Article Text


On Archimedes
  1. T Lopez1
  1. 1Bristol Royal Hospital for Children; tonylopez{at}
  1. B Phillips2
  1. 2Evidence-based On Call, Centre for Evidence-based Medicine, University Dept of Psychiatry, Warneford Hospital, Headington OX3 7JX, UK; bob.phillips{at}

Statistics from

Using the best available evidence is expected of us in clinical practice. How should clinicians get such evidence? Should we all be formulating questions, searching for the evidence and then appraising it? Or as busy clinicians are we forced to rely on the evidence provided for us in published systematic reviews. Rudolf’s recent paper puts one side of the argument.1 Nine doctors attending at MmedSc Course each spent an average of five hours analysing a clinical problem “in accordance with the principles of evidence based practice”. As a result of this work they judged themselves to have improved in structuring clinical questions, searching electronic databases, and in critical appraisal. In addition they succeeded in highlighting the poor evidence upon which we base much of our practice. I have no doubt that their efforts had an educational value, but would they be right to base their clinical practice on the conclusions of five hours work?

In November 2001, as part of the Archimedes series, two middle grade paediatricians attempted to answer the following question: in a feverish infant, how accurately does tympanic thermometry measure core temperature?2 They took rectal temperature to reflect core temperature and restricted their search to work on children. They found two directly relevant studies and one systematic review. On this they based their analysis.

In August 2002, Craig and her colleagues published a systematic review comparing tympanic thermometry and rectal thermometry in children.3 They searched eight databases and checked through numerous reference lists. They contacted authors and suppliers of clinical thermometers. They found 44 studies eligible for inclusion, including two unpublished papers and five written in languages other than English. The process of searching for and identifying eligible articles took approximately 80 hours spread over several months (V Craig, personal communication). Given the huge disparity in the number of identified papers, it is surprising that the results of both reviews were similar: in an individual patient, tympanic thermometry may not accurately mirror the rectal temperature. However in the details they differed. Riddell and Eppich tell us that “age and presence of fever significantly affected the rectal tympanic difference”.2 Craig et al showed that this was not the case, “there was no systematic relation between the temperature difference and the underlying temperature” and similarly “we found no association between temperature difference and the age of the children”.3

Riddell and Eppich found 3 papers. Craig et al had found 44. What does this tell us?

Answering clinical questions by appraising the available evidence is justifiably the new creed. But done quickly, it risks being done badly. The search for evidence, and its analysis, is best left to those with the necessary time and expertise. The urge to join in is understandable. It should be resisted. Those of us in busy clinical posts should assess the results of thorough systematic reviews and then, in the words of Sackett and his colleagues, conscientiously, explicitly and judiciously, use them to make decisions about the care of our patients.4 If we are honest with ourselves, we really haven’t time for anything else.


Author’s reply

As editor of Archimedes, and victim of his play,1 I have the pleasure in responding to the concerns raised by Dr Lopez. I think there are two—a concern with the philosophy of Archimedes and a problem with the tympanic topic.

There is a firm and widely held belief that evidence based practice can be achieved only by those “with the necessary time and expertise”, and that we should only change our practice after assessing “the results of thorough systematic reviews and . . .conscientiously, explicitly and judiciously, use them to make decisions about the care of our patients”.

The position of Archimedes, and I’d guess many clinicians who believe they practice in an evidence based fashion,2 is nearly the opposite. Finding questions, seeking answers, assessing the best of the evidence to answer that question, applying it, and assessing the results is a cycle we all (should) perform in some way or another. I think that’s all “evidence based” practice really is. The famous Sackett quote about evidence based medicine being “the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients”3 relies on the current best evidence: not just systematic reviews (which might be categorised as “the best evidence which can be produced”).

Dr Lopez asks “Riddell and Eppich found three papers. Craig et al had found 44. What does this tell us?”. Well, possibly that Dr Lopez didn’t read the Riddell report4 very closely, as their commentary starts with “The systematic review . . .[had] 44 studies addressing the use of different methods of temperature measurement” Accordingly, it’s not at all surprising that the papers come to the same bottom lines. Any minor differences—reported as comments on subsidiary papers with a much lower level of evidence—wouldn’t be that clinically exciting, would they?

But anyway, I think that Dr Lopez and myself would agree on one thing. Compared to starvation in Zimbabwe, an impending Gulf war, and children being raped by their parents, this isn’t all that important.


View Abstract

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.