Statistically validated evaluation of clinical trials
Lugosi, L.; Molnár, I.
Orvosi Hetilap 140(41): 2303-2307
1999
ISSN/ISBN: 0030-6002 PMID: 10603745 Document Number: 501428
Data of clinical trials of medicinal products must be evaluated in statistically valid models. The statistical validity criteria are defined. Statistically invalid models will result in biased parameter and confidence interval estimations, erroneous statistical inferences and clinical interpretations. Finally, wrong decisions will call forth deleterious consequences in the judgement of the therapeutic effect and the frequency and severity of the adverse reactions of the tested new medicinal, and generic products. Statistically validated analyses will promote the international harmonization of the scientific evaluation of medicinal products according to the idea of the evidence-based-medicine. The study presents examples of clinical trials evaluated with a software checking statistical validity assumptions while performing evaluation of data.