attached to the temple (B) while worn by a test person (C) on the back of an ocular occlusion patch, and (D) while worn by a test person. Note that the microsensor is hardly visible. In this study, we tried to principally prove that TheraMon® microsensors can measure wearing times of glasses and of patches and that the microsensors can be used for monitoring compliance in the ophthalmological field. We also tested if it would be possible to discriminate temperature transitions when the sensors were attached to glasses or patches versus being carried in a bag or box or pocket. Commercially available TheraMon® microsensors and the TheraMon® reading station were lent by the Handelsagentur Gschladt (Hargelsberg, Austria). TheraMon® microsensor measures 9 × 13 × 4.5 mm in size and samples in regular intervals of 15 min with a storage capacity of 100 days and a battery capacity of 2 years. The microsensors use radio frequency identification technology (RFID) to send data to the reading station; the antenna can only be activated if positioned near the reading station (2–3 cm). Sampling intervals are fixed at 15 min. Temperature sensitivity was set to 0.2°C. Different sampling intervals can be chosen, between 1 and 4096 seconds and different levels of accuracy resulting in different data storage times. For measuring wearing times of glasses, microsensors were attached to the earpieces of the glasses of test persons (Fig. 1A,B) using commercially available orthodontic separating rings with a diameter of 5 mm (Dentaurum GmbH & Co. KG, Ispringen, Germany). In order to monitor patching time, microsensors were attached on the inside of commercially available occlusion patches (Piratoplast®; Dr. Ausbüttel & Co. GmbH, Witten, Germany) just above the inferior margin of the orbita without applying pressure (Fig. 1C,D). Three individuals from our group tested TheraMon® microsensors on glasses and patches on themselves. The sensor measures 96 data points per 24-hr period and is inactivated between measurements. Measurements were performed over 2 months. A wearing time diary was kept by all individuals, and wearing times were written down as accurately as possible. For the TheraMon® microsensors, a sudden increase in temperature close to the body temperature was defined as the starting point of the wearing time and a sudden drop in temperature as the end-point. In case of high surrounding temperatures or a starting point that was not able to be clearly measured by body temperature changes, sinusoid temperature curves were used to identify application time. For inconclusive data of the microsensors, 0 min was entered as the wearing time. Test persons were from our study group and they were fully aware of the implications of this study. The importance of keeping a good wearing times diary was known. Test persons were asked to specifically report discomfort while wearing the device. Temperature measurements were also performed in an incubator (Biocenter 2001, Salvis-Lab™; Renggli AG, Rotkreutz, Switzerland) at 37°C or while being carried in a pocket. Ten blinded test persons were asked to compare the temperature curves and to determine if the microsensor was worn on patches/glasses and whether it was kept in an incubator or in a pocket. Data were logged by connecting the microsensors to the reading station via antenna at 2–3 cm distance from the antenna. A USB cable transferred the data to a PC for storage; wearing times were evaluated using the TheraMon® Software and compared to the wearing times protocol. Mean differences of wearing times between protocol and as measured by the microsensors were calculated and compared to their means as described previously by Altman & Bland (1983). Ocular patching and wearing times of glasses were looked at as overall wearing times and separately. For statistical analysis, jmp® 8 Statistical Discovery Software (SAS, Institute, Cary, NC, USA) was used and the 95% limits of agreement were estimated by mean difference ±2 standard deviations from the differences. All individuals returned their TheraMon® microsensors after a 2-month period with their wearing times manual. The group consisted of two females (29 years) and one male (30 years). The test persons were at any time blind to the recorded microsensor data, the wearing times protocol; the temperature measurements evaluation were performed independently by two different persons as was the subsequent statistical analysis. In order to reveal surrounding temperature as a factor potentially influencing wearing times, measurements were performed under mild temperature conditions with a mean temperature of 22.6°C (n = 25 measurements; Tmin of 20°C and Tmax of 24°C) and under warmer conditions with an average surrounding temperature of 26.8°C (n = 17 measurements; Tmin of 24°C and Tmax of 32°C). Over a 2-month period, wearing times of patches or glasses were protocolled as 9556 min, while the microsensors detected 8891 min (Table 1); the mean wearing time was 137.0 min (protocol) and 133.3 min (microsensors) per day. There was an observed mean difference in overall wearing times of −5.7 min (95% CI −21.2 to 9.8, Fig. 2A). For ocular patching, the mean difference between measured wearing times of the microsensor and the wearing protocol was −21.2 min (95% CI −72.8 to −21.3, Fig. 2B). For glasses, the mean difference was −4.4 min (95% CI −25.1 to 16.3, Fig. 2C). All data are shown in Fig. 2. Inconclusive data of the microsensors had to be noted as 0-min wearing time in five single measurements (Fig. 2). Bland–Altman Plots for original data respecting distortion of results. Mean of the differences of wearing times of single measurements (wearing time protocol – microsensor measurement) and limits of agreement mean differences ± 2× s (s = standard deviation of differences) for (A) overall wearing times (B) glasses, and (C) ocular patching. The dotted red lines represent 95% limits of agreement. The red lines represent the mean difference of the measurements. The test subjects did not report any discomfort when wearing either patches or glasses. No reaction of the microsensors to the glue of the patch was noticed. As shown in Fig. 3A, it was always possible to distinguish the beginning- and end-point of the wearing time, with the exception of the five inconclusive measurements. Different temperature transitions (e.g. pocket, or box, or worn above 30°C) showed different temperature profiles (Fig. 3B,C). Temperature measurements in an incubator at 37°C (Fig. 4A,B) or while being carried in a pocket (Fig. 4C,D) could be distinguished from ocular patching/glasses in 100% of blinded persons, as well as temperature transitions while worn in a pocket. They were insecure trying to distinguish temperature profiles of glasses and patching in 50%. Original data recording from the detailed analysis function of the TheraMon® software for a limited time interval of 4 days; white and gray background marks 24 hr. Y-axis indicated temperature in °C. Different background temperatures (20 and 27°C) are shown. (A) shows an overview, (B) Shows a detailed temperature curve of patching, (C) shows a detailed temperature curve during refractive correction. Note the steep in- and decrease clearly indicating wearing time as shown in (B) and (C). Note also the sinusoid temperature character that only occurs during – even rather short term – application of glasses or patches; characteristically, this is more pronounced during patching (B) than during refractive correction (C). Original data recording from the detailed analysis function of the TheraMon® software showing temperature curves of (A), (B) an incubator, and (C, D) pocket temperatures. Note the characteristic temperature profile of the single measurements (B, D). The goal of this study was to show well-documented wearing time protocols as the standard method agree sufficiently with the wearing times measured by the TheraMon® microsensors. Moreover, we tried to principally prove that TheraMon® microsensors are a reliable, comfortable and safe tool for measuring wearing times of glasses and patches and can be applicable for compliance and dose–response studies. We also tested if it will be possible to discriminate accidental temperature changes from occlusion or refractive therapy. For the analysis of method comparison studies, neither the correlation coefficient nor techniques such as regression analysis are appropriate, instead we used the well-established Bland–Altman method for analysis (Altman & Bland 1983; Bland & Altman 1986, 1995, 2003; Grouven et al. 2007). Our results show that the temperature measurement methods (wearing time protocol and microsensor) agree sufficiently enough to replace wearing time protocols in the future studies using the microsensors. Because of the high correlation and reasonably large sample size, the confidence intervals for the limits of agreement are narrow (Fig. 2A–C). Even the occasional large differences between sensors and protocols in Fig. 2A,B did not influence these results. The inconclusive data were either measurements that did not show a steep temperature (three measurements) or did display multiple dips and inconclusive temperature curves (two measurements) and neither influenced agreement of both measurement methods. The fact that those inconclusive data did not strongly influence the confidence intervals of the narrow limits of agreement further support the reliability of the sensors. Our data are supported by earlier studies using the same or a similar method in orthodontics (Ackerman et al. 2009; Schott et al. 2011) and by studies in ophthalmology using ODM, with the difference that the TheraMon® microsensor does not measure temperature difference, as does the ODM (Fronius et al. 2006, 2009; Loudon et al. 2009; Gharebaghi et al. 2011; Roefs et al. 2012), thus being more susceptible to manipulation. However, this is the first study to correctly state that wearing time diaries and temperature measurement are indeed interchangeable. Earlier studies in ophthalmology have applied ODM with great success but were describing the procedure of data collection and the data evaluation as laborious and time consuming (Stewart et al. 2004, 2007; Fronius et al. 2006, 2009), moreover, battery capacity did sometimes require home visits from the study team to exchange ODM devices (Fronius et al. 2009). The software from the microsensors will enable us to define a certain occlusion time in the future and has the data automatically evaluated and easily visualized (Schott et al. 2011). This might be an interesting option for children and parents undergoing occlusion/refraction therapy in the future as a compliance supporting tool and should be considered in the clinical routine. In this study, all the test persons came from the same study group and knew the implications of this study as well as the importance of good record keeping. Therefore, results of this study can only be carefully transferred to the clinical situation and no comments on compliance can be made based on it. Wearing time diaries of this selected group of test persons can and should be considered as very reliable, as shown before (Fronius et al. 2009). Results of analysis of agreement should be considered more valuable. Moreover, it has to be mentioned that in the clinical setting, for example occlusion is measured by the hour. Although ODM development has improved and led to smaller ODMs with more potential wearing comfort, the TheraMon® microsensors are significantly smaller and therefore can be considered more comfortable (Chopovska et al. 2005). Still, one slight limitation in wearing comfort is the height of the sensor with 4.5 mm regarding patching of the eye. In cases of discomfort, the battery could be transferred to the outside through a hole in the occlusion patch thus reducing the contact height of 0.5 mm. Because batteries are designed to last over 24 months, in the future, a slight reduction in size is thinkable as well. The polyurethane encapsulation prevents skin irritation upon direct contact. The part of the microsensor in direct contact with the skin is plain and therefore should not be a danger to children or influence the correct position of glasses (Fig. 1A,B). The long lasting battery is an advantage of the new microsensor: it could be re-used for different tests and over longer testing periods. Another limitation would be outside temperature higher than 33°C resulting in less accuracy of the measurements. For future application, as a compliance monitoring device, we think it is rather difficult to fake compliance by using a heater or wearing it in a pocket, because of the mentioned characteristic sinusoid temperature curve while correctly applied. This is supported by our results from blinded test persons being able to clearly distinguish between real wearing temperature measurements and artificial measurements. Nevertheless, the ‘sinusoid character’ of the curves could also be caused by slight temperature fluctuations either of the body or in the environment, although this temperature curve has been described for ODM in the past (Chopovska et al. 2005) and for use in orthodontics (Schott et al. 2011). We also think that it will be important for the first studies on patients that single measurements should be checked for plausibility. The microsensors or temperature profile of the microsensors could not distinguish between ocular patching and wearing times of glasses. Therefore, for future studies, it would be advisable to have different microsensors for patching and refractive adaption. Thus, ocular patching or refractive therapy can securely and easily be distinguished. Additionally, it would be wise to set the sampling intervals to 5 min. It has to be further pointed out that the new microsensors with only one thermosensor will display a lack of accuracy between temperatures of 33 and 37°C, like the ODM (Fronius et al. 2006, 2009). In conclusion, it has to be stated that the data from the microsensors are not unquestionable. It is unlikely that the RFID technology would cause any impact on the person wearing the microsensor because the antenna is only activated 2–3 cm near the reading station. So far, no negative reports considering safety issues exist. The results of this study give the proof of principle that TheraMon® microsensor technique is applicable to the ophthalmological field and that they are as valuable as using wearing time protocols. We would like to thank Regina Hofer for her support and David Schilling, Oliver Borst, Andreas Schmidt and Sascha Hoffmann for fruitful discussions and valuable advice while preparing the manuscript. We also like to thank Gerhard Gschladt for providing the microsensors and his enthusiasm for the project and Gershom Spengler for editorial assistance. Kai Januschowski would especially like to thank the department of ophthalmology of the hospital Nuestra Senora de la Candelaria and the team of Tocuyo for their incredible support.