Send letter to journal:
Re: Accessing electronic information for clinical decisions

Dear Editor,

I enjoyed the recent leading article on accessing electronic information for clinical decisions.[1] In this article, R S Phillips describes a scenario in which a child presents with presumed pneumococcal meningitis. The admitting paediatrician attempts to make an evidence-based decision on the use of dexamethasone to try and prevent hearing loss. The paper provides a useful overview of how information technology can be accessed in a medical emergency. However, I would like to comment on a number of points regarding both the electronic search and the management of the scenario.

Phillips mainly discusses the use of resources in which the evidence has already been reviewed and evidence-based guidelines have been produced. At one point Phillips claims that the use of sources other than those with these “preprocessed” formats is unrealistic. I took this as a challenge, and conducted the “pneumoccal meningitis and dexamethasone” search on Medline. By restricting the search to sytematic reviews and meta-analyses, I found one paper and then accessed the abstract.[2] I presume that the paper identified is the same meta-analysis as that mentioned by Phillips. The whole search took three minutes, which is well within the time constraint applied in the scenario. I believe that it is important to realise that Medline searches can be used in an emergency. Unlike some of the resources mentioned by Phillips, Medline is widely and freely available, both on the internet and on hospital computer networks. Furthermore, as Phillips himself found, the number of scenarios covered by the preprocessed formats is limited. Medline, on the other hand, should be able to offer some information on almost every subject.

In the clinical scenario, the registrar conducts the computerised search after the CSF microscopy results have been phoned through and while the antiboitics are being drawn up. In reality, I suspect that many paediatricians would start treatment as soon as the lumbar puncture had been performed. This would certainly be my practice where, as in the scenario, the child has convincing signs of bacterial meningitis. This presents a problem because the abstract of the meta-analysis does recommend the use of steroids in pneumococcal meningitis, but only if given early.[2] Reading the meta-analysis later, I realised that early treatment means giving dexamethasone before, or at the same time as, the antibiotics. Thus, by the time the registrar decides to give dexamethasone, it will be too late. This is, perhaps, an example of a limitation on the use of evidence-based medicine in real life. It would probably be preferable to have evidence-based guidelines available within the department for such an eventuality.

Finally, at the end of the article Phillips leaves us with the child discharged from hospital and awaiting an appointment with the audiologist. The registrar does not know if his intervention may, or may not, have been effective. This need not be the case. A further literature search would reveal that hearing tests can safely and effectively be performed before discharge from hospital.[3]

Martin Richardson
Peterborough District Hospital
Thorpe Road, Peterborough PE3 6DA

References
(1) Phillips RS. Accessing electronic information for clinical decisions. Arch Dis Child 2000;83:373-4.

(2) McIntyre PB, Berkey CS, King SM, Schaad UB, Kilpi T, Kanra GY, Perez CM. Dexamethasone as adjunctive therapy in bacterial meningitis: a meta- analysis of randomized clinical trials since 1988. JAMA 1997;278:925-31.