[Stragegies for evaluation of non-randomized therapeutic comparisons exemplified by prehospital volume therapy after trauma].
Langenbecks Archiv fur Chirurgie. Supplement. Kongressband. Deutsche Gesellschaft fur Chirurgie. Kongress · 1998-01-01 · Observational study
Abstract
Non-randomized comparisons of therapeutic approaches require extensive description of the patients. Classical subgroup analysis suffers from small sample sizes. Estimation of outcome based on known prognostic factors (e.g. scores) allow for an indirect comparison of treatment effects. Matched pair analysis generates two homogenous subgroups of patients. The propensity score approach classifies all patients according to their chance of receiving a certain therapy or not, thus giving comparable situations for stratified analysis. The preclinical fluid resuscitation for trauma patients can hardly be assessed by randomized trials. Application of the above mentioned methods shows no overall effect but polytraumatized patients seem to have a worse outcome with high volume resuscitation.