Selection of the Most Accurate Thermometer Devices for Clinical Practice: Part 1

Meta-Analysis of the Accuracy of Non-Core Thermometer Devices Compared to Core Body Temperature

Nancy A. Ryan-Wenger; Maureen A. Sims; Rebecca A. Patton; Jayme Williamson


Pediatr Nurs. 2018;44(3):116-133. 

In This Article


Accuracy of non-core Thermometer Devices

This is the first meta-analysis of carefully designed studies that used appropriate sites for core body temperatures compared to temperatures from non-core thermometers devices. We identified 34 high-quality studies in which 97 total sample comparisons were made between core and non-core temperatures. Forest plots in Figures 2 to 6 illustrated many dramatic differences in the accuracy of temperatures from a variety of non-core thermometer devices when compared to core body temperatures. Non-core devices are not expected to exactly reflect core body temperature because of their relative distance from the core; thus, it is important to know the extent of under- or over-estimation and the consistency of measurements as reflected by the width of confidence intervals. The risk associated with wide variability in temperatures is the potential for unintended consequences associated with the measurements; underestimation of actual temperature may lead to missed diagnosis or treatment, while overestimation of actual temperature may lead to unjustified diagnosis or treatment.

The pooled mean differences (ES) between core body temperature and temperature from electronic oral and rectal temperatures were closest to the line of reference (zero difference) and had smaller confidence limits compared to the other devices (see Figure 6). Thus, oral and rectal electronic thermometers, when used according to manufacturer recommendations, are most likely the safest devices for use in clinical practice, but meticulous attention to the procedural steps is essential every time these devices are used. It is important to note that the thermometer devices themselves are not always at fault. Manufacturers report their devices are tested and highly accurate in controlled laboratory settings, and are approved by the Federal Drug Administration. In these highly controlled clinical studies, in which data collectors were expected to be meticulous about following protocols for each device, there were still wide variations in accuracy. We are concerned that in clinical settings, where nurses and unlicensed assistive personnel are initially trained to use new types of thermometer devices, temperature-taking may become routine, and attention to all details of operating a device may wane over time. Just as reliability is a pre-requi-site for validity in psychometrics (Cohen, 1988), repeatability is prerequisite for accuracy in temperature measurement. Only one of six authors who evaluated intra- and inter-rater reliability used appropriate, meaningful statistics; however, Bland and Altman's (1986) classic article was cited as a reference.

It was disappointing to find that 83 research studies identified in our literature search were ineligible for meta-analysis because non-core temperature sites, primarily rectal, were used as the gold standard for comparing other non-core thermometer devices. Researchers continue to generate more unusable data by replicating these studies, and journal editors continue to publish them. Our concern is that with the increased emphasis on evidence-based practice, healthcare providers may change practice and policy based on "evidence" from these studies.

Data from 97 comparisons of six non-core thermometer devices with core body temperature in this meta-analysis are compelling; thus, we see little need for more studies of this type. However, oral chemical thermometers are widely used in health-care settings, but we found only one study on their accuracy. More research is needed on that device.