This study aimed at detecting any differences in BT measurements between the old gold-standard AXLMER thermometer and the new digital and infrared thermometers in a hospital pediatric setting in Albania where mercury thermometers were still allowed despite the Minamata Convention due to a transition phase related to the in-country resources and capacities13. This study could not be currently conducted since on 26 May 2020, the Government of Albania deposited its instrument of ratification, becoming the 120th Party to the Minamata Convention. For this reason, the window of opportunity exploited in this study has been closed and the mercury thermometer cannot longer be used in Albania resulting in beneficial effects for the environment, children, and community.
Following the ban on mercury, in the last 10 years numerous environment-friendly devices have been introduced into clinical settings in many countries around the world to replace mercury thermometers, and a great amount of research has been conducted to explore their validity and reliability16-18. Some recent meta-analyses carried out to compare peripheral electronic devices (e.g. AXLDGT, FHDIR, and TYMIR) with central devices (e.g. bladder, esophageal, and rectal) indicated that the peripheral devices showed poor accuracy when used to estimate core BT and inadequate sensitivity when used for fever detection in adults and children8,9,22. However, using a different research methodology, other authors, comparing the peripheral electronic devices, achieved contradictory evidence32-35. For example, some studies found that FHDIR thermometers could serve as a good alternative to AXLDGT due to their user-friendliness and speed of use36, while others, reporting great mean differences between the investigated devices, did not consider the FHDIR device as accurate as the AXLDGT thermometer24,34,35.
The research approaches in the above-mentioned studies had two distinct goals: comparing peripheral devices with central ones and exploring the level of agreement among the alternative thermometers. Surprisingly, available evidence has not completely dissolved concerns about the validity and reliability of the new thermometers and have not fully allayed nurses’ doubts emerging from clinical contexts. In fact, when nurses need to make their clinical decisions on the basis of BT values detected with the alternative peripheral devices, they need to know if these new thermometers (e.g. AXLDGT, FHDIR, and TYMIR) are as reliable at least as the old mercury device they used for a long time. To tackle this issue and contribute to the global debate, we conducted this perspective pragmatic research.
In the light of our results, the research hypothesis of no clinically significant differences between the old gold-standard mercury thermometer and the new devices should be partially rejected.
In fact, although moderate to strong significant direct correlations were found between BT values detected by all the alternative devices and those by the AXLMER (r = from 0.623 for TYMIR to 0.844 for AXLDGT), also significant mean differences were found for the paired BT comparisons between each of AXLDGT and TYMIR and the mercury thermometer (-0.04 °C and -0.12 °C, respectively). However, in this case statistically significant differences cannot correspond to a strong clinical significance since it is unlikely that a maximum difference of -0.12 °C can affect some clinical judgements, such as drugs administration or caring interventions. Nevertheless, considering the poor clinical significance of these differences, it would be hazardous to state that the new alternative devices are all equally reliable tools for BT measurement in children.
In fact, beyond the statistical significance, in the visual analysis of difference between measurements (Bland-Altman scatterplots) it is noticeable that the AXLDGT values are the closest to the gold standard’s since 95% of their differences fell within the narrowest range (95% LoA = -0.62 °C to +0.53 °C). In addition, the AXLDGT showed the highest percentage (94.0%) of BT differences within the clinical acceptable value of 0.5 °C8. For these reasons, AXLDGT thermometer should be considered most similar to the old gold-standard mercury thermometer and, in accordance with current guidelines, may be preferable for measuring BT in children3-5, 20.
FHDIR and TYMIR devices showed poor agreement with AXLMER thermometer, exhibiting broader 95% LoAs (-0.92 °C to 0.87 °C and -1.04 °C to 0.81 °C, respectively) and more differences higher than 0.5 °C compared to AXLDGT thermometer. In the light of this findings, it would not be cautious to indicate FHDIR and TYMIR as the first choices for BT measurements in children. However, in pediatric clinical settings where FHDIR and TYMIR are used32,36, we suggest that BT values be validated through clinical observation and AXLDGT thermometer be used in case of discrepancy.
Furthermore, AXLDGT thermometer also showed better performance in screening for fever than FHDIR and TYMIR devices, even if in this study, as in some other studies, all the investigated devices showed moderate to low sensitivity and high specificity in detecting fever8,9,22. In fact, even if to a lesser extent than the other alternative devices, also the AXLDGT thermometer resulted in a higher proportion of false-negative than false-positive readings, and this could be explained by the tendency to underestimate the gold-standard measurements adopted in this study. In practice, also using the AXLDGT thermometer, in a minimal proportion of children found to be non-febrile, fever could not be ruled out with certainty. In this regard, it is always advisable, in accordance with current guidelines, a conservative approach to protect children from missed care, by repeating BT measurements or using rectal thermometers to confirm fever, especially when clinical signs and symptoms contrast with detected BT values5,20.
Practice implication
The results of this study suggest that AXLDGT thermometer may be the best choice for BT measurement in pediatric settings, considering especially that this device adequately balances accuracy, safety, and children’s comfort. Even if there are practical reasons suggesting the use of TYMIR and FHDIR thermometers in pediatric clinical practice, such as their ease of use, speed of measurements and improved hygiene, the results of this study show that these devices cannot be considered interchangeable with the gold standard adopted in this study, and AXLDGT should be preferred, especially if there are clinical doubts. However, considering the performance of AXLDGT in the screening of fever, clinical decisions should not be based exclusively on BT values, but, in accordance with current guidelines, it is always advisable assessing children for the presence or absence of signs and symptoms potentially associable with fever5. Assessing the skin colour and turgor, respiratory function, cardio-circulatory condition, the child’s activity, and the presence of headache, shiver, and nausea provide excellent criteria to confirm or doubt a BT value5,20. This last recommendation should be strongly considered for clinical practice especially during epidemic events, such as the current Coronavirus disease 2019 (COVID-19) pandemic. In this regard, one of the special accommodations made in clinical practice and other contexts is the use of the infrared thermometers. For its ‘no contact’ process aimed at limiting the virus spread, the FHDIR device has become the most widely accepted thermometer in this pandemic. However, considering that in this study fever (≥ 37.5 °C), in about one-third of febrile children could not be detected with the FHDIR device, temperature screening alone should be avoided in every context since it may not be very effective, as also supported by previous evidence9. Those signs and symptoms commonly present in children with Coronavirus disease 2019, such as fatigue, dry cough, and other respiratory symptoms, should be considered along with BT values37,38.
Limitations of the study
The strengths of this study included the adequate sample size of pediatric patients, the measurement of BT in a real clinical setting, the use of axillary mercury thermometers as the gold standard, and the use of appropriate statistical methods for data analysis. However, the results of this study should be accepted bearing in mind the monocentric approach of the study and the differences between core and peripheral BTs.
In conclusion, the results of this study confirmed the AXLDGT device as the best alternative to the AXLMER thermometer in detecting children’s BT both in cases of fever or not.
However, according to current guidelines, when clinical signs and symptoms contrast with detected BT values, it is recommended repeating BT measurements or using rectal thermometers.