We thank Ng et al for their response to our paper.1 They raise the interesting issue of gold standards for the measurement of fever. To our knowledge, there is no such single accepted and universally agreed gold standard. In their review article, they summarise the differences between axillary, rectal, tympanic, and pulmonary artery measures using different types of thermometers, but do not present original data to support their assertion that rectal thermometry represents such a ‘gold standard’.2 We were conscious of this issue when we analysed our data and wrote the paper; for this reason we discussed in our introduction the rationale for why we chose to use both measures of agreement and diagnostic accuracy.
The issue of rest prior to thermometry would be important if this was routine in clinical practice. However, we see the results of our study as generalisable to the real world in which children arrive in surgery in varying states of unrest.
It appears that we agree with Ng et al more than we disagree: we agree that thermometry using different instruments in different body sites and in different age groups leads to measurement differences. Given that treating fever in the home is widespread,3 the questions that still need addressing are: 1) rather like the differences between clinic and ambulatory blood pressure measures,4 which instrument and site best predicts the child's prognosis as opposed to the immediate diagnosis of fever? 2) when should a thermometer be used? and 3) when, and if, a gold standard is established, which instrument and site offers the best agreement?
- © British Journal of General Practice, 2004.