Temperature Monitoring

Transcript

RESEARCH © 2004 NNA All rights reserved Temperature monitoring in newborns: A comparison of thermometry and measurement sites Linda S Smith MS, DSN, RN, Assistant Professor of Nursing, Oregon Health and Science University, Klamath Falls, Oregon Newborns have poor thermal stability and therefore infants are extremely vulnerable to body temperature variations. The purpose of choosing a temperature site in neonates is to minimise the effect of extraneous variables such as environment, on the readings. Moreover, the most accurate, least invasive and obtrusive method should be chosen. Traditional, low-tech mercury-glass thermometers are no longer acceptable. This study sought to examine the effect of age, weight, sex, race and ambient temperature on body temperature in newborns and also to compare the level of agreement and relationship between temperature readings from DataTherm, SolarTherm devices and mercury-glass and hard-wire Mallinckrodt devices. Keywords: temperature; thermometry; clinical monitoring; newborn; DataTherm, SolarTherm; mercury Background As a measure of thermal balance and thermoregulation, accurate monitoring of body temperature in neonates is essential. Neonates also have decreased subcutaneous fat, a thin epidermis, and when compared with adults, a greater body surface related to body mass, and blood vessels that are closer to skin surfaces. Newborns have poor thermal stability mostly due to excessive heat loss – about four times the heat loss compared to adults (Ladewig et al, 1998). Thus, infants are extremely vulnerable to body temperature variations. In addition newborns are especially susceptible to serious infection (McKenzie, 1998). Variables that influence temperature measurement The purpose of choosing a temperature site in neonates is to minimise effects of extraneous variables on readings. Moreover, the most accurate, least invasive and obtrusive method should be chosen, given the neonate’s condition. For this study, rectal and skin sites (axilla, groin) were chosen. Rectal Healthcare professionals continue to consider the rectal temperature site as the gold standard for measurement of core body temperature, especially in children (Bliss-Holtz, 1989; Carlson, 1996; Cattaneo, et al., 2000; Eoff, et al, 1974; Greenes and Fleisher, 2001), and rectal temperature measurement has been most often used and requested by physicians (Clarke, 1992). Rectal temperatures, as identified via mercury-glass thermometers in neonates, were studied by Browne et al (2000) and found to have the highest mean and lowest variability. Robinson and colleagues (1998) reported that urinary bladder site temperatures were no more accurate than rectal temperatures in children. Recent concerns regarding the rectal temperature site have surfaced. Rectal lag – the lag time needed for rectal 6 Temperature monitoring temperatures to ‘catch up’ to core temperature readings – was identified by a number of scientists (Ash, et al 1992; Buck and Zaritsky, 1989; Erickson and Woo, 1994; Robinson, et al 1998). The rectal site carries additional concerns of cross contamination, discomfort, possible presence of stool, and limited social acceptability. Another concern relates to rectal injury risk during assessment. To diminish this risk, recommendations were made for the use of flexible thin lubricated catheter probes, inserted to a depth of 3 cm or less (Keeling, 1992; Kunnel et al, 1988; Lodha et al, 2000; Zehner and Terndrup, 1991) during rectal temperature assessment. Axilla For infants, Lodha et al (2000) identified the axillary site as less likely to be affected by technique variances. Payne et al (1994) Key points Smith, L.S. (2004) Temperature monitoring in newborns: A comparison of thermometry and measurement sites. Journal of Neonatal Nursing 10(5): xx-xx. 1. -Neonates have poor thermal stability, therefore accurate monitoring of body temperature is important. 2. Mean groin temperatures were lower than axillary and rectal; skin temperatures were lower than rectal temperatures. 3. When compared with mean body temperature, newborn temperatures related to weight and sex: heavier newborns had higher temperatures; female and non-white newborns tended to have lower mean temperatures. 4. Axilla sites demonstrated good comparisons with rectal. 5. Axilla skin sites performed better than groin skin sites and DataTherm and SolarTherm devices performed better than mercury-glass. 6. Parental education regarding temperature warning signs, safe use of temperature devices, and appropriate temperature sites for newborns, is essential. JOURNAL OF neonatal NURSING VOLUME 10 ISSUE 5 2004 RESEARCH in their child study of two devices at the axillary site, used a mercury-glass reference device and 5 minute dwell times. When identifying the two axilla sides in simultaneous measures (each device on both sides), these investigators found a smaller sideto-side difference with the mercury device, when compared with the disposable chemical thermometer (Tempa DOT). With 57 neonates and comparing tympanic, axillary, and rectal sites, Browne et al (2000) identified the axillary site, with temperatures measured via an infrared axillary thermometer (correlating well with rectal mercury) as the best due to saved nursing time and non-invasive nature of the site. They concluded that the electronic axillary site/device was a viable alternative to the mercury axilla and rectal methods. These data also validate the work of Torrance (1968) who studied simultaneously assessed rectal and axillary temperature readings in 120 preterm infants. Torrance found that when maximum dwell times were imposed (4.5 minutes) and highest readings identified, the mean difference between sites was 0.05°C (0.09°F). These findings differ from the work of Haddock et al (1986) who suggested the ten minute axilla dwell time after studying 31 infants. Browne and colleagues (2000) used mercury-glass at the rectal site (3 minute dwell time) and compared readings with ear-based and axillary sites. Axillary measures were made with mercury-glass and infrared (Lightouch Neonate) thermometers. These researchers studied neonates and found the axillary sites compared well with rectal. Importantly, concern over the asymmetrical structure of the human thoracic cavity (Crafts, 1985) has prompted questions regarding normal side to side differences in body temperature assessed at skin sites. Groin The groin (inguinal, femoral) temperature site is another common location for infants and neonates. Kunnel and colleagues (1988) took simultaneous measures of rectal, groin, axillary, and skin-to mattress temperatures in 99 stable neonates. They found optimal dwell times with the mercuryglass device to be 5 minutes for rectal, 11 minutes for axilla, 11 minutes for groin, and 13 minutes for mattress. They wrote that both skin sites were good choices in the newborn nursery based on time and safety considerations. These data differ from the work of Bliss-Holz (1989) who found maximum temperatures reached for rectal and groin after 5 minutes and axillary after 5.5 minutes. Using mercury-glass thermometers, Bliss-Holz compared rectal, axillary, and groin temperatures in 120 fullterm infants. She reported using the inguinal site because it was safe, well supplied with blood vessels, easily accessed, and lacked brown adipose tissue usually found in newborns’ axillary pockets. In her discussion, Bliss-Holtz wrote that though the inguinal/rectal pairs had the greatest mean differences (0.8°F) they also had the strongest correlations. Eleven minute dwell times also contradicted the work of Stephen and Sexton (1987) who concluded that for neonates, axillary dwell times of 3 minutes were adequate. Mercury issues Mercury-glass thermometers have been used as reference in a number of paediatric and newborn research studies comparing JOURNAL OF neonatal NURSING VOLUME 10 ISSUE 5 2004 temperature sites and devices (Bliss-Holtz, 1989). The mercuryglass temperature device has long been considered the ‘gold standard’ for human temperature assessment (Clarke, 1992; Hooker, et al, 1996; Sganga et al, 2000). Clarke (1992) surveyed 145 UK general practitioners and found that mercuryglass was the most commonly used thermometer type. Unfortunately, only one physician reported using thermometer sheaths and six doctors denied cleaning thermometers between patients. Concerns over glass breakage and mercury toxicity have influenced government and resulted in organisational bans on the use, sale, and distribution of these devices, making way for mercury-free glass and electronic digital-display thermometers. Parent education Helping parents understand body temperature and infant fever is an essential nursing activity. Parents may be unaware of fever definition, which safe body sites and devices to use with infants, and what to do when suspicion of illness exists. Parents/child caretakers need to know how to accurately and safely take, read, and report an infant’s temperature (Haddock et al, 1986; Whaley and Wong, 1989). Healthcare researchers (Broome et al, 2003) studied efficient, cost effective fever management educational intervention strategies that successfully increased parents’ knowledge about fever management in children. Study group participants were given video and written information plus a fever management pack containing a fever measuring device (thermometer) and a pen or pencil. In addition, some study participants were given one-on-one verbal information and reinforcement from a healthcare professional. Following their multi-site investigation, they wrote that every parent/grandparent could benefit from fever management materials for children. Though informational brochures and video effectively increased short-term fever management knowledge, personalised instruction helped sustain that knowledge over time (Broome, et al, 2003). Purpose The purpose of this study was to determine how temperature readings taken in the axilla, groin, and rectum compare to each other in newborns. An additional purpose was to compare readings from four different types of thermometers. Methods The need for safe, accessible, trustworthy, convenient, and costeffective thermometer devices continues to dictate site and device selection in homes and healthcare facilities. It becomes critically important, however, to demonstrate device trustworthiness through clinical application and study. Design A descriptive study design was used to determine within-subject mean differences between and among sites and instruments. Descriptive statistics identified the sample, and described temperature measures and difference scores. Subjects were their own controls. Alpha was set at 0.05. The Statistical Package for the Social Sciences (SPSS, 2002) 11.5 was used for all statistical analyses. Temperature monitoring 7 RESEARCH Setting The Family Birthing Unit of a rural, pacific-northwest medical centre was the study setting. This medical centre is a 176-bed accredited healthcare facility with 12 birthing units where labour, delivery, and recovery take place in one room. Participants settings (range between 36°C and 40°C). Devices that did not meet these requirements were eliminated. Mallinckrodt probes were also water bath tested and found to record temperatures within ± 0.1°C. Instruments This was a convenience sample of neonates (n=44). Newborn infants and mothers were medical centre in-patients. In addition to Oregon Health & Science University Institutional Review Board (IRB) approval, IRB approval was obtained through the active medical center review board. Parent(s)/guardian(s) of all newborn participants signed informed consent documents prior to data collection. PI or research assistant (RA) provided consenting parents with 1:1 instructions and demonstrations. These instructions were given to parents with the help of a flexible demonstration doll, new thermometer and thermometer kit, printed literature, and 1:1 teaching/learning session. Parents took the thermometer kit and literature home for future use and reference. Instruction included: 1. How to use and read thermometer 2. How to safely and correctly place thermometer (demonstration doll) 3. “Normal” temperature ranges at skin sites in newborns DataTherm (study device) Procedures In their study of neonates, Leick-Rude and Bloom (1998) used the Mon-a-Therm Model 1000 by Mallinckrodt to assess skin temperatures. For this study, the Mallinckrodt/Mon-a-Therm model 4070 Thermistor Temperature Monitor (catalogue # 501-0070) was used. This instrument provides temperature readings at two sites when cables are inserted and temperature probes are connected to cables. Probes were flexible catheters with thermistor tips. Sterile, single use rectal general-purpose (Thermistor 400 series, Mon-a-therm) thermistor probes (temperature sensors at tips) size 9 FR were used for rectal readings. For groin readings, sterile single use Mallinckrodt Skin Temperature Probes (Thermistor 400 Series), each with hypoallergenic adhesive and foam pads, were used. For this study trained RA and PI performed thermometer placement and all data collection procedures for each newborn participant. Collaboration of temperature readings took place between trained RA and PI. Temperature measurements were obtained from: • Groin probes (DataTherm, Mallinckrodt) simultaneously placed (random) • Both axillary (DataTherm, mercury-glass) thermometers placed (random) • Rectal probe placed (2 cm depth) by PI Readings were made every two minutes after placement. After ten minutes, the glass thermometer was removed. Recordings were taken at the axillary site with 2 minutes rest. The SolarTherm was placed in the rested axilla, and held until end beep (approximately 60 seconds). After 16 minutes, all thermometer probes were removed and the ambient temperature was recorded. Temperature readings were recorded in degrees Celsius since the Celsius scale is internationally encouraged for use in healthcare situations (Holtzclaw, 1993). No attempt was made to irritate skin surface areas. DataTherm, SolarTherm, and mercury-glass temperature devices were assessed for accuracy prior to study use. A swirling water bath was used with identified temperature levels and controls, consistent with procedures implemented by previous researchers (Bliss-Holtz, 1989; Kunnel et al, 1988; Nichols and Verhonick, 1967). An ASTM mercury-glass thermometer with traceable accuracy certification accuracy (against NIST standards) identified water bath temperatures. All DataTherm, SolarTherm, and mercury-glass devices used in this study read within ± 0.1°C at three different water bath temperature 8 Temperature monitoring The Geratherm DataTherm is a small (50 gram), portable unit that continuously monitors (every four seconds) and stores (at programmable intervals) temperature data of up to 70 readings. Digital readouts are continuous (increments of 0.01°C) once the device is turned on. It has a wide temperature range (17°C to 45°C). The DataTherm has a slim, flexible sensor probe with thermistor (encased in radio-opaque, polyurethane tubing) located at the distal tip. For this study, each DataTherm probe was fitted with a single-use, thin, latex-free probe sheath, prior to use. To prevent cross-contamination, the DataTherm box was completely contained in sealed, removable single-use clear plastic. After each use, this plastic wrap was discarded and the DataTherm box and probe cleaned with isopropyl alcohol. Probes were held in place with a single-use hydro-gel adhesive patch. Mallinckrodt (reference device) BD Mercury-in-glass (reference device) BD (Becton-Dickinson) mercury-glass oral/skin thermometers were used to assess participant axilla site temperature. Thermometers were covered with single use sheaths and cleaned with alcohol between applications. The BD mercuryglass thermometer has a temperature range between 35-41°C (calibrated to 0.1°C). (Note: This thermometer is no longer available for general sale and distribution in US). All mercuryglass thermometers were shaken below < 35°C (lowest recordable temperature calibration). SolarTherm (study device) The Geratherm SolarTherm is a digital thermometer powered by ambient light with a 72 hour capacitor for holding the solar charge in darkness. The device is light (12 gram), small (13.75 cm) and battery-free with a temperature range between 32 and 43.8°C. It is portable, with a liquid crystal display and can be cleaned, sheathed, and used for oral, skin, or rectal sites. Measurement of temperature is complete in about 60 seconds JOURNAL OF neonatal NURSING VOLUME 10 ISSUE 5 2004 RESEARCH Variable Mean SD Minimum Maximum Age (hours) 21.65 14.51 1.5 79 Weight (kg) 3.42 0.56 2.396 4.69 Ambient temperature (°C) 23.56 2.20 20.1 30.0 Mean temperature (°C) 36.41 0.317 35.55 37.08 Race/ Culture White Asian Hispanic Black Native Am Total Gender Male Female 15 12 2 0 6 3 2 0 1 3 26 18 Total 27 2 9 2 4 44 Table 1. Description of sample. Site device (maximum reading) Mean(°C) SD Range DataTherm Axilla (maximum reading) 36.48 0.35 1.5 DataTherm Groin (maximum reading) Mallinckrodt Groin (maximum reading) 36.29 35.99 0.44 0.44 2.05 1.9 SolarTherm Axilla 36.53 0.44 2.0 Mercury-glass Axilla (BD) 36.70 0.45 2.0 Rectal (maximum reading) 36.82 0.31 1.4 Table 2. Mean temperatures by device/site. dwell time after which an end-beep sounds. Each SolarTherm was fitted with a single-use thin thermometer sheath, prior to use and thoroughly cleaned after each use. Results Description of the sample The 44 participating newborns averaged 21.65 (SD 14.5) hours of age and weighed, on average, 3.42 (SD 0.57) kg with 61.3% of study participants being white and male (59%). Average temperature for all participants was 36.41°C (SD 0.317) (Table 1). Comparing sites and devices Clinically, temperature variability is always present, especially when measures are displayed to hundredths, as is true with the DataTherm monitoring device. To establish the point at which temperatures recorded by DataTherm stabilised and thermal equilibrium was reached, temperature readings were made and recorded at regular intervals from the first 2 minutes to the 16th minute following probe placement. An optimum mean dwell time of 10.68 (SD 1.77) minutes for DataTherm at the axilla site and 10.82 (SD 1.33) minutes for DataTherm at the groin site was established. Mean Mallinckrodt at the groin skin site was 12.19 (SD 2.15) minutes. These skin dwell times are not notably different from recommendations of other scientists (Fulbrook, 1993; Haddock et al, 1986; Keeling, 1992). On average, groin temperatures were lower than axillae and rectal (Table 2). Skin temperatures were lower than rectal temperatures. In order to investigate site differences more completely, within subject difference scores were calculated. JOURNAL OF neonatal NURSING VOLUME 10 ISSUE 5 2004 Rectal readings remained higher than axillae readings and axillae readings were somewhat higher than groin temperature assessments. These findings were as expected (Bliss-Holtz, 1989; Craig et al, 2000; Cusson et al, 1997; Eoff et al, 1974; Smith, 2003). Eoff, Meier, and Miller in their study of 30 newborns found a mean difference of –0.49°F as the average difference between axilla and rectal sites. Similar to the findings of Browne and colleagues (2000) rectal means had the lowest standard deviation (0.31), followed by the DataTherm Axilla readings with a standard deviation of 0.35. Bland and Altman (1986) described statistical issues surrounding clinical measurements. When two measures are obtained and compared, neither measure can be assumed to be an absolute valid representation of that physical process. Therefore, looking for the degree of agreement between the two measures is beneficial and information explaining how different methods are from each other, is a good approach. Bland and Altman recommended the method of plotting difference scores (between two methods) against their mean. Agreement is considered by calculating the bias (mean difference between methods) and standard deviation of differences. Identifying confidence intervals as a range of values is a useful approach to methods comparison studies (Gardner and Altman, 1986). In contrast, correlation statistics measure strength of a linear relationship between two variables. Tests of correlation have been and continue to be widely used as a statistical technique for temperature device/site comparison investigations (Erickson and Woo, 1994; Fulbrook, 1993; Smitz, et al, 2000). Strong correlations imply that two variables change Temperature monitoring 9 RESEARCH similarly (McKenzie, 2001). Correlations are also needed for a complete picture of clinical comparisons between temperature devices and sites. To test for the assumption of normality on difference scores, the Kolmogorov-Smirnov and Shapiro Wilk tests were applied. For this investigation, there is generally insufficient evidence to doubt assumptions of normality for difference scores. Normality was not confirmed, however, for Mallinckrodt groin and rectal reading difference scores at 16 minutes. Following work with 57 neonates and using mercury-glass to assess axilla and rectal temperatures, Browne et al (2000) reported a correlation of 0.497 between sites. Working with Variable Pair Pearson r electronic instruments, Cusson and colleagues (1997) evaluated site temperature differences between tympanic, groin, axilla and rectal (n=63 neonates), They reported Pearson r correlation statistics between groin, rectal, and axillary as ranging from 0.54 to 0.58. Correlations between tympanic and rectal and axillary sites ranged between 0.32 and 0.35. For the present investigation, two study devices were compared with rectal readings with strong, statistically significant correlations. DataTherm axilla/rectal readings (r=0.741); SolarTherm axilla/rectal readings (r=0.714) (Table 3). Correlations were also statistically significant at the axilla Significance (2-tailed) N DataTherm Axilla & Rectal 16 minutes Maximum readings 0.634 0.741 <0.001 <0.001 44 44 DataTherm Groin & Rectal 16 minutes Maximum readings 0.630 0.500 <0.001 0.001 41 42 Mallinckrodt Groin & Rectal 16 minutes Maximum readings 0.282 0.472 Mercury Axilla & DataTherm Axilla 10/12 minutes Maximum readings 0.518 0.517 <0.001 <0.001 43 43 Mercury Axilla & SolarTherm Axilla 10/15 minutes 0.893 <0.001 43 Mercury Axilla & Rectal 10 minutes Maximum readings 0.657 0.680 <0.001 <0.001 43 43 SolarTherm Axilla & Rectal 16 minutes Maximum readings 0.694 0.714 <0.001 <0.001 44 44 DataTherm Axilla & Device/Site mean 16 minutes 0.758 <0.001 39 DataTherm Groin & Device/Site mean 16 minutes 0.756 <0.001 39 SolarTherm Axilla & Device/Site mean 16 minutes 0.813 <0.001 39 Mercury Axilla & Device/Site mean 16 minutes 0.849 <0.001 39 0.067 (N/S) 0.001 Comparisons Between Demographic Variables and Mean of Six Readings at 16 minutes Age -0.142 0.390 (N/S) Race -0.350 0.029 Sex -0.431 0.006 Weight (kg) 0.390 0.016 Ambient temperature 0.119 0.226 (N/S) 43 43 39 39 39 39 39 Table 3. Inter-correlations between variable pairs. 10 Temperature monitoring JOURNAL OF neonatal NURSING VOLUME 10 ISSUE 5 2004 RESEARCH site between mercury-glass and study devices (DataTherm r=0.518; SolarTherm r=0.893) and mercury-glass/rectal (r=0.680). Thus, with these newborns, stronger correlations with rectal readings were demonstrated with the two study devices. Lodha et al (2000) reported 95% limits of agreement between rectal and axillary infant temperatures between -0.3°C and 1.4°C. For this study, 95% confidence intervals for DataTherm axilla/rectal readings (-0.40 to -0.24), between DataTherm groin and rectal readings (-0.58 to -0.24), between mercury axilla/ rectal (-0.24 to -0.03) and between SolarTherm axilla/rectal (-0.35 to -0.16), were established. Comparing demographic variables For these newborns, age, race, and weight had no relationship with difference scores. This finding differs from work by Haddock et al (1986) who reported greater difference scores for black neonates. Neither was there a statistically significant relationship between difference scores and ambient temperature and difference scores and mean body temperature. That is, no systematic effect was noted between the newborn’s temperature and the device difference scores, similar to findings of LeickRude and Bloom (1998). Nagy (2001), in a study of 101 newborns, found that body temperatures of neonates did not stabilise until day five of life. Nagy identified lower temperature means day one, higher temperatures day two and three and stability after day five. These findings may explain the seemingly low mean temperature of this sample group as the mean age was less than one day. When compared with mean body temperature, these newborn temperatures related to weight and sex. Though extremes of body temperature were not identified (mean 36.41°C), heavier newborns had higher mean temperatures. Female newborns in this sample tended to have lower temperature scores. Furthermore, non-white participants tended to have lower mean body temperatures. That finding seems to concur with work by Gillum (1992) who reported that, among a sample of 7,119 children, whites had greater body temperatures. Additionally, randomly assigned right/left axilla and groin device placements did not demonstrate statistically significant relationships with temperature readings or bias scores. To further describe difference scores (mean bias) between DataTherm and rectal temperatures at three different time intervals (12, 14, 16 minutes), and compare these three bias scores (within-subjects factor) with demographic variables (between-subjects factors), repeated measures of analysis of variance were performed. Results must be viewed with extreme caution, however, because test assumptions (random, normal, sphericity) cannot be established and sample sub-sets are small. Categorical variables were created for the demographic variables of age, weight, room temperature, and mean body temperature. No statistically significant differences between subject effects were noted among variables of mean body temperature (p=0.648), age (p=0.780), ambient temperature (p=0.710), sex (p=0.119), and race (p=0.824). No statistically significant difference was demonstrated relative to DataTherm axilla/rectal bias scores at 12, 14, or 16 minute dwell times. JOURNAL OF neonatal NURSING VOLUME 10 ISSUE 5 2004 Discussion of site and device comparisons In their meta-analysis of research studies investigating temperatures measured at the axilla and rectum in children and young people, Craig and colleagues (2000) wrote that when electronic devices were used, the mean difference between axillae and rectal sites was 0.85°C (95% limits of agreement –0.19°C -1.90°C). For this study, mean differences between rectal probe and other sites were generally as expected. (Table 4). The mean difference between DataTherm axilla and rectal was -0.32°C (SD 0.24). Mean difference between mercury-glass axilla and rectal was -0.14°C (SD 0.33). Mean difference between SolarTherm axilla and rectal at 16 minutes was -0.26°C (SD 0.32), and mean difference between DataTherm groin and rectal at 16 minutes was -0.48°C (SD 0.34). Mean differences between two study references (Mallinckrodt groin and rectal) at 16 minutes were much greater at -0.86°C (SD 0.57). The greatest variability (identified with SD scores and 95% confidence intervals of difference scores) by device and site, occurred for the Mallinckrodt groin and rectal devices. When comparing maximum readings for DataTherm axilla and rectal sites, a SD of 0.32 and confidence interval range of 0.16 was established. This is a narrower range than the mercury and rectal (maximum reading) confidence interval range of 0.21. Mallinckrodt thermistor probes, used as reference for this study in groin and rectal sites, were unsheathed single use devices in direct contact with either skin or rectum. For infection control and safety, however, thin, plastic, single-use disposable sheaths were used on SolarTherm, DataTherm, and mercury-glass devices. Though sheaths have previously demonstrated negligible effects on temperature recording accuracy (Graves and Markarian, 1980), they cannot be ruled out as a possible negative influence. During the process of describing temperature study results, when more than two sites/devices were used, investigators (Molton et al, 2001) have also used the technique of comparing bias against a mean of three or more measures. This technique makes it obvious that there is no one absolute temperature reading upon which to compare test results. Thus, a mean temperature score was calculated with six temperature readings (rectal, mercury axilla, DataTherm axilla, SolarTherm axilla, DataTherm groin, Mallinckrodt groin) at the 16 minute time point. This mean temperature score was compared with DataTherm and SolarTherm device information. DataTherm axilla and DataTherm groin readings strongly correlated (r=0.758 and 0.756) with this mean temperature score. SolarTherm axilla and mercury axilla also correlated strongly (r=0.813 and 0.858) (Table 4). Notably, for these participants, the DataTherm axilla site, when compared with this mean body temperature, demonstrated a 0.00 (SD 0.24) mean bias with a negligible 95% confidence interval between -0.08 and 0.07 (Table 4). These results need to be considered with extreme caution due to the non-independence of readings. It was expected that the Mallinckrodt two referents in this study, groin skin site and rectal, would strongly correlate with each other but correlations between this site and rectal were statistically non-significant. An important consideration could Temperature monitoring 11 RESEARCH be the design of the adhesive backed sponge into which the probe rests. The 5 by 2.5 cm size made placement into the newborn groin easy, but less defined regarding direct femoral artery pulse contact with thermistor sensors. Conclusions For these study participants, the axilla site performed better than the groin skin site. Therefore, it is recommended that when skin sites are appropriate, newborn axillae be used with these devices. Importantly, when temperature comparisons are Device/Site/Time made from one point in time to another, comparisons are site and device specific. However, differences between rectal readings and devices, when available, need to be understood and communicated to neonatal clinicians and user patients. Device literature/training information needs to indicate approximate differences between readings at various sites and their considered reference to rectal temperatures in newborns. Implications Temperatures of newborn infants can fall 2-3°C after birth. It is Mean Bias °C SD DataTherm Axilla & Rectal 16 minutes Maximum readings -0.32 -0.32 0.34 0.24 -0.43 -0.40 -0.21 -0.24 -0.55 -0.39 DataTherm Groin & Rectal 16 minutes Maximum readings -0.48 -0.47 0.32 0.36 -0.58 -0.58 -0.37 -0.35 0.13 0.61 0.16 0.14 0.32 0.30 0.06 0.05 0.27 0.24 -0.94 -0.35 -0.86 -0.82 0.57 0.44 -1.0 -0.96 -0.68 -0.68 -0.90 -0.48 Mercury Axilla & DataTherm Axilla 10/16 minutes 10/10 minutes 10/12 Maximum readings 0.28 0.34 0.30 0.20 0.40 0.41 0.41 0.36 0.15 0.21 0.17 0.08 0.41 0.47 0.43 0.32 0.02 -0.04 -0.22 -0.35 Mercury Axilla & DataTherm Groin 10/16 minutes 0.44 0.40 0.31 0.57 0.10 Mercury Axilla & SolarTherm Axilla 10/15 minutes 0.14 0.20 0.07 0.21 0.16 Mercury Axilla & Rectal Maximum readings -0.14 0.33 -0.24 -0.30 -0.23 SolarTherm Axilla & Rectal 16 minutes Maximum readings -0.20 -0.26 0.32 0.29 -0.20 -0.35 -0.30 -0.16 0.01 0.10 DataTherm Axilla & Device/Site mean 16 minutes -0.00 0.24 -0.08 0.07 -1.3 DataTherm Groin & Device/Site mean 16 minutes -0.16 0.23 -0.24 -0.08 -0.41 SolarTherm Axilla & Device/Site mean 16 minutes 0.13 0.24 0.05 0.21 0.63 Mercury Axilla & Device/Site mean 16 minutes 0.27 0.25 0.19 0.36 -0.03 -0.55 0.46 -0.70 -0.40 -0.56 DataTherm Axilla & DataTherm Groin 16 minutes Maximum readings Mallinckrodt Groin & Rectal 16 minutes Maximum readings Mallinckrodt Groin & Device/Site mean 16 minutes 95% CI Lower Band 95% CI Upper Band Skew Table 4. Mean temperature (°C) differences (BIAS), SD, and 95% confidence intervals (CI) between devices. 12 Temperature monitoring JOURNAL OF neonatal NURSING VOLUME 10 ISSUE 5 2004 RESEARCH important, therefore, to accurately and carefully monitor temperatures (Sganga et al, 2000). The average participant age was 21.6 hours (SD 14.5), less than a day old. Because body temperature is not considered stable until a child is five days old (Nagy, 2001), long after most newborns have been discharged to home, parental education regarding temperature warning signs, safe use of temperature devices, and appropriate temperature sites for newborns, is essential. Parents need to be taught how to take temperatures, when to take them, and when and how to report abnormal findings. Our findings suggest that parents, in a newborn birthing setting, are ready and willing to learn more about their child’s health. Haddock et al (1986) asserted that skin temperature sites are easily learned and more acceptable sites over rectal. Limitations Study limitations include an unblinded, convenience sample of in-patient newborns. Extremes of temperature were not established; for these participants, temperatures ranged between 35.55°C and 37.08°C (a range of about 1.5°C). The draw-down effect of hydro-gel adhesive patch used to secure DataTherm thermistor probes may be a consideration. Also, rest time between axillary devices (two minutes) may not have been enough to stabilise and re-warm the axilla area. Not all data points were obtained. DataTherm probe covers fit loosely over the small, flexible probes. Upon correct positioning, adhesive was applied to the probe, with probe cover in place and therefore, though adhesion remained secure, the probe itself could move. Thus, probe movement could affect thermistor position and device readings. Though DataTherm monitoring at the axilla and groin sites needed a stabilising dwell time of about 10.5 minutes, variables such as side-to-side differences, probe adhesive, and probe cover may have been contributing factors. Further research Research is needed to develop teaching/learning interventions with parents of newborns. Quantification of learning outcomes will be essential in order to establish practice standards for this important healthcare activity. All parents in this study expressed appreciation for the additional temperature-specific information provided (written, verbal, demonstration). However, learning outcomes were not measured. With additional research that includes a larger, more diverse sample, the DataTherm monitoring device and SolarTherm thermometer may replace intermittent use long-established mercury tools. Acknowledgements The author of this investigation acknowledges financial support, as received by the Oregon Health & Science University, of two study sponsors: Geratherm Medical Diagnostic Systems and RG Medical Diagnostics. The donation of thermometer devices, sheaths, and hydro-gel adhesive patches from RG Medical Diagnostics is acknowledged and sincerely appreciated. Thanks go to the Mallinckrodt Corporation for their generous donation of single-use thermistors and on-loan temperature monitor. JOURNAL OF neonatal NURSING VOLUME 10 ISSUE 5 2004 References Ash, C.J., Cook, J.R., McMurry, T.A., Auner, C.R. (1992) The use of rectal temperature to monitor heat stroke. Missouri Medicine 89(5): 283-88. Bland, J.M. and Altman, D.G. (1986) Statistical methods for assessing agreement between two methods of clinical measurement. The Lancet 1(8476): 307-10. Bliss-Holtz, J. (1989) Comparison of rectal, axillary, and inguinal temperatures in full-term newborn infants. Nursing Research 38(2): 85-87. Broome, M.E., Dokken, D.L., Broome, C.D., Woodring, B., Stegelman, M.F. (2003) A study of parent/grandparent education for managing a febrile illness using the CALM approach. Journal of Pediatric Health Care 17(4): 176-83. Browne, S., Coleman, H., Geary, E. et al. (2000) Accurate measurement of body temperature in the neonate: A comparative study. Journal of Neonatal Nursing 6(5): 165-68. Buck, S.H. and Zaritsky, A.L. (1989) Occult core hyperthermia complicating cardiogenic shock. Pediatrics 83(5): 782-84. Carlson, K.K. (1996) Vital sign assessment. In: Craven, R. F. and Hirnle, C. J., ed. Fundamentals of Nursing: Human Health and Function: Second Edition. Philadelphia: Lippincott-Raven Publishers. 423-33. Cattaneo, C.G., Frank, S.M., Hesel, T.W., El-Rahmany, H.K., Kim, L.J., Tran, K.M. (2000) The accuracy and precision of body temperature monitoring methods during regional and general anesthesia. Anesthesia & Analgesia 90(4): 938-45. Clarke, S. (1992) Use of thermometers in general practice. British Medical Journal 304(6832): 961-63. Crafts, R.C. (1985) A Textbook of Human Anatomy: Third Edition. New York: John Wiley & Sons. Craig, J.V., Lancaster, G.A., Williamson, P.R., Smyth, R.L. (2000) Temperature measured at the axilla compared with rectum in children and young people: Systematic review. British Medical Journal 320(7243): 1174-78. Cusson, R.M., Madonia, J.A., Taekman, J.B. (1997) The effect of environment on body site temperatures in full-term neonates. Nursing Research 46(4): 202-07. Eoff, M.J.F., Meier, R.S., Miller, C. (1974) Temperature measurement in infants. Nursing Research 23(6): 457-60. Erickson, R.S. and Woo, T.M. (1994) Critical care in children: Accuracy of infrared ear thermometry and traditional temperature methods in young children. Heart & Lung The Journal of Critical Care 23(3): 181-95. Fulbrook, P. (1993) Core temperature measurement: A comparison of rectal, axillary, and pulmonary artery blood temperature. Intensive and Critical Care Nursing 9(4): 217-25. Gardner, M.J. and Altman, D.G. (1986). Confidence intervals rather than P values: Estimation rather than hypothesis testing. British Medical Journal 292: 746-50. Gillum, R.F. (1992) Body temperature and its relationship to demographic and cardiovascular risk factors in a national sample of children and adolescents. Journal of the National Medical Association 84(7): 591-99. Graves, R.D. and Markarian, M.F. (1980) Three-minute time interval when using an oral mercury-in-glass thermometer with or without J-temp sheaths. Nursing Research 29(5): 323-24. Greenes, D.S. and Fleisher, G.R. (2001) Accuracy of a noninvasive temporal artery thermometer for use in infants. Archives of Pediatrics & Adolescent Medicine 155(3): 376-81. Haddock, B., Vincent, P., Merrow, D. ( 1986) Axillary and rectal Temperature monitoring 13 RESEARCH temperatures of full-term neonates: Are they different? Neonatal Network 5(1): 36-40. Holtzclaw, B.J. (1993) Monitoring body temperature. AACN Clinical Issues in Critical Care Nursing 4(1): 44-55. Hooker, E.A., Smith, S.W., Miles, T., King, L. (1996) Subjective assessment of fever by parents: Comparison with measurement by non-contact tympanic thermometer and calibrated rectal glass mercury thermometer. Annals of Emergency Medicine 28(3): 313-17. Keeling, E.B. (1992) Thermoregulation and axillary temperature measurements in neonates: A review of the literature. MaternalChild Nursing Journal 20(3-4): 124-40. Kunnel, M.T., O’Brien, C., Munro, B.H., Medoff-Cooper, B. (1988) Comparisons of rectal, femoral, axillary, and skin-to-mattress temperatures in stable neonates. Nursing Research 37(3): 162-164, 189. Ladewig, P.W., London, M.L., Olds, S.B. (1998) Maternal-newborn nursing care: The nurse, the family, and the community: Fourth Edition. Menlo Park, CA: Addison Wesley Longman, Inc. Leick-Rude, M.K. and Bloom, L.F. (1998) A comparison of temperature-taking methods in neonates. Neonatal Network 17(5): 21-37. Lodha, R., Mukerji, N., Sinha, N., Pandey, R.M., Jain, Y. (2000) Is axillary temperature an appropriate surrogate for core temperature? Indian Journal of Pediatrics 67(8): 571-74. McKenzie, N.E. (1998) Fever: Upping the body’s thermostat. Nursing 28(10): 41-45. McKenzie, N.E. (2001) The thermometry page: Sites and issues. McKenzie & Associates, LLC. www.graduateresearch.com/thermometry/sites.htm (Jan 4 2002). Molton, A.H., Blacktop, J., Hall, C.M. (2001) Temperature taking in children. Journal of Child Health Care 5(1): 5-10. Nagy, E. (2001). Gender-related differences in rectal temperature in human neonates. Early Human Development 64(2001): 37-43. Nichols, G.A. and Verhonick, P.J. (1967) Time and temperature. American Journal of Nursing 67(11): 2304-06. Payne, D., Johnson, A., McKenzie, S., Rogers, M. (1994) Chemical and glass thermometers for axillary temperatures: How do they compare? Archives of Disease in Childhood 71(3): 259-60. Robinson, J.L., Seal, R.F., Spady, D.W., Joffres, M.R. (1998) Comparison of esophageal, rectal, axillary, bladder, tympanic, and pulmonary artery temperatures in children. The Journal of Pediatrics 133(4): 553-56. Sganga, A., Wallace, R., Kiehl, E., Irving, T., Witter, L. (2000) A comparison of four methods of normal newborn temperature measurement. MCN, The American Journal of Maternal/Child Nursing 25(2): 76-79. Smith, L.S. (2003) Reexamining age, race, site, and thermometer type as variables affecting temperature measurement in adults – a comparison study. BMC Nursing 2:1. Smitz, S., Giagoultsis, T., Dewe’, W., Albert, A. (2000) Comparison of rectal and infrared ear temperatures in older hospital inpatients. Journal of the American Geriatrics Society 48(1): 63-66. Statistical Package for the Social Sciences [SPSS]. (2002) SPSS 11.5 for Windows. Chicago, Illinois: SPSS, Inc. Stephen, S.B. and Sexton, P.R. (1987) Neonatal axillary temperatures: Increases in readings over time. Neonatal Network 5(6): 25-28. Torrance, J.T. (1968) Temperature readings of premature infants. Nursing Research 17(4): 312-20. Whaley, L.F. and Wong, D.L. (1989) Essentials of Pediatric Nursing: Third Edition. St. Louis: The C. V. Mosby Company. Zehner, W.J. and Terndrup, T.E. (1991) The impact of moderate ambient temperature variance on the relationship between oral, rectal, and tympanic membrane temperatures. Clinical Pediatrics 30(4 suppl): 61-64. 14 Temperature monitoring JOURNAL OF neonatal NURSING VOLUME 10 ISSUE 5 2004