Validity of Wrist and Forehead Temperature in Temperature
Transcript Of Validity of Wrist and Forehead Temperature in Temperature
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
1 Validity of Wrist and Forehead Temperature in Temperature Screening in the General 2 Population During the Outbreak of 2019 Novel Coronavirus: a prospective real-world 3 study 4 5 (Validity of Wrist and Forehead Temperature in Temperature Screening in the General 6 Population During the Outbreak of COVID-19) 7 8 Ge Chen1,8, Jiarong Xie2,3,8, Guangli Dai1, Peijun Zheng4, Xiaqing Hu5, Hongpeng Lu2,3, Lei 9 Xu2,3, Xueqin Chen6*, Xiaomin Chen2,7* 10 11 1Department of Clinical Engineering, Ningbo First Hospital, Ningbo, Zhejiang Province, 12 China; 13 2Department of General Internal Medicine, Ningbo First Hospital, Ningbo, Zhejiang Province, 14 China; 15 3Department of Gastroenterology, Ningbo First Hospital, Ningbo, Zhejiang Province, China; 16 4Department of Nursing, Ningbo First Hospital, Ningbo, Zhejiang Province, China; 17 5Department of Emergency, Ningbo First Hospital, Ningbo, Zhejiang Province, China; 18 6Department of Chinese Traditional Medicine, Ningbo First Hospital, Ningbo, Zhejiang 19 Province, China; 20 7Department of Cardiology, Ningbo First Hospital, Ningbo, Zhejiang Province, China; 21 8These authors contributed equally to this work. 22 * Corresponding authors 23 24 Abbreviations: 2019-nCoV, 2019 Novel Coronavirus; NCIT, Non-contact infrared 25 thermometer; IRTT, infrared tympanic thermometers; ROC, receiver–operator characteristic.
1
NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
26 27 Key words: 2019 Novel Coronavirus; wrist temperature; non-contact infrared thermometer; 28
2
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
29 Abstract° 30 Aims: Temperature screening is important in the population during the outbreak of 2019 31 Novel Coronavirus (COVID-19). This study aimed to compare the accuracy and precision of 32 wrist and forehead temperature with tympanic temperature under different circumstances. 33 Methods: We performed a prospective observational study in a real-life population. We 34 consecutively collected wrist and forehead temperatures in Celsius (°C) using a non-contact 35 infrared thermometer (NCIT). We also measured the tympanic temperature using a tympanic
36 thermometers (IRTT) and defined fever as a tympanic temperature ≥37.3°C.
37 Results: We enrolled a total of 528 participants including 261 indoor and 267 outdoor 38 participants. We divided outdoor participants into four types according to their means of 39 transportation to the hospital as walk, bicycle, electric vehicle, car, and inside the car. Under 40 different circumstance, the mean difference ranged from -1.72 to -0.56°C in different groups 41 for the forehead measurements, and -0.96 to -0.61°C for the wrist measurements. Both 42 measurements had high fever screening abilities in inpatients (wrist: AUC 0.790; 95% CI: 43 0.725-0.854, P <0.001; forehead: AUC 0.816; 95% CI: 0.757-0.876, P <0.001). The cut-off
44 value of wrist measurement for detecting tympanic temperature ≥37.3°C was 36.2°C with a
45 86.4% sensitivity and a 67.0% specificity, and the best threshold of forehead measurement 46 was also 36.2°C with a 93.2% sensitivity and a 60.0% specificity. 47 Conclusions: Wrist measurement is more stable than forehead measurement under different 48 circumstance. Both measurements have great fever screening abilities for indoor patients. The 49 cut-off value of both measurements was 36.2°C. (ClinicalTrials.gov number: NCT04274621) 50
3
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
51 Introduction
52
The outbreaks of 2019 novel coronavirus COVID-19 (previously known as 2019-nCoV)
53 has attracted global attention, due to its strong transmission ability and certain fatality (1, 2).
54 Some studies reported that fever, fatigue and dry cough are common symptoms of
55 COVID-19 patients (3, 4), and 43.8% of the patients showed fever before admission with it
56 largely being the first symptom (5). Therefore, temperature screening in the high-risk
57 population is important for early identification of COVID-19 infection and thereby reducing
58 the risk of cross infection.
59
During the epidemic, infrared tympanic thermometers (IRTT) and non-contact infrared
60 thermometer (NCIT) are being applied to temperature screening in the general population (6).
61 As a screening tool, it is quick for mass screening and allows a faster triage (7). However, we
62 need to consume a lot of disposable plastic covers when we use IRTT. It may increase the
63 financial burden in the widespread use of population screening. Furthermore, indirect
64 contacts with infected individuals may increase the risk of cross infection. NCIT meets the
65 clinical requirements for mass screening in terms of detection efficiency, safety and
66 cost-performance. Besides, it takes less time than IRTT. Forehead is one of the key targets of
67 thermography. However, forehead temperature is affected by physiological and
68 environmental conditions (8). It should be measured in a relatively temperature-controlled
69 environment. A previous study suggested to acclimate to the indoor temperature for at least
70 10 min for those who were exposed to the cold before taking body temperature readings (8).
71 However, it is not practical for mass screening in winter during the outbreak of COVID-19.
72
Wrist temperature in this outbreak is under consideration. Before testing, they just need
73 to roll up their sleeves at 10 cm above the palmar side of the wrist. Considering this area is
74 covered with clothing, the wrist temperatures may keep stable. Previous study showed
75 wearable devices (WD) on the wrist were applied in temperature monitoring in clinical
4
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
76 practice (9). It brings a challenge whether it can be used as an accurate, safe and
77 cost-effective screening tool in this outbreak.
78
In this study, we explored the accuracy and advantages of wrist temperature
79 measurement in a real-life population in different environments and conditions. We aimed to
80 find the thresholds of this key technique for diagnosis of fever. It may assist to improve the
81 standardization of both practical use and performance, especially indispensable in the
82 pandemic 2019-nCoV situation.
83
84
5
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
85 Materials and Methods
86 Study population
87
This was a prospective observational study in a real-life population. We consecutively
88 enrolled a total of 572 participants at Ningbo First Hospital in China in this study (Figure 1).
89 The exclusion criteria included: (i) Age ≤ 18 years (n = 6); (ii) Wearing hearing aid, or
90 having a cerumen (n = 7); (iii) Participants with soft tissue infection or trauma (n =3); (iv)
91 Missing data of wrist, forehead, and tympanic temperature (n = 4); and (v) Participants whose
92 forehead temperature measurements showed “low” (n = 23). We finally enrolled 528 eligible
93 participants for the final analysis, including 261 indoor and 267 outdoor participants. The 261
94 indoor participants were from the fever clinic and emergency department, and the 267
95 outdoor participants included patients and accompanying family members. The data of indoor
96 participants were collected consecutively between February 14th and February 20th, 2020.
97 The data of outdoor participants were collected on February 14th, 15th, 17th, 2020.
98 Temperature readings were taken by trained and experienced nurses. Each participant was
99 measured for wrist, forehead, and tympanic temperature twice. The temperatures were
100 recorded by mean wrist temperature, forehead, and tympanic temperature, respectively. Data
101 regarding age, gender, transportation, occupation, and temperature were recorded
102 immediately by the nurse to pre-printed files.
103
The study was approved by Ningbo First Hospital Ethics Committee. All participants
104 were asked verbally. They gave their oral informed consent in this study. The study was
105 registered in ClinicalTrials.gov with identifier number: NCT04274621.
106
107 Assessment of environment
108
Indoor patients at the fever clinic and emergency department were those who has been
109 indoors for at least a few minutes. The outdoor participants were divided into four type
6
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
110 according to their means of transportation to the hospital as walk, bicycle/electric vehicle, car,
111 and inside the car.
112
113 Measurement of temperature
114
Tympanic temperature was measured using IRTT (Braun ThermoScan PRO 6000). Wrist
115 and forehead temperature were measured using NCIT. The NCIT was ranged 32.0–42.9°C.
116 The accuracy was ± 0.2°C. NCIT measurements were taken following the manufacturer's
117 instructions in the mid-forehead and a region at 10 cm above the palmar side of the wrist.
118 After pulling the pinna backward, the nurse inserted IRTT into the external auditory meatus.
119 The probe was held in the same position until the “beep” was heard. Temperature readings
120 were taken by the same trained nurse in the following order: forehead, forehead (the second
121 time), left wrist, right wrist, left tympanic, and right tympanic. The data were recorded by
122 another researcher in pre-printed files. Tympanic membrane is in close proximity to the
123 hypothalamus and the internal carotid artery (10). Thus, tympanic temperature is considered
124 to directly reflect core temperature (11), and was defined as the gold standard in this study.
125 These thermometers were stabilized before measurements. Calibration of thermometers were
126 checked by the Quality and Technology Supervision Bureau, Ningbo, China. It was
127 according to Calibration Specification of Infrared Thermometers for Measurement of Human
128 Temperature (JJF1107-2003).
129
130 Statistical analysis
131
Power calculation was performed for sample size. The following parameters were used: a
132 power of 90%, an α-error level of 0.05, estimating a standard deviation of 1°C and a potential
133 allowable error of 0.2°C. Considering a 10% possibility of dropouts and otherwise missing
134 data, at least 293 subjects were needed in our study.
7
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
135
Continuous variables were expressed as mean ± standard deviation (SD), and categorical
136 data in frequency and proportion. The agreements for each method versus
137 tympanic temperature were analyzed by Bland–Altman analysis (12). It also showed three
138 superimposed horizontal lines. Red dashed line highlighted mean bias among all the paired
139 measurements. Black dashed line marked upper and lower 95% Limits of Agreement (LoA).
140 A temperature deviation of 0.5°C was considered as clinically acceptable (13). A tympanic
141 temperature of ≥ 37.3°C was defined as the cut-off point for fever. Statistical analyses were
142 conducted using R version 3.5.1 (The R Foundation for Statistical Computing, Vienna,
143 Austria).
144
145
8
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
146 Results
147 Participants
148
In this prospective observational study, a total of 528 participants were enrolled. Figure 1
149 summarizes characteristics of the participants. The mean age was 46.7 ± 16.4 years. 69.4% (n
150 = 297) of participants were males, and 78.2% (n = 413) were patients (Table 1). Mean
151 forehead, wrist, tympanic measurements were 35.6 ± 1.2°C, 35.7 ± 0.8°C, and 36.6 ± 0.6°C,
152 respectively. There were 44 patients with fever in indoor patients. The data of outdoor
153 participants were collected on February 14th, 15th, 17th, 2020. Mean weather temperatures
154 were 13°C, 14°C, and 7°C, respectively.
155
156 Bland-Altman comparison among the participants under different environment
157
Table 2 showed mean temperatures and Bland-Altman analysis among the participants
158 under different environment. Compared with tympanic temperature as golden standard, the
159 mean difference ranged from -1.72 to -0.56°C for the forehead measurement, and -0.96 to
160 -0.61°C for the wrist measurement. We observed a lower variation in wrist than forehead
161 temperature measurements.
162
Outdoor participants were divided into four types as walk, bicycle or electric vehicle, car,
163 and inside the car. For those who walked, the agreement limits for wrist and tympanic was
164 between -2.05 and 0.34°C; -4.07 and 0.64°C for forehead and tympanic (Figure 2A, B). For
165 those who used bicycle or electric vehicle, the agreement limits for wrist and tympanic was
166 between -2.14 and 0.93°C; -3.82 and 0.84°C for forehead and tympanic (Figure 2C, D). For
167 those who were transported by car, the agreement limits for wrist and tympanic was between
168 -1.43 and -0.44°C; -1.47 and -0.36°C for forehead and tympanic (Figure 2E, F). For those
169 who were inside the car, the agreement limits for wrist and tympanic was between -1.54 and
170 -0.15°C; -2.41 and 0.16°C for forehead and tympanic (Figure 2G, H). It highlighted that wrist
9
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
171 temperature had narrower 95% limits of agreement than forehead. Wrist measurements had
172 the higher percentage of differences falling within ± 0.5°C than forehead measurements in
173 these four types.
174
For indoor patients, the agreement limits for wrist and tympanic was between -2.70 and
175 -0.77°C; -1.91 and 0.80°C for forehead and tympanic (Figure 3). 57.1% of forehead values
176 were included within ± 0.5°C, followed by wrist values (41.4%). We also explore the
177 agreement of left and right wrists (Figure S1). The mean bias is 0.00. The agreement limits
178 for wrist and tympanic was between -0.74 and 0.74°C. It showed good agreement between
179 right and left wrists.
180
181 The receiver–operator characteristic (ROC) curves for detection of fever
182
We performed a ROC curves in indoor patients for detecting tympanic temperature
183 ≥37.3°C. Figure 4 shows the comparison between wrist and forehead measurements for
184 detection of fever. Both measurement had significantly great abilities of screening patients
185 for fever (wrist: AUC 0.790; 95% CI: 0.725–0.854, P <0.001; forehead: AUC 0.816; 95% CI:
186 0.757–0.876, P <0.0001). The cut-off value of wrist measurement for detecting tympanic
187 temperature ≥37.3°C was 36.2°3 with a 86.4% sensitivity and a 67.0% specificity. And the
188 best threshold of forehead measurement was also 36.2°6 with a 93.2% sensitivity and a 60.0%
189 specificity.
190
10
1 Validity of Wrist and Forehead Temperature in Temperature Screening in the General 2 Population During the Outbreak of 2019 Novel Coronavirus: a prospective real-world 3 study 4 5 (Validity of Wrist and Forehead Temperature in Temperature Screening in the General 6 Population During the Outbreak of COVID-19) 7 8 Ge Chen1,8, Jiarong Xie2,3,8, Guangli Dai1, Peijun Zheng4, Xiaqing Hu5, Hongpeng Lu2,3, Lei 9 Xu2,3, Xueqin Chen6*, Xiaomin Chen2,7* 10 11 1Department of Clinical Engineering, Ningbo First Hospital, Ningbo, Zhejiang Province, 12 China; 13 2Department of General Internal Medicine, Ningbo First Hospital, Ningbo, Zhejiang Province, 14 China; 15 3Department of Gastroenterology, Ningbo First Hospital, Ningbo, Zhejiang Province, China; 16 4Department of Nursing, Ningbo First Hospital, Ningbo, Zhejiang Province, China; 17 5Department of Emergency, Ningbo First Hospital, Ningbo, Zhejiang Province, China; 18 6Department of Chinese Traditional Medicine, Ningbo First Hospital, Ningbo, Zhejiang 19 Province, China; 20 7Department of Cardiology, Ningbo First Hospital, Ningbo, Zhejiang Province, China; 21 8These authors contributed equally to this work. 22 * Corresponding authors 23 24 Abbreviations: 2019-nCoV, 2019 Novel Coronavirus; NCIT, Non-contact infrared 25 thermometer; IRTT, infrared tympanic thermometers; ROC, receiver–operator characteristic.
1
NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
26 27 Key words: 2019 Novel Coronavirus; wrist temperature; non-contact infrared thermometer; 28
2
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
29 Abstract° 30 Aims: Temperature screening is important in the population during the outbreak of 2019 31 Novel Coronavirus (COVID-19). This study aimed to compare the accuracy and precision of 32 wrist and forehead temperature with tympanic temperature under different circumstances. 33 Methods: We performed a prospective observational study in a real-life population. We 34 consecutively collected wrist and forehead temperatures in Celsius (°C) using a non-contact 35 infrared thermometer (NCIT). We also measured the tympanic temperature using a tympanic
36 thermometers (IRTT) and defined fever as a tympanic temperature ≥37.3°C.
37 Results: We enrolled a total of 528 participants including 261 indoor and 267 outdoor 38 participants. We divided outdoor participants into four types according to their means of 39 transportation to the hospital as walk, bicycle, electric vehicle, car, and inside the car. Under 40 different circumstance, the mean difference ranged from -1.72 to -0.56°C in different groups 41 for the forehead measurements, and -0.96 to -0.61°C for the wrist measurements. Both 42 measurements had high fever screening abilities in inpatients (wrist: AUC 0.790; 95% CI: 43 0.725-0.854, P <0.001; forehead: AUC 0.816; 95% CI: 0.757-0.876, P <0.001). The cut-off
44 value of wrist measurement for detecting tympanic temperature ≥37.3°C was 36.2°C with a
45 86.4% sensitivity and a 67.0% specificity, and the best threshold of forehead measurement 46 was also 36.2°C with a 93.2% sensitivity and a 60.0% specificity. 47 Conclusions: Wrist measurement is more stable than forehead measurement under different 48 circumstance. Both measurements have great fever screening abilities for indoor patients. The 49 cut-off value of both measurements was 36.2°C. (ClinicalTrials.gov number: NCT04274621) 50
3
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
51 Introduction
52
The outbreaks of 2019 novel coronavirus COVID-19 (previously known as 2019-nCoV)
53 has attracted global attention, due to its strong transmission ability and certain fatality (1, 2).
54 Some studies reported that fever, fatigue and dry cough are common symptoms of
55 COVID-19 patients (3, 4), and 43.8% of the patients showed fever before admission with it
56 largely being the first symptom (5). Therefore, temperature screening in the high-risk
57 population is important for early identification of COVID-19 infection and thereby reducing
58 the risk of cross infection.
59
During the epidemic, infrared tympanic thermometers (IRTT) and non-contact infrared
60 thermometer (NCIT) are being applied to temperature screening in the general population (6).
61 As a screening tool, it is quick for mass screening and allows a faster triage (7). However, we
62 need to consume a lot of disposable plastic covers when we use IRTT. It may increase the
63 financial burden in the widespread use of population screening. Furthermore, indirect
64 contacts with infected individuals may increase the risk of cross infection. NCIT meets the
65 clinical requirements for mass screening in terms of detection efficiency, safety and
66 cost-performance. Besides, it takes less time than IRTT. Forehead is one of the key targets of
67 thermography. However, forehead temperature is affected by physiological and
68 environmental conditions (8). It should be measured in a relatively temperature-controlled
69 environment. A previous study suggested to acclimate to the indoor temperature for at least
70 10 min for those who were exposed to the cold before taking body temperature readings (8).
71 However, it is not practical for mass screening in winter during the outbreak of COVID-19.
72
Wrist temperature in this outbreak is under consideration. Before testing, they just need
73 to roll up their sleeves at 10 cm above the palmar side of the wrist. Considering this area is
74 covered with clothing, the wrist temperatures may keep stable. Previous study showed
75 wearable devices (WD) on the wrist were applied in temperature monitoring in clinical
4
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
76 practice (9). It brings a challenge whether it can be used as an accurate, safe and
77 cost-effective screening tool in this outbreak.
78
In this study, we explored the accuracy and advantages of wrist temperature
79 measurement in a real-life population in different environments and conditions. We aimed to
80 find the thresholds of this key technique for diagnosis of fever. It may assist to improve the
81 standardization of both practical use and performance, especially indispensable in the
82 pandemic 2019-nCoV situation.
83
84
5
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
85 Materials and Methods
86 Study population
87
This was a prospective observational study in a real-life population. We consecutively
88 enrolled a total of 572 participants at Ningbo First Hospital in China in this study (Figure 1).
89 The exclusion criteria included: (i) Age ≤ 18 years (n = 6); (ii) Wearing hearing aid, or
90 having a cerumen (n = 7); (iii) Participants with soft tissue infection or trauma (n =3); (iv)
91 Missing data of wrist, forehead, and tympanic temperature (n = 4); and (v) Participants whose
92 forehead temperature measurements showed “low” (n = 23). We finally enrolled 528 eligible
93 participants for the final analysis, including 261 indoor and 267 outdoor participants. The 261
94 indoor participants were from the fever clinic and emergency department, and the 267
95 outdoor participants included patients and accompanying family members. The data of indoor
96 participants were collected consecutively between February 14th and February 20th, 2020.
97 The data of outdoor participants were collected on February 14th, 15th, 17th, 2020.
98 Temperature readings were taken by trained and experienced nurses. Each participant was
99 measured for wrist, forehead, and tympanic temperature twice. The temperatures were
100 recorded by mean wrist temperature, forehead, and tympanic temperature, respectively. Data
101 regarding age, gender, transportation, occupation, and temperature were recorded
102 immediately by the nurse to pre-printed files.
103
The study was approved by Ningbo First Hospital Ethics Committee. All participants
104 were asked verbally. They gave their oral informed consent in this study. The study was
105 registered in ClinicalTrials.gov with identifier number: NCT04274621.
106
107 Assessment of environment
108
Indoor patients at the fever clinic and emergency department were those who has been
109 indoors for at least a few minutes. The outdoor participants were divided into four type
6
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
110 according to their means of transportation to the hospital as walk, bicycle/electric vehicle, car,
111 and inside the car.
112
113 Measurement of temperature
114
Tympanic temperature was measured using IRTT (Braun ThermoScan PRO 6000). Wrist
115 and forehead temperature were measured using NCIT. The NCIT was ranged 32.0–42.9°C.
116 The accuracy was ± 0.2°C. NCIT measurements were taken following the manufacturer's
117 instructions in the mid-forehead and a region at 10 cm above the palmar side of the wrist.
118 After pulling the pinna backward, the nurse inserted IRTT into the external auditory meatus.
119 The probe was held in the same position until the “beep” was heard. Temperature readings
120 were taken by the same trained nurse in the following order: forehead, forehead (the second
121 time), left wrist, right wrist, left tympanic, and right tympanic. The data were recorded by
122 another researcher in pre-printed files. Tympanic membrane is in close proximity to the
123 hypothalamus and the internal carotid artery (10). Thus, tympanic temperature is considered
124 to directly reflect core temperature (11), and was defined as the gold standard in this study.
125 These thermometers were stabilized before measurements. Calibration of thermometers were
126 checked by the Quality and Technology Supervision Bureau, Ningbo, China. It was
127 according to Calibration Specification of Infrared Thermometers for Measurement of Human
128 Temperature (JJF1107-2003).
129
130 Statistical analysis
131
Power calculation was performed for sample size. The following parameters were used: a
132 power of 90%, an α-error level of 0.05, estimating a standard deviation of 1°C and a potential
133 allowable error of 0.2°C. Considering a 10% possibility of dropouts and otherwise missing
134 data, at least 293 subjects were needed in our study.
7
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
135
Continuous variables were expressed as mean ± standard deviation (SD), and categorical
136 data in frequency and proportion. The agreements for each method versus
137 tympanic temperature were analyzed by Bland–Altman analysis (12). It also showed three
138 superimposed horizontal lines. Red dashed line highlighted mean bias among all the paired
139 measurements. Black dashed line marked upper and lower 95% Limits of Agreement (LoA).
140 A temperature deviation of 0.5°C was considered as clinically acceptable (13). A tympanic
141 temperature of ≥ 37.3°C was defined as the cut-off point for fever. Statistical analyses were
142 conducted using R version 3.5.1 (The R Foundation for Statistical Computing, Vienna,
143 Austria).
144
145
8
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
146 Results
147 Participants
148
In this prospective observational study, a total of 528 participants were enrolled. Figure 1
149 summarizes characteristics of the participants. The mean age was 46.7 ± 16.4 years. 69.4% (n
150 = 297) of participants were males, and 78.2% (n = 413) were patients (Table 1). Mean
151 forehead, wrist, tympanic measurements were 35.6 ± 1.2°C, 35.7 ± 0.8°C, and 36.6 ± 0.6°C,
152 respectively. There were 44 patients with fever in indoor patients. The data of outdoor
153 participants were collected on February 14th, 15th, 17th, 2020. Mean weather temperatures
154 were 13°C, 14°C, and 7°C, respectively.
155
156 Bland-Altman comparison among the participants under different environment
157
Table 2 showed mean temperatures and Bland-Altman analysis among the participants
158 under different environment. Compared with tympanic temperature as golden standard, the
159 mean difference ranged from -1.72 to -0.56°C for the forehead measurement, and -0.96 to
160 -0.61°C for the wrist measurement. We observed a lower variation in wrist than forehead
161 temperature measurements.
162
Outdoor participants were divided into four types as walk, bicycle or electric vehicle, car,
163 and inside the car. For those who walked, the agreement limits for wrist and tympanic was
164 between -2.05 and 0.34°C; -4.07 and 0.64°C for forehead and tympanic (Figure 2A, B). For
165 those who used bicycle or electric vehicle, the agreement limits for wrist and tympanic was
166 between -2.14 and 0.93°C; -3.82 and 0.84°C for forehead and tympanic (Figure 2C, D). For
167 those who were transported by car, the agreement limits for wrist and tympanic was between
168 -1.43 and -0.44°C; -1.47 and -0.36°C for forehead and tympanic (Figure 2E, F). For those
169 who were inside the car, the agreement limits for wrist and tympanic was between -1.54 and
170 -0.15°C; -2.41 and 0.16°C for forehead and tympanic (Figure 2G, H). It highlighted that wrist
9
medRxiv preprint doi: https://doi.org/10.1101/2020.03.02.20030148; this version posted March 6, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.
171 temperature had narrower 95% limits of agreement than forehead. Wrist measurements had
172 the higher percentage of differences falling within ± 0.5°C than forehead measurements in
173 these four types.
174
For indoor patients, the agreement limits for wrist and tympanic was between -2.70 and
175 -0.77°C; -1.91 and 0.80°C for forehead and tympanic (Figure 3). 57.1% of forehead values
176 were included within ± 0.5°C, followed by wrist values (41.4%). We also explore the
177 agreement of left and right wrists (Figure S1). The mean bias is 0.00. The agreement limits
178 for wrist and tympanic was between -0.74 and 0.74°C. It showed good agreement between
179 right and left wrists.
180
181 The receiver–operator characteristic (ROC) curves for detection of fever
182
We performed a ROC curves in indoor patients for detecting tympanic temperature
183 ≥37.3°C. Figure 4 shows the comparison between wrist and forehead measurements for
184 detection of fever. Both measurement had significantly great abilities of screening patients
185 for fever (wrist: AUC 0.790; 95% CI: 0.725–0.854, P <0.001; forehead: AUC 0.816; 95% CI:
186 0.757–0.876, P <0.0001). The cut-off value of wrist measurement for detecting tympanic
187 temperature ≥37.3°C was 36.2°3 with a 86.4% sensitivity and a 67.0% specificity. And the
188 best threshold of forehead measurement was also 36.2°6 with a 93.2% sensitivity and a 60.0%
189 specificity.
190
10