Towards evidence based medicine for paediatricians
Archimedes - Towards Evidence-Based Paediatrics
The principles of evidence-based healthcare have been widely accepted across a variety of health care settings, and a textbook of Evidence-based Pediatrics and Child Health is available to educate and inform paediatricians about its' practice. Their definition is "the integration of clinical information obtained from a patient with the best evidence available from clinical research and experience, and the application of this knowledge to the prevention, diagnosis or management of disease in that child" . In order that this may be performed efficiently, summaries of the best evidence are helpful to guide practice.
The format we are using is based on the Best BETS (http://www.bestbets.org) template. To register a question, and to submit completed Archimedes topics, please check the BestBETS site or email . The creation of such a topic summary follows this process:
- Selection of a clinical scenario
Scenarios which trigger searches for evidence-based medicine are best if they address common questions, which are important for the patients, doctor, health system or all three.
- Definition of the clinical question
Structured clinical questions improve one's efficiency when searching for answers, and narrow down the area of enquiry. What might start as "How do you treat croup?" becomes focussed into "In a child with croup, does nebulised budesonide work better or faster than oral dexamethasone?" The structured clinical question has three (or four) parts: the patients, the intervention/test, possibly a comparison, and the outcome(s) of interest.
- Search for answers
Searches are best if they look for the best quality evidence first, then work down the hierarchy of evidence (http://cebm.jr2.ox.ac.uk/docs/levels.htm). In order to do this, searches are first performed on secondary sources (Cochrane Library, Best Evidence and Clinical Evidence), then repeated for either more recent information or any information on a Medline system. PubMed, the NLM's freely available Medline service, is the commonly used search engine. If a systematic review is found (and of good quality), then a search forwards in time, to see if anything has been published after the review, is necessary.
The technique is to use a broad search on the secondary sources - for instance the patients' condition. If this produces an unreasonably large number hits, narrow with the intervention. If this is still an unreasonably large number, try using the comparator or an outcome measure until a reasonable number are listed, then scan the abstracts. The search words can be free text (if only a few hits are found) or MeSH terms (indexed words) if too many articles are discovered.
The search on Medline follows a similar strategy, but with the use of Clinical Question Filters.
Another person should repeat this search, and a consensus about the results reached.
- Appraise the evidence
The appraisal of evidence is the important step in the process. All studies require appraisal - from Cochrane reviews to case series from Orthopaedics Weekly. There are many texts assisting the critical evaluation of studies, the classic text is by Sackett et al , or the first chapters in the Evidence-based Pediatrics and Child Health textbook .
- Create a critically appraised topic (CAT)
The distillation of the critical appraisal into a CAT is important to allow others to evaluate the conclusions reached. A CAT is a detailed summary of the critical appraisal of the individual paper. In order to do this, CATMaker, a programme from the Oxford Centre for Evidence-based Medicine has been developed (cebm.jr2.ox.ac.uk), although any format is acceptable.
- Summarise as a best evidence topic (BET)
The best evidence topic is the final accumulation of the critical appraisal. The strict format allows the casual reader to extract important information quickly and easily. Some notes on the template are below:
A question to tease the reader into your article.
Short description of the clinical episode.
Structured clinical question
Identifying the 3 or 4 parts of the question
Details of the search strategy used, and the number of studies found. This is abbreviated greatly in the paper version.
As a table; this identifies the studies, the population group and size, the study design (and level of evidence), the key outcomes and their results. Any vitally important points of critical appraisal are also noted here.
A short narrative, summary of the evidence and its application.
Clinical bottom lines
A summary of the important messages from the BET - 3 at the most.
In general, the past tense is preferred for all answers based on studies of a lower quality than a good systematic review, and the use of terms such as 'may' or 'might' for information of very low quality.
Name and designations, only please.
The BETs within Archimedes are the best evidence summaries a real clinician can produce. They represent the best of the available evidence applied in real paediatric settings. They can be created within journal clubs, training schemes or by interested clinicians at any level of training. In time, this archive will demonstrate the strengths and weaknesses of the evidence upon which paediatric care is based.
To register a question, and to submit completed Archimedes topics, please check the BestBETS site or email.
BestBETs are an original format developed by Kevin Mackway-Jones and Simon Carley at the Manchester Royal Infirmary.
(1) Evidence-based Pediatrics and Child Health. Moyer V, Elliot EJ, Davis RL et al. BMJ Books: London 2000
(2) Evidence-based Medicine: How to Practise and Teach EBM. Sackett DL, Straus SE, Richardson WS, Rosenberg W, Haynes RB Second Edition. Churchill Livingstone: Edinburgh, 2000.
About the toc
doi: 10.1136/adc.85.3.252 ADC 1 September 2001 vol. 85 no. 3 252-255
- Instructions for Contributors
- Interactive Table Index