Commentary: Mona Okasha

Mona Okasha takes you through this paper and explains what it means

Over recent decades, doctors have been more and more encouraged to practise evidence based medicine - to base their medical decisions on research evidence - although clinical experience is still important. One of the foremost efforts in this area is the Cochrane Collaboration (www.cochrane.org), which publishes systematic reviews covering all areas of health care. But do doctors at the front line of treating patients change their beliefs or practice in the light of published research?

Phanie agency/rex

Wake and Hesketh describe the perceptions of five groups of health professionals with regard to the prevalence of symptoms associated with teething. They note in the introduction that previous studies have found that teething is associated with minor and infrequent symptoms.

Study design issues

The authors did a cross-sectional survey, using published state registers of professional bodies to identify their target population. The groups targeted were maternal and child health nurses, pharmacists, general practitioners, dentists, and paediatricians. A postal questionnaire was sent to 550 health professionals; 85% responded.

Asking individuals about their perceptions or beliefs is rarely ideally studied by questionnaire. This method tends to force respondents to tick a certain box and characterises their beliefs according to the researchers who designed the questions. Qualitative work - in which the health professionals would have been interviewed by a trained researcher - tends to elicit a broader range of answers, and reasons behind a persons beliefs can be explored in a way that is not possible quantitatively.

How evidence based were the health professionals beliefs?

The results indicate that what health professionals believe about teething does not concur with the current evidence base. The average number of symptoms associated with teething ranged from 2.8 (reported by paediatricians) to 9.8 (reported by nurses). The proportion of professionals who thought that all or most children have teething problems ranged from about 25% (dentists and paediatricians) to 75% (nurses). These results are not compatible with those referenced by the authors which suggest that teething problems are relatively infrequent and minor.

Why do health professionals report beliefs so divergent from scientific research? Perhaps the studies suggesting that the symptoms of teething are infrequent and minor were not well designed or executed. Perhaps the questionnaire did not ask valid questions about teething. Perhaps health professionals do not keep up to date with the evidence. This is likely, since thousands of medical journals are published each week. You cannot blame a health professional for keeping abreast of only some of the current research, and it is unlikely that teething is a top priority.

Presenting your results

The authors presented their results in two rather wordy paragraphs because of strict space restrictions in the BMJ (for a short report only 600 words, 5 references and 1 table or graph are allowed). Generally, using graphs or tables rather than text is best. Remember to choose the right kind of graph for the type of data that you are presenting. If you are presenting percentages, it is often better to use pie charts rather than bar charts as this shows a better representation, making it easier to compare.

Reproducing data collection instruments When interpreting data from questionnaires, it is important to know what questions were asked. Two questions aiming to ask the same thing may elicit different answers from respondents, so a suitable critical appraisal of a paper requires an understanding of how the research was collected. For example, it is not clear from this study whether respondents were given a list of symptoms from which to choose what they would ascribe to teething, or whether they were asked to create such a list themselves.

Implications of this work

On the face of it, this study may not be earth shattering, but it brings two important points to light. Firstly, the authors note that if certain symptoms are ascribed to teething, suitable management of the true underlying illness may be delayed, with potentially serious consequences. Secondly, health professionals may prescribe or advise unnecessary medication in the face of a misdiagnosis. Dont forget that evidence based medicine is not just an ideal suggested by academic health researchers but helps you make decisions and ultimately benefit your patients