The New alpha value for statistical significance in clinical trials

Authors

DOI:

https://doi.org/10.28998/rpss.e02207011

Abstract

Nowadays the null hypotheses is used to show statistical significance, but all the hypotheses of the statistical model must be used to disseminate information closer to the results that occur in the population. The objective of this article is to describe the bases for understanding the best reference value for statistical significance to be adopted in clinical research. A handsearch was conducted to select papers on Medline via Pubmed. Only papers in English were considered. After analyzing the selected bibliography, the findings were sumirized to present the results. The similarity of the statistical hypothesis tests with the diagnostic tests and the evaluation of α and β at the same time provide a formula for the value of α. This formula is: α = statistical power – (positive predicted value x statistical power)/positive predicted value. The new alpha value for the interpretation of statistical significance should be 0.042 considering the sample size calculation with a statistical power of 0.8 and a confidence level of 95%. The alpha value must be adjusted if there is variation in statistical power.

Downloads

Author Biography

Fabiano Timbó Barbosa, Centro Universitário Tiradentes

Médico formado pela Universidade Estadual de Ciências da Saúde de Alagoas em 1999 e especialista em Anestesiologia com pós-graduação a nível de doutorado em Ciências da Saúde

Published

2022-12-03

How to Cite

Barbosa, F. T. (2022). The New alpha value for statistical significance in clinical trials. Revista Portal: Saúde E Sociedade, 7(Fluxo contínuo), 9. https://doi.org/10.28998/rpss.e02207011