Current Problems in
iatrics" Volume 27
Number 8
September 1997
Repeat Pediatric Visits to a General Emergency Department Deena R. Zimmerman,MD, MPH, KevanA. McCarten-Gibbs,MD, DeniseH. DeNoble, Caryn Borger,JacquelineFleming,MD, Margaret Hsieh,JessicaC. Langer,and Mary B. Breckenriclge,PhD Departments of Pediatrics and Family Medicine University of Medicine and Dentistry of New Jersey--Robert Wood Johnson Medical School New Brunswick, New Jersey
Subjective Assessment of Fever by Parents: Comparison with Measurement by Noncontact Tympanic Thermometer and Calibrated Rectal Glass Mercury Thermometer EdmondA. Hooker,MD, StevenW. Smith,MD, Tim Miles, MD, and LonnieKing, MD Departments of Emergency Medicine and Pediatrics University of Louisville School of Medicine Louisville, Kentucky
Vasculitis in Children: A Diagnostic Challenge Rub6n J. Cuttica, MD Rheumatology Section Hospital de Pediatrfa "Dr. Pedro de Elizalde" Buenos Aires, Argentina
~v~ Mosby ATimes Mirror Company
Subjective Assessment of Fever by Parents: Comparison With Measurement by Noncontact Tympanic Thermometer and Calibrated Rectal Glass Mercury Thermometer Edmond A. Hooker, MD,* Steven W. Smith, MD,* Tim Miles, MD,* and LonnieKing, MD~ Study objective: To assess the ability of parents to subjectively evaluate their children for fever and to compare their assessments with temperature measurements made with the use of a noncontact tympanic (NCT) or rectal glass mercury thermometer. A secondary goal was to assess how well a recently developed definition of fever for NCT thermometers, when used in the ear-equivalent mode (temperature of 37.7 ~ C or more), performed in a clinical situation. Methods: This 6-month prospective observational study employed a convenience sample of 180 children, aged birth to 4 years, who presented to the emergency department of a tertiary care children's hospital. Parents were asked to subjectively assess whether their child had a fever. The child's temperature was then measured with an NCT thermometer (three times in the rectal-equivalent mode and three times in the actual-ear mode). Both the subjective assessment and the NCT temperatures were compared with the rectal temperature measured by a rectal glass mercury thermometer. Results: The mean age of participants was 14.6 + 11.8 months (range, 2 days to 48 months); 56% were boys. The sensitivity of parental detection of fever by subjective means was 81.8% and the specificity 76.5%. The parent and the rectal glass thermometer agreed 79% of the time (95% confidence interval [CI], 73% to 85%). The sensitivity of the first temperature reading obtained with the NCT thermometer in rectal-equivalent mode was 74.7%, and the specificity was 96.3%. The NCT thermometer and the rectal glass thermometer agreed 84% of the time (95% CI, 78% to 89%). Use of the proposed definition of fever for NCT thermometers, when used in the ear-equivalent mode, caused sensitivity of a single measurement for fever to drop to 53.5%. Conclusion: Parental subjective assessment of fever agreed with the presence of fever as measured by rectal glass thermometer in 79% of cases. Specificity was improved with the use of the NCT thermometer. The recently proposed definition for fever for NCT thermometers, when they are used in the ear-equivalent mode, does not appear to be validated by the current data. (Hooker EA, Smith SW, Miles T, King L: Subjective assessment of fever by parents: Comparison with measurement by noncontact tympanic thermometer and calibrated rectal glass mercury thermometer. Ann Emerg Med September 1996;28:313-317.) Reprinted from Annals of Emergency Medicine 1996;28:313-7, copyright 9 1996 by the American College of Emergency Physicians. From the Departments of Emergency Medicine* and Pediatrics,~f University of Louisville School of Medicine, Louisville, Kentucky. Received for publication December 27, 1995. Revision received March 5, 1996. Accepted for publication April 3, 1996. The noncontact tympanic thermometers and calibrating instruments used in this study were provided by Thermoscan Incorporated. Address for reprints: Edmond A. Hooker, MD, 11609 Stonewall Jackson Drive, Spotsylvania, VA 22553. 0196-0644/96 $5.00 + .10 5 3 / 1 / 8 4 3 9 5
304
he optimal method for the assessment of fever in outpatients has been widely debated in recent years with the introduction of external auditory canal thermometers, also referred to as noncontact tympanic (NCT) thermometers. 1 In clinical practice, it is not uncommon to find that parents have not objectively assessed their child's temperature before presenting the child for care. However, many parents report that the child has a fever based on their subjective assessment. There has been very little research into the
T
Curr Probl Pediatr, September 1997
sensitivity of subjective parental assessment of the presence of fever in children. Singhi showed that mothers detected 88.9% of fevers that were subsequently detected by oral or axillary thermometry.2 Banco found assessment of fever by mothers to have a sensitivity of 73.9%; however, he used a mixture of rectal and oral thermometry as his gold standard. 3NCT thermometers are easy to use and are well accepted by patients and their parents .4 Although many researchers have shown these devices to be poorly sensitive for fever, the devices are available to consumers. 5,6 The purpose of this study was to evaluate how well parents detected fever compared with the current gold standard, rectal temperature measured with a calibrated glass mercury thermometer. Because many consumers are purchasing the NCT thermometers, we evaluated whether the NCT thermometer was as good or better than subjective assessment. A last goal was to evaluate how well the recently developed definition of fever for NCT thermometers performed in a clinical situation.7
Materials and Methods A convenience sample of children younger than 5 years who presented for care to a tertiary care children's hospital were eligible for enrollment in the study. Patients who arrived at the emergency department by ambulance and those judged to be critically ill were excluded from the study. One of four investigators was at the triage desk and identified eligible children before vital signs were measured. The investigator briefly explained the study and asked the parent whether he or she would consent to the child's enrollment in the study. If the parent agreed, written informed consent was obtained. The parent was asked a series of eight questions, including whether he or she believed the child had a fever at the time of the study. After the parent performed a subjective assessment of the child, the child's temperature was measured with a Thermoscan PRO-LT NCT thermometer (Thermoscan Incorporated). Three ear temperature measurements, separated by 30 seconds, were obtained with the device in the rectal-equivalent mode, and a second series of three measurements was obtained with the device in the actual-ear mode. The child's temperature was then measured with a Tern-Con rectal glass mercury clinical thermometer. Each rectal glass mercury thermometer was checked in a stirred water bath at three different temperatures against a certified mercury laboratory thermometer (Ever Ready Thermometer) with accuracy traceable to the National Bureau of Standards. Those
Curr Probl Pediatr, September 1997
thermometers displaying accuracy within 0.2 ~ F (0.1 ~ C) were retained for study use. All investigators were given in-service instruction before commencement of the study. The in-service training included review of the instruction manual for the NCT thermometer, hands-on demonstration of the NCT thermometer, description of the correct method of rectal temperature measurement, and review of the survey instrument. Calibration of the NCT thermometer was performed at least weekly using the IR-3000 Infrared Temperature Reference (Thermoscan Inc), which is a cavity-type black body. Any NCT thermometer that was found to vary more than _+0.2 ~ F (0.1 ~ C) from the reference thermometer was returned to the manufacturer for replacement. Fever was defined as a temperature equal to or greater than 100.4 ~ F (38,0 ~ C) by rectal glass mercury thermometer or by NCT thermometer in the rectal-equivalent mode, or a temperature equal to or greater than 99.9 ~ F (37.7 ~ C) by NCT thermometer in the actual-ear mode. 7,8The data were analyzed to test the reliability of subjective parental assessment and of NCT temperature measurement in the detection of fever. Sensitivity, specificity, and limits of agreement with corresponding 95% confidence intervals (CIs) were calculated for each method. The significance of level of education of the parents and that of family size was assessed with the Mann Whitney U test, with significance set at a P value of .05. The study was approved by the Human Studies Committee of the University of Louisville School of Medicine, and informed consent was obtained before study enrollment.
Results Over a 6-month study period, 180 patients were enrolled in the study. Their mean age was 14.6 _+ 11.8 months (range, 2 days to 48 months); 56% were boys. The mean temperatures obtained were as follows: rectal, 101.0 ~ _+ 1.8 ~ F (38.3 ~ _ 1.0 ~ C); first NCT measurement in rectal-equivalent mode, 100.5 ~ _ 2.0 ~ F (38.1 ~ _+ 1.1 ~ C); maximum of three NCT measurements in rectal-equivalent mode, 100.8 ~ _ 2.0 ~ F (38.2 ~ _+ 1.1 ~ C); first NCT measurement in actual-ear mode, 99.0 ~ _+ 1.8 ~ F (37.2 ~ + 1.0 ~ C); maximum of three NCT measurements in actual-ear mode, 99.2 ~ _+ 1.8 ~ F (37.3 ~ _+ 1.0 ~ C). Of the 180 patients, 99 (55%) were febrile, as assessed by rectal glass mercury thermometer. The performance of parents and the NCT thermometer in the detection of fever is shown in Table 1. The
305
Sensitivity
Sensitivity 1.0
.....__~r___a__u__a~a--a-ca-a~a-a-aam
Wf
.80
1.0
.80
un
.60
.60
.40
.40
.20
.20 0
,
0
.20
.40
.60
.80
0
I
0
1,0
I
.20
I
.40
.60
I
I
.80
1.0
1-SPECIFICITY
1-SPECIFICITY FIG. 1. ROC curve for the NCT thermometer in the rectal-equivalent mode.
FIG. 2. ROC curve for the NCT thermometer in the actual-ear mode.
TABLE 1. Performance of parents and NCT thermometers in the detection of fever as measured with a rectal glass mercury thermometer (temperature _>100.4 ~ F [38.0 ~ C])
Temperature Reading Subjective assessment by parent NCT thermometer* First rectal-equivalent measurement Maximum rectal-equivalent measurement First actual-ear measurement Maximum actual-ear measu rement
(%)
(%)
Positive Predictive Value (%)
81,8
76.5
81.0
77.5
74.7
96.3
96.1
75.7
78.8
96.3
96.3
78.8
Sensitivity
Specificity
Negative Predictive Value (%)
53.5
100
100
63.8
58.6
100
100
66.4
*Fever was defined as 100.4 ~ F (38.0 ~ C) for rectal-equivalent mode, 99,9 o F (37.7 ~ C) for actual-ear mode.
sensitivity for parents in detecting fever by subjective means was 81.8%, and the specificity was 76.5%. The parent and the rectal glass mercury thermometer agreed 79% of the time (95% CI, 73% to 85%). The sensitivity of the first rectal-equivalent measurement made with the NCT thermometer to detect a fever was 74.7%, and the specificity was 96.3%. The NCT thermometer and the rectal thermometer agreed as to the presence of fever 84% of the time (95% CI, 78% to 89%). If the maximum of three consecutive NCT temperatures was used, the sensitivity was 78.8% and the specificity 96.3%. This measurement and the rectal thermometer agreed 87% of the time. Receiver operating characteristic (ROC) curves were constructed from the data for the NCT thermometer in the rectal-equivalent mode (Figure 1). Each point represents the diagnostic utility of the NCT thermometer
306
as the definition of fever is varied (in increments of 0.1 ~ F). It appears that a cutoff of 100.1 ~ F (37.8 ~ C) would provide a better definition for fever when the thermometer is used in the rectal-equivalent mode. We also evaluated the proposed definition of fever for the NCT thermometer when used in the ear-equivalent mode (temperature 99.9 ~ F [37.7 ~ C]). 7 This definition causes the sensitivity for fever to drop to 53.5% for a single measurement or 58.6% for the maximum of three measurements. An ROC curve was constructed for the ear-equivalent mode (Figure 2). It appears that a temperature of 98.3 ~ F (36.8 ~ C) or more would be a better definition for fever when the NCT thermometer is used in the ear-equivalent mode. The questionnaire results indicated that 91.1% of parents believed that they could tell whether their child had a fever by touching the child. Fifty-six percent believed that
Curr Probl Pediatr, September 1997
their child had a fever at the time of the evaluation. Eighty-eight percent of parents did own a thermometer; however, only 56% had used it before presentation of the child to the ED (Table 2). There was no significant effect of number of children in the family (z = -1.074, P = .2829), level of education (z = -.042, P = .9666), or site of assessment (z = .991, P = .3219) on the ability of parents to subjectively detect fever. During the study period, two of the NCT instruments were returned to the factory because they were no longer in calibration. Neither instrument gave any indication of being out of calibration except when it was tested against the black body.
Discussion In clinical emergency medicine, parents frequently report that a child had a fever at home. Although some parents have objectively measured the temperature, it is not uncommon to have only a subjective assessment by the parent. Because the workup may be modified on the basis of history of fever, physicians must know whether they should believe the parent's subjective assessment. Two previous authors have looked at the subjective assessment of fever by parents. Singhi showed that mothers detected 88.9% of fevers that were subsequently detected by oral or axillary thermometry.2 Banco found assessment by mothers to have a sensitivity of 73.9%; however, he used a mixture of rectal and oral measurements as his standard3 Our study found parental subjective assessment to have a sensitivity of 81.8% and a specificity of 76.5%. The parental assessment and the rectal glass mercury thermometer agreed 79% of the time. These results indicate that although parents are not perfect at detecting fevers, a parental assessment that the child had a fever at home should not be ignored. NCT thermometers are sold for home use. The instrument we chose is similar to the model sold for home use by Thermoscan. The sensitivity of the NCT thermometer if only the first rectal-equivalent temperature measurement was used was 74.7%, and the specificity was 96.3%. The NCT thermometer and the rectal thermometer agreed 84% of the time (95% CI, 78% to 89%). The sensitivity increased if the maximum of three consecutive measurements was used. These results indicate that the NCT readings are almost as sensitive and more specific than subjective assessment, if the built-in offset were to be increased, the sensitivity of the NCT thermometer could be increased without sacrificin~ much specificity. We also evaluated the proposed definition of fever for
Curr Probl Pediatr, September 1997
TABLE 2. Questionnaire results (N=180) % Question Do you believe your child has a fever now? Do you own a thermometer? Did you use a thermometer to assess your child's fever? Can you tell whether your child has a fever by touch or feel? Did you give your child antipyretics? Where do you touch to evaluate for fever? Forehead Forehead plus another body part Trunk Other How many children do you have? One Two Three or more What is your level of education? Did not graduate from high school High school graduate College graduate Graduate school
Responding Poslt!ve!y 55.6 87.8 55.6 91.1 46.7 67,4 19.7 7.3 5.6 42.1 32.6 25.3 34.5 41.2 22,6 1.7
the NCT thermometer when used in the actual-ear mode (temperature >99.9 ~ F [37.7 ~ C]). 7 Some authors have proposed that the NCT instruments would be better if used in the actual-ear mode. Using the definition of fever proposed by Chamberlain caused the sensitivity for fever to drop to 53.5% for a single measurement or 58.6% for the maximum of three measurements. The results of the current study do not support the definition of fever proposed for the NCT thermometers when used in the actual-ear mode. Our study has a few limitations. The NCT readings were not performed by the parent; these instruments may have performed worse or better when used by the parent. The study was also limited by having the parent assess the fever only once; repeat assessments over time may have resulted in higher accuracy. Some authors have shown that, in patients being rapidly cooled for cardiac bypass surgery or for hyperthermia, rectal temperature readings may lag behind changes in core temperature. T M Although it is difficult to extrapolate their results to outpatients seen in an ED, it is possible that a child may be afebrile in the ear and yet febrile in the rectum. The final limitation is the definition of fever; for some children, a lower temperature (e.g., 100.0 ~ F [37.7 ~ C]) may represent a fever. Therefore, the parental assessment may have a better specificity than reported. The use of a convenience sample design may also have introduced bias. Patients were entered when one of the principal
307
investigators could be present. Because consent was required and not all children who presented for care were entered, these results may not apply to all parents. Also, most patients were entered in the evening hours, and their parents may have performed differently than parents bringing their children for care in the middle of the night. Parental subjective assessments of the presence of fever agreed with measurement by rectal glass thermometer in 79% of cases. Specificity was improved with the use of an NCT thermometer. The authors thank L. Jane Goldsmith, PhD, senior statistical consultant, for statistical support and review of the manuscript.
References 1. Chamberlain JM. Tympanic thermometers. South Med J 1994;87:1059-60. 2. Singhi S, Sood V. Reliability of subjective assessment of fever by mothers. Indian Pediatrics 1990;27:811-5. 3. Banco L, Veltri D. Ability of mothers to subjectively assess the
308
presence of fever in their children. Am J Dis Child 1984;138: 976-8. 4. Alexander D, Kelly B. Responses of children, parents, and nurses to tympanic thermometry in the pediatric office, Clin Pediatr 1991;30(suppl):53-6. 5. Hooker EA. The use of tympanic thermometers to screen for fever in pediatric emergency department patients. South Med J 1993;86:855-8. 6. Brennan DE 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. 7. Chamberlain JM, Terndrup TE, Alexander DT, et al. Determination of normal ear temperature with an infrared emission detection thermometer. Ann Emerg Med 1995;25:15-20. 8. Anagnostakis D, Matsaniotis N, Grafakos S, et al. Rectal-axillary temperature difference in febrile infants and children. Clin Pediatr 1993;32:268-72. 9. Ramsay JG, Ralley FE, Whalley DG, et al. Site of temperature monitoring and prediction of afterdrop after open heart surgery. Can Anaesth Soc J 1985;32:607-12. 10. Edwards RJ, Belyavin AJ, Harrison MH. Core temperature measurement in man. Aviat Space Environ Med 1978;49:128494. 11. Ash CJ, Cook JR, McMurry TA, et al. The use of rectal temperature to monitor heat stroke. Missouri Med 1992;89:283-8.
Curr Probl Pediatr, September 1997