Aula historia da epidemiologia

O cálculo amostral deve ser realizado para o planejamento do estudo, para isso, devemos escolher a variável primária (ou mais importante) antes do início da pesquisa. Caso o revisor ou o leitor opte por acreditar que tivemos um erro tipo II em nosso estudo, esse provavelmente será mais perceptível para a variável % de recob, o que não foi nossa variável primária. E seguindo o que já existe na literatura, optou-se pelo planejamento estatístico com a Recred, para esse parâmetro, acredito que realmente não haja diferença entre os grupos.

Although the primary outcome was established as recession reduction, as previous studies, at the beginning of this trial, the null result of root coverage percentage difference seems particularly relevant to the overall interpretation of the results. We therefore conducted a post hoc power analysis to find out whether the design for that parameter had enough power to detect an effect of negation. The effect size of this particular contrast was 0.69 (i.e., a large effect, according to Cohen’s, 1977, effect size conventions) and the power to detect a difference between the two groups was determined to be 0.58. Thus, we cannot completely rule out the presence of an effect of negation in the present conditions for root coverage percentage.

Uma resposta que poderia ser utilizada para o editor

“Power analysis (a priori or prospective power analysis) is usually performed before planning a study but, as you suggested, it can be done either after data collection (post hoc or retrospective power analysis). While the utility of prospective power analysis in experimental design is universally accepted, the usefulness of retrospective techniques is controversial[1]. The use of the statistical analysis of the already collected data to estimate the power, may result in uninformative and misleading values. In particular, it has been shown [2] that post-hoc power in its simplest form is a one-to-one function of the obtained p-value. This has been extended[2] to show that all post-hoc power analyses suffer from what is called the "power approach paradox" (PAP), which is presented in the paper cited. For all these reasons and for the complex calculation of several simple sizes related to every variable examined, we considered unwise to perform a post hoc analysis. Furthermore we are confident that a statistically significant p value still can offer a reliable and useful information even if the power analysis has not been performed.”