RESEARCH

NON-CONTACT INFRARED THERMOMETRY VERSUS RECTAL THERMOMETRY IN YOUNG CHILDREN EVALUATED IN THE EMERGENCY DEPARTMENT FOR FEVER
OF
Authors: Ezio L. Fortuna, RN, Michele M. Carney, MD, Michelle Macy, MD, Rachel M. Stanley, MD, John G. Younger, MD, and Stuart A. Bradin, DO, Ann Arbor, MI

ACCURACY

Earn Up to 8.5 CE Hours. See page 187.

Objective: We evaluated the accuracy of a non-contact infrared thermometer compared with a rectal thermometer. Methods: Two hundred patients, ages 1 month to 4 years, were Results: A linear relationship between rectal and infrared temperature measurements was observed; however, the coefficient of determination (r 2) value between was only 0.48 (P < 0.01). Infrared thermometry tended to overestimate the temperature of afebrile children and underestimate the temperature of febrile patients (P < .01). Ambient temperature and child age did not affect the accuracy of the device. Conclusion: In this study, non-contact infrared thermometry did not sufficiently agree with rectal thermometer to indicate its routine use. Key words: Rectal thermometry; Non-contact infrared

included in the study. Each child underwent contemporaneous standard rectal thermometry and mid forehead non-contact infrared thermometry. Clinical features, including chief complaint, recently administered antipyretic agents, and ambient temperature at the time of measurement, were included.
Analysis: Linear models were used to compare agreement

between the 2 techniques, as well as to determine bias of infrared thermometry at different rectal temperatures. Multivariate linear models were used to evaluate the impact of clinical variables and ambient temperature.

ever is a common chief complaint in the pediatric emergency department. Studies suggest that the height of fever, in combination with other clinical features, is a reliable predictor of occult bacteremia.1-3 Oral and rectal temperatures are the most reliable predictors of core temperature.4,5 Children who are able to have an oral temperature obtained have the advantage of a simple and reliable measurement tool. Children younger than 4 years have difficulty keeping an oral thermometer under their tongues, making reliable temperature measurements difficult.
Ezio L. Fortuna is Staff Nurse, Childrens Emergency Services, and Level E Nurse, Department of Emergency Medicine, Ann Arbor, MI. Michele M. Carney is Clinical Instructor, Department of Emergency Medicine, University of Michigan, Ann Arbor, MI. Michelle Macy is Fellow, Pediatric Emergency Medicine and Pediatric Health Services Research, Department of Emergency Medicine, University of Michigan, Ann Arbor, MI. Rachel M. Stanley is Assistant Professor, Emergency Medicine and Pediatrics, Department of Emergency Medicine, University of Michigan, Ann Arbor, MI. John G. Younger is Associate Professor, Associate Chair for Research, Department of Emergency Medicine, University of Michigan, Ann Arbor, MI.

At present, rectal thermometry is the clinical standard when measuring temperature in young children4-6 because it correlates highly with core body temperature. While axillary and aural thermometry may be less invasive, measurements provided by these devices are not sufficiently accurate or reliable.4,6-14 In fact, Dodd and colleagues9 suggest that tympanic infrared thermometry fails to diagnose fever in 3 to 4 out of 10 febrile children. This inaccuracy is unacceptable in clinical settings wherein fever guidelines often are initiated at specific fever thresholds.4
Stuart A. Bradin is Assistant Professor of Pediatrics and Emergency Medicine, Department of Emergency Medicine, University of Michigan, Ann Arbor, MI. For correspondence, write: Ezio Fortuna, RN, CEN, EMT-P, Department of Emergency Medicine, 1500 E Medical Center Dr, Ann Arbor MI 48109-5305; E-mail: monnnmn@umich.edu. J Emerg Nurs 2010;36:101-4. Available online 3 September 2009. 0099-1767/$36.00 Copyright 2010 Emergency Nurses Association. Published by Elsevier Inc. All rights reserved. doi: 10.1016/j.jen.2009.07.017

March 2010

VOLUME 36 ISSUE 2

WWW.JENONLINE.ORG

101

RESEARCH/Fortuna et al

The situation is further complicated in some children by the presence of iatrogenic neutropenia, where rectal thermometry is relatively contraindicated as to avoid potential injury to the mucosal barrier in an immunocompromised host.15 In these patients there are no good alternatives. Infrared thermometry is a technology in which a detector samples blackbody radiation emitted from a warm surface and provides an estimate of that surfaces temperature. The technique is the basis of tympanic thermometry. Recently, economical hand-held devices have become available that estimate core temperature based on surface temperature determinations. Should these devices prove reliable, they could become an important part of initial and ongoing assessment of children in whom other temperature measurement strategies are insufficient. In the current work, we evaluated a commercially available non-contact infrared thermometer in children younger than 4 years. We specifically sought to determine the agreement between measurements taken by this device and standard rectal thermometry.
Methodology

TABLE

Characteristics of patients enrolled

Characteristic N (%)

Medication prior to arrival None Tylenol Motrin Unknown Five most common chief complaints Fever Cough Vomiting Laceration Rash

147 (53.8) 44 (16.1) 16 (5.9) 66 (24.2) 119 (43.6) 17 (6.2) 15 (5.5) 14 (5.1) 11 (4.0)

We studied a prospective convenience sample of children ages 1 month through 4 years presenting to a tertiary pediatric emergency department. Patients were included if their presenting acuity was not too high as to preclude participation (ie, triage levels 3-5), had an English-speaking parent or guardian, had no contraindication to rectal thermometry, and had no skin abnormalities on the forehead (where the measurements were taken). Informed consent was obtained in all cases, and the protocol was approved by the local institutional review board. Standard demographic information collected from each patients guardian included age, any antipyretics given prior to measurement, the temperature of the room at the time of measurement, and the chief complaints of the participants.
TEMPERATURE MEASUREMENTS

converging the thermometers 2 range-finding light emitting diodes into one luminous point. The operator held the device still until signaled that a reliable reading had been taken. The operator then recorded the temperature visualized on the display. Two registered nurses, 2 pediatric emergency physicians, and 2 ED paramedics were trained by the one of the authors (EF) following the manufacturers guidelines regarding obtaining rectal and infrared temperatures. These individuals conducted all measurements in the study.
STATISTICAL METHODS

A Welch Allen SureTemp thermometer, model 678, was used to obtain rectal temperatures. It was calibrated using a manufacturer-supplied calibration key. The thermometer was then introduced into the rectum to a depth of 1.5 cm from the anal margin. The temperature was read 15 seconds after insertion. Immediately after the measurement of the rectal temperature, the same operator then recorded the skin temperature using the Thermofocus non-contact infrared thermometer model 1500 (Tecnimed srl Vedano Olona [VA] Italy) on the central part of the forehead. The thermometer was first calibrated to the room temperature and then held perpendicular to the center of the forehead,

Agreement between infrared and rectal thermometry was made in the standard fashion as described by Altman and Bland.16 Rectal temperature is considered the gold standard. Analysis began by determining the correlation between the 2 measurements. Next, bias (greater disagreement between the 2 techniques at some temperature ranges more than others) was considered by determining the correlation of rectal temperature to the difference between infrared and rectal temperature. First-order linear models were used to consider the impact of patient age and ambient temperature on measurement agreement. All analyses were performed using the statistical package R 2.6.0 (www.cran.r-project.org).
Results

Two hundred patients were enrolled. The mean age of children studied was 1.4 years (interquartile range, 0.7, 2.0 years). The average rectal and infrared temperatures of all participants were 99.6F (98.7, 100.5) and 99.5F (98.6, 100.3), respectively. The ambient temperature during measurements

102

JOURNAL OF EMERGENCY NURSING

VOLUME 36 ISSUE 2

March 2010

Fortuna et al/RESEARCH

the regression line was far from unity (0.697 + 0.05, r2 = 0.48, P < .01), and the 95% prediction band for an infrared measurement given a rectal temperature was unacceptably broad, on the order of 4F. An ideal agreement between the 2 techniques should have a slope of one and an intercept of zero. In addition to only moderate agreement between methods, evidence also existed of infrared thermometry overestimating rectal temperature in patients with lower temperatures and underestimating rectal temperature in patients with fever (r2 = 0.149, P < .01; see Figure, part B). More sophisticated linear models incorporating patient age, ambient temperature, and the additive combination of these 2 factors did not find statistically significant contributions of these data on the level of agreement (data not shown).
Discussion

FIGURE Agreement between rectal and non-contact infrared thermometry. A, The correlation between rectal temperature and infrared temperature is shown. Correlation was modest. The diagonal dotted curves are the upper and lower 95% confidence interval for future predictions (eg, for a patient with a rectal temperature of 100.0F, 19 times out of 20 the non-contact measurement will fall between 97.5 and 101.5F. B, The bias of infrared versus rectal thermometry is shown. In an ideal instrument, the difference between infrared and rectal thermometry would be 0F at any temperature measured. For the device studied, there was significant device overestimation of temperature at low rectal temperatures and significant underestimation at elevated rectal temperatures (P < .01).

was 73.6F (72.6, 75). Antipyretic agents provided by caregivers prior to measurement and the most common presenting chief complaints are shown in the Table. A diagnostic plot of measurement agreement is shown in part A of the Figure. As expected, there was a monotonic linear relationship between rectal temperature measurements and infrared thermometry that was highly statistically significant (P < .01). However, the slope of

In the current study, we compared a non-contact infrared thermometer with a standard rectal thermometer in the evaluation of young children in a pediatric emergency department. Although measurements of surface temperature correlated modestly with rectal temperatures taken contemporaneously, the agreement between the 2 methods was not sufficiently strong to recommend the use of the tested infrared device in clinical practice. Fever, defined by consensus as a rectal temperature greater than 38.0C (100.4F), is a very common finding in children and accounts for close to 20% of all pediatric ED visits.17 Children younger than 2 years average 4 to 6 acute fever episodes annually.6 Presence and degree of fever have been correlated with an increased risk for bacteremia.17 Fever with no identifiable source remains both a diagnostic and management challenge; interventions are dictated by detection of an elevated temperature, especially when stratified by age.2,6 In the child younger than 2 months of age, and certainly in the neonate (aged 30 days or less), a history of and/or documented temperature of greater than 38C will trigger an evaluation for sepsis, including blood and urine cultures, lumbar puncture, antibiotics, and potential hospital admission. Prior to the Haemophilus influenzae and conjugate pneumococcal vaccines, fever greater than 39C in the 3- to 36-month-old population with no identifiable source often requires invasive investigation because of the concern for occult bacteremia and its sequelae.6,17 Clinical guidelines, initially published in 1993 by Baraff and colleagues2 to help identify children thought to be at most risk for serious bacterial infection (pneumonia, urinary tract infection, meningitis, and osteomyelitis), have evolved but still guide clinical practice.3,4 Our study attempted to demonstrate the accuracy of a non-contact infrared thermometer when compared with the accepted standard of rectal thermometry. We found poor

March 2010

VOLUME 36 ISSUE 2

WWW.JENONLINE.ORG

103

RESEARCH/Fortuna et al

agreement between the 2 methods. Concordance between rectal and infrared temperatures was low, illustrated as well by DeCurtis and colleagues.5 Similar to findings by Devrim and colleagues8 and Schuh and colleagues,18 infrared thermometry showed exaggerated disagreement at temperature extremes. Infrared thermometry overestimated low temperatures and underestimated temperatures in children with documented (rectal temperature) fever. Our results are striking in that we observed for any given rectal temperature; the infrared measurement can be expected to lay more than 2C above or below a simultaneous rectal temperature once every 20 measurements. Ambient temperature and child age did not affect the accuracy of the device. When clinical decisions are based on the presence of true fever, despite its speed, lack of invasiveness, and ease of use, infrared thermometry is too unreliable and inaccurate for use in the ED setting.
Limitations

REFERENCES
1. Lee GM, Harper MB. Risk of bacteremia for febrile young children in the post-Haemophilus influenzae type b era. Arch Pediatr Adolesc Med. 1998;152:624. Baraff LJ, Bass JW, Fleisher GR, et al. Practice guidelines for the management of infants and children 0-36 months of age with fever without source. Pediatrics. 1993;92:1-12. Baraff LJ. Management of fever without source in infants and children. Ann Emerg Med. 2000;36:602-14. Ishimine P. Fever without source in children 0-36 months of age. Pediatr Clin North Am. 2006;53:167-94. De Curtis M, Calzolari F, Marciano A, et al. Comparison between rectal and infrared skin temperature in the newborn. Arch Dis Child (Fetal Neonatal Ed). 2008;93:F55F57. McCarthy PL. Fever. Pediatr Rev. 1998;19:401-8. Yaron M, Lowenstein SR, Koziol-McLain J. Measuring the accuracy of the infrared tympanic thermometer: correlation does not signify agreement. J Emerg Med. 1995;13:617-21. Devrim I, Kara A, Ceyhan M, et al. Measurement accuracy of fever by tympanic and axillary thermometry. Pediatr Emerg Care. 2007;23:16-9. Dodd SR, Lancaster GA, Craig JV, et al. In a systematic review, infrared ear thermometry for fever diagnosis in children finds poor sensitivity. J Clin Epidemiol. 2006;59:354-7.

2.

3. 4. 5.

6. 7.

8. 9.

Convenience recruiting strategies must always be carefully considered when interpreting prospectively gathered data. It is possible that immeasurable selection bias against the test device was present, although we do not find evidence in our analysis to suggest such bias. Inter-rater accuracy was not evaluated in this study, and although detailed device training was provided to each of the caregivers collecting data, it is possible that relative inexperience may have been a disadvantage for the test device. However, we do not believe that use of the non-contact device in our study differs significantly from how we would deploy the device in daily non-experimental use.
Implications for Emergency Nurses

10. Asher C, Northington LK. Position statement for measurement of temperature/fever in children. J Pediatr Nurs.. 2008;23:234-6. 11. Craig JV, Lancaster GA, Taylor S, et al. Infrared ear thermometry compared with rectal thermometry in children: a systematic review. Lancet. 2002;360:603-9. 12. Haldeman LA, Allen TL, Chan KJ, et al. Feasibility and utility of infrared thermography as a thermometer: comparison with oral and tympanic thermometry. Ann Emerg Med. 2004;44:s53-s170. 13. Romanovsky AA, Quint PA, Benskova Y, et al. A difference of 5 degree C between ear and rectal temperatures in a febrile patient. Am J Emerg Med. 1997;15:383-5. 14. Muma BK, Treloar DJ, Wurmlinger K, et al. Comparison of rectal, axillary, and tympanic membrane temperatures in infants and young children. Ann Emerg Med. 1991;20:41-4. 15. Wolff LJ, Ablin AR, Altman AJ, Johnson FL. The management of fever. In: Ablin AR, ed. Supportive Care of Children With Cancer: Current Therapy and Guidelines From the Childrens Cancer Group. Baltimore, Md: Johns Hopkins University Press; 1997:23. 16. Altman DG, Bland JM. Measurement in medicine: the analysis of method comparison studies. Statistician. 1983;32:307-17. 17. Brennan DF, Falk JL, Rothrock SG, et al. Reliability of infrared tympanic thermometry in the detection of rectal fever in children. Ann Emerg Med. 1995;25:21-30. 18. Schuh S, Komar L, Stephens D, et al. Comparison of the temporal artery and rectal thermometry in children in the emergency department. Pediatr Emerg Care. 2004;20:736-41.

The non-contact infrared device studied in this report did not perform sufficiently well to warrant its adoption as a standard means of estimating core temperature in young children. A reliable alternative to rectal thermometry remains an important engineering problem in the field.
Conclusions

When used in a prospectively enrolled cohort of young children in a pediatric emergency department by caregivers specifically trained in its use, a non-contact infrared thermometerbased measurement failed to a accurately estimate contemporaneously collected rectal temperatures.

104

JOURNAL OF EMERGENCY NURSING

VOLUME 36 ISSUE 2

March 2010