article
05.06.2019

Interim Analysis as a Key Decision-Making Tool

Good Clinical Practice dictates stricter compliance with the fundamental concepts of conducting interim analysis by Russian pharmaсeutical companies, mainly because of the additional organizational resources it requires. On the other hand, the methodology and principles of deciding on the necessity of carrying out an interim analysis are not always properly reflected in a research protocol.

The topic of Interim analysis and its role within clinical trials have always been raising plenty of discussion among the industry stakeholders. The subject was thoroughly covered during a dedicated panel session at the 6th Annual OCT Conference on Clinical Trials. 

“Russian regulatory entities have a rather straightforward standpoint regarding interim analysis. Its results are deemed legitimate both from a regulatory and a scientific point of views provided a necessity of such an analysis is announced and substantiated in a research protocol”. - Dmitry Goryachev, head of the FSBI Scientific Centre for Expert Evaluation of Medicinal Products of the Ministry of Health of the Russian Federation commented during the 6th annual OCT Conference on Clinical Trials.
"Interim analysis should not be viewed as a panacea guarantying the success of a study. Obtaining statistically significant results during the final analysis even with the noticeable results of the interim analysis should be viewed only in terms of probability. Therefore, there should be a certain level of confidence, that a trial will benefit from using such an adaptation, which may create certain organization, logistic and statistical difficulties.
The decision on carrying out an interim analysis, as well as defining its strategy, timeframe and adaptive design elements is made during the trial planning stage. Different options are thoroughly examined in order to find the most optimal one for each individual trial. 
Traditional non-adaptive approach with a fixed sample size assumes that during a trial planning stage, a calculation of a sample size is performed based on the expectation effect, considering the minimal clinically important difference (MCID). However, there are cases, when there is no certainty in the effect estimation, while when using a conservative sample size estimation, the effect becomes too large from both ethical and financial points of view. There is an alternative approach, however, adaptive design. It is a sequential trial design or a recalculation of the number of patients based on the results of an interim analysis. Classic group sequential design assumes a rather significant planned number of patients allowing for high accuracy when calculating a relatively small but clinically important difference. This approach allows to stop a trial during early stage effectiveness demonstration or in case of futility during a regular planned interim analysis, which in turns lowers the final sample size. The second approach is to start a trial with a relatively small sample size, planned based on the realistic expectations towards the effect (a little bigger than an MCID) and when necessary increase the sample size based on the interim analysis results." - Data MATRIX Lead Biostatistician, PhD in Biology, Irina Bondareva noted. 


"We place value of quality of  the necessary procedures during trials above deadlines and budgets, since any inaccuracies occurred during interim analysis, for instance, might jeopardize the overall study results. High level expertise from all the contributing parties along with  knowledge of international and local directives, understanding of  regulator’s outlook and compliance with data confidentiality requirements are the primary elements to take note of” - OCT clinical operations director Irina Petrova added.