This paper (Farnell et al. 2005) is an interesting study that makes a useful contribution to the debate surrounding the use of different methods of clinical thermometry. However, in doing this, the authors show a somewhat overzealous belief in the inherent accuracy and usefulness of any method of clinical thermometry. This leads them to make a very general statement about the use of tympanic thermometers, which is not supported by these data and is certainly not generalizable to other clinical or non-clinical settings from either this study or the studies quoted.
The problems with their conclusion are twofold. The first is the idea that there should necessarily be an obvious relationship between different thermometry sites. The authors note that only 11·1% of TempaDOT and 5·2% of tympanic temperature measurements were in exact agreement with the pulmonary artery temperature. The temperature in the pulmonary artery, in the absence of hypothalamic measurement, is probably the nearest that there is to a central temperature. Heat is produced within the body, primarily from the muscles and other metabolically active tissues, and is then transported around the body to the peripheries, mainly through the bloodstream (Boulant 1997, Stanier et al. 1984).
Therefore, it is to be expected that a temperature found deep within the body is very likely to be different to that found in other parts of the body because of a temperature gradient from metabolically active to less active tissues. However, it is not necessarily the case that this difference should be predictable or necessarily consistent. There is evidence of a lack of consistency between core and peripheral sites in both adults (Thomas et al. 2004) and young children (Wailoo et al. 1989) for a wide variety of reasons (Mackowiak 1997). There seems little point in worrying about this, what is important is acknowledging that this is the case and modifying decision-making processes accordingly.
The second major issue that the paper (Farnell et al. 2005) discusses is the ability of the different techniques to inform clinical decision making. However, there are issues both with lack of a single normal, and so abnormal temperature, and the necessity of some of the interventions mentioned. With regard to ‘normal’ temperature, Wunderlich noted as far back as the 1870s that there was variation between the temperatures of healthy people (Wunderlich 1871), a finding supported by the research some 120 years later, which found a range of oral temperatures among healthy American adults of 35·6–38·2 °C (Mackowiak et al. 1992). Although it may appear to make life easier to think that there is one normal temperature, it is more accurate to consider temperature as being analogous to blood pressure and many other physiological indicators that vary between individuals and over time. Although this study did not aim to define normal temperature, it did assume levels of abnormality that necessitated the commencement of certain clinical interventions.
Although the precise details of some of these interventions are not clear, a number of these are, for the generality of patients, not necessary. It is important to remember that fever is a normal physiological response to some kind of insult, normally infection. As such, fever itself (as opposed to the underlying condition) rarely requires treatment, although there may be particular cases where this is necessary. Therefore, there is no indication for the routine administration of paracetamol, removing blankets, tepid washing, or the use of ice, let alone at any specific temperature. Furthermore, as fever is the result of an interaction between circulating prostaglandin E (PGE), produced in response to pyrogens, and the hypothalamus (Aronoff & Neilson 2001), there is never any benefit in removing blankets, tepid washing or using ice in the absence of antipyretic medication to reduce the effect of PGE. If the body does cool thorough such measures, it will simply institute heat promotion measures to restore the body temperature.
Because of this normal variation in temperature, it is not satisfactory for nurses to base clinical decision making on temperature alone and if a more thorough assessment is taken, there is less chance of patients receiving delayed or unnecessary treatment, one of the outcomes used in the paper. The problem with this particular approach and subsequent analysis is not, therefore, just its subjectivity (which the authors acknowledge), but its validity as a measure of the effectiveness of different methods of thermometry.
The authors conclude, somewhat boldly with the statement, that ‘it is becoming increasingly difficult to defend the continued use of tympanic thermometry in clinical practice’. On the contrary, it is very easy to defend its use as long as the limitation of this method is remembered, especially as the authors themselves note the potential problems associated with the TempaDOT, which included seven cases where no temperature could be recorded at all. The defence is that they are quick, popular with patients (Barton et al. 2003) and even in this study, 74% were within 0·6°C of the pulmonary artery temperature. There may be particular settings where a higher level of accuracy is required, but when all of the other variables that affect temperature recording are taken into account, such as natural variation in normal temperature, activity, time of day and gender (Mackowiak et al. 1992), this level of accuracy may actually be sufficient.
Although this critique has necessarily been of this paper, in many ways the problem with the current clinical use of thermometry is much wider, because this paper (Farnell et al. 2005) does represent the way the medical and nursing establishments interpret different temperatures. The important thing is to treat the patient, not the thermometer, if this is remembered many of the problems that the authors discuss will be alleviated.