Research can produce false-positive results just as can diagnostic tests.
Uncontrolled studies have a specificity of only 11%, versus 88% for randomized controlled trials (RCTs), which have been designed to minimize the bias of investigators toward a positive outcome.
A search of all the scientific studies in Medline since 1985 revealed 5,842 publications on prehospital EMS, but only 54 were RCTs (and therefore unlikely to produce false-positive results).
By way of comparison, during the same time hundreds of RCTs have been conducted on major medical emergency conditions, and RCTs on even minor topics such as urticaria and constipation exceed the scientific database on all of EMS.
Of the 54 EMS RCTs, 4 (7%) reported harm from the new therapy, and 74% reported no effect of the new therapy at all.
Only 7 (13%) RCTs showing a positive outcome of the intervention were uncontradicted ; of these only 1 examined a major outcome such as survival, and only 1 compared the intervention with a placebo and could therefore evaluate the efficacy of EMS itself.
Because there is such a paucity of scientific support for EMS interventions and because monitoring of outcomes and adverse effects is so poor, a serious reexamination of EMS practice is indicated.
Mots-clés Pascal : Médecine, Urgence, Service santé, Période préhospitalière, Epidémiologie, Evaluation, Méthodologie, Biais méthodologique, Homme, Etats Unis, Amérique du Nord, Amérique, Soin intensif
Mots-clés Pascal anglais : Medicine, Emergency, Health service, Prehospital period, Epidemiology, Evaluation, Methodology, Methodological bias, Human, United States, North America, America, Intensive care
Notice produite par :
Inist-CNRS - Institut de l'Information Scientifique et Technique
Cote : 98-0052259
Code Inist : 002B30A01A1. Création : 14/05/1998.