The setting

It's early on a Friday evening, and you're working as registrar in A&E. A 6 year old girl is rushed in as an emergency, complaining of headache; she is febrile with convincing nuchal rigidity. Your colleague performs a lumbar puncture, then you attempt to achieve intravenous access, and by the time the anaesthetic cream has cooked the microbiologist rings from the lab. The CSF contains 50 polymorphs, two red cells and is “teeming with Gram positive diplococci”.

The question

Before you give an antibiotic though, you wonder about the steroids in meningitis debate. Should this girl get dexamethasone prior to antibiotics? Will it reduce her chance of hearing loss—or just increase her chance of continued infection? You reckon the five minutes it will take to draw up the antibiotic is enough time to try to find some information to help.

The problem

In the setting of acute paediatrics, information needs to be delivered to those who provide care as quickly as possible. It has been suggested that immediate information should be accessible within 15 seconds, further information within three minutes, and a digest of some detail in around 10 minutes.1 The only way this is possible is by utilising electronic information sources. Furthermore, the information should be “evidence based”; informed by the most valid clinical research available. Finally, the information needs to be applicable in the local settting, taking into account local services, populations, and peculiarities.2

Nearly all hospitals have an internal computer system, but these are primarily “datanets”—systems designed to transport laboratory data. Some systems are also “knowledge nets”—providing support for clinical decision making. The transition between a simple datanet to a knowledge net can be achieved by allowing computers to access the Internet—no mean feat for most hospital information management and technology departments.

Assuming there is access to electronic information resources, there is then a bewildering array of databases, knowledge banks, and repositories from which to choose. Within the proliferation, certain names stand out for their comprehensive nature, ease of use, or high quality of information. Unfortunately, there is yet no location which can be said to harbour all three qualities. Table 1 highlights some strengths and weaknesses.

The search

Back to the plot, though. In the setting of A&E, with access to a computer linked to the Internet, we can trawl through some databases to seek an answer. PIER, the Royal College of Paediatrics and Child Health sponsored collection of paediatric information at Sheffield, is a good example of a guideline repository.3 When looking for information in a hurry, there is a good search engine with clear results. In the case of “meningitis”, three guidelines are found: two describe steroids as useful only in HiB meningitis, the third implies dexamethasone is a standard part of the protocol. No guidelines reference their source material, nor state how they were collated (are they consensus statements? “evidence based” guides?).

The BMJ Publishing Group has a dedicated team writing and updating biannually Clinical evidence.4This is a book, also available via the Internet, which aims to provide succinct summaries of high quality evidence relating to the treatment of common conditions. It is aimed at a very general audience, and currently contains few chapters relating to acute paediatrics. If, as the registrar in A&E, you could remember your password, there would be no information of relevance to your current plight.

With time running out, another option may be to use a portal which accesses multiple information sources: you type in a question, and the portal looks up the answer on a number of different databases. “SumSearch” from the University of Texas in San Antonio,5 or Ovid's “Evidence Based Medicine Reviews”6 are such products. SumSearch allows a single search request to access a myriad of databases: a textbook of medicine, the abstracts of the Cochrane Collaboration reviews, full text of the “DARE” review collection, and even focused searches of Medline. It is freely available over the Internet, and although a little slow at peak times, provides high quality answers quickly. In response to “pneumococcal meningitis and dexamethasone”, SumSearch tidied up the search and provided links to a systematic review and meta-analysis. Similar information may soon be available within the NHS from the National Electronic Library of Health.7

Evidence Based Medicine Reviews from Ovid6 performs a similar task, with the addition of full text Cochrane and searching Best Evidence too, but is a subscription access product. The search “pneumococcal meningitis and dexamethasone” provides a reference to the same review, but in a predigested format with clinical commentary attached (from Best Evidence).

The future

It is unrealistic to look for answers in clinical time in anything but these “preprocessed” formats. As hinted above, there are a selection of databases which back up the process—Cochrane, DARE, Best Evidence—and there are new projects developing such resources.

These developing knowledge banks contain easily digested summaries of the evidence, with the facility for more detailed investigation if time allows. The format used in many centres practising evidence based child heath is the “critically appraised topic” (CAT). This is in effect a highly structured abstract, written independently, with commentaries on both methodological and clinical issues.

Evidence Based On Call (EBOC),8 a team from the Centre of Evidence Based Medicine (Oxford), has produced such a resource for general medicine, and is working on expanding this into paediatric practice. These CATs are independently verified by two researchers and reviewed by a clinician working in the appropriate field. There are also a number of sites in the USA and UK which have begun to collate such summaries (see table 2). Many of these sites incorporate mechanisms by which practising clinicians can contribute to the emerging wealth of knowledge.9

The answer

The abstracted version of the meta-analysis convinces you to give intravenous dexamethasone with the antibiotic. The child is confirmed to have pneumococcal meningitis (penicillin sensitive) by microbiological testing, and is discharged home a fortnight later. There is an appointment with the local audiologist coming up—but no hearing test as yet.