China
Africa
Explore chapters and articles related to this topic
Statistical Approaches in the Development of Digital Therapeutics
Published in Oleksandr Sverdlov, Joris van Dam, Digital Therapeutics, 2023
Oleksandr Sverdlov, Yevgen Ryeznik, Sergei Leonov, Valerii Fedorov
Statistical analysis issues of CER studies merit careful considerations as well. Due to the real-world nature of these studies, post-randomization events such as non-adherence, treatment switching, dropouts, etc., are likely to occur and need to be handled appropriately. The estimand framework, initially proposed in 2014 and adopted in 2019 as an ICH (International Conference on Harmonisation) E9 (R1) Addendum38 provides a systematic approach to define estimands (target objects to be estimated) in clinical trials (Mallinckrodt et al., 2020). It has been increasingly utilized in biopharmaceutical research, and it should also be applicable in the DTx research settings such as mobile health intervention trials. The importance of rigorously defining estimands is magnified in decentralized clinical trials (De Brouwer et al., 2021), where clinical outcome data are acquired either entirely remotely or in a combined manner—both during in-clinic visits and remotely.
Selected Statistical Topics of Regulatory Importance
Published in Demissie Alemayehu, Birol Emir, Michael Gaffney, Interface between Regulation and Statistics in Drug Development, 2020
Demissie Alemayehu, Birol Emir, Michael Gaffney
In a broad sense, an estimand is the quantity that is the target of inference in order to address the scientific question of interest posed by the trial objective (ICH E9 (R1) 2017). As such, it may be characterized by various attributes, including the population of interest, the variable (or endpoint), the handling of intercurrent or post-randomization events, and the summary statistics associated with the outcome variable.
Missing Data
Published in Shein-Chung Chow, Innovative Statistics in Regulatory Science, 2019
An estimand is a parameter that is to be estimated in a statistical analysis. The term is used to more clearly distinguish the target of inference from the function to obtain this parameter, i.e., the estimator and the specific value obtained from a given data set, i.e., the estimate (Mosteller and Tukey, 1987). To distinguish the terms of estimator and estimand, consider the following example. Let X be a normally distributed random variable with mean μ and variance The variance is often estimated by sample variance , which is an estimator of and is called the estimand. An estimand reflects what is to be estimated to address the scientific question of interest posed by a trial. In practice, the choice of an estimand involves population of interest, endpoint of interest, and measure of intervention effect. Measure of intervention effect may take into account the impact of post-randomization events such as dropouts, non-compliance, discontinuation of study, discontinuation of intervention, treatment switching, rescue medication, death and so on.
The estimand framework and its application in substance use disorder clinical trials: a case study
Published in The American Journal of Drug and Alcohol Abuse, 2021
Jessica K. Roydhouse, Lysbeth Floden, Rachel L. Tomko, Kevin M. Gray, Melanie L. Bell
Per the ICH, an estimand has five attributes: treatment condition, population, variable, intercurrent events (ICEs), and the summary measure of the variable (15). ICEs are post-randomization events that can affect the measurement or interpretation of the treatment effect (15). Although ICEs are highly relevant for understanding treatment effects, they may have only been considered in statistical analysis plans rather than forming part of the scientific question (21). Different ICEs may arise in different trial and treatment contexts, even within SUD. For example, non-adherence was high in the MATCH trial (6). In the COMPASS trial, 12% of participants still smoking after treatment discontinued varenicline as they perceived it was not working (22). Other potential ICEs may include co-use of other recreational substances or concomitant medications that can affect treatment response.
Custom Epoch Estimation for Surveys
Published in Journal of Applied Statistics, 2019
Tucker McElroy, Osbert Pang, George Sheldon
In describing two sources of uncertainty – the superpopulation and the sample – we follow the approach delineated in [1]. We utilize lower case letters for realizations of random variables on a particular ω in the probability space. For example, the sampling mechanism (which governs the selection of sampling units from the population) is denoted S, with X denotes the epoch estimand – a random variable with realization f, and
Pragmatic clinical trials in the context of regulation of medicines
Published in Upsala Journal of Medical Sciences, 2019
Rolf Gedeborg, Charles Cline, Björn Zethelius, Tomas Salmonson
The completeness of data collection is a key issue for any type of trial and in focus for GCP (6). An example is when data on outcomes routinely collected in the electronic patient records are not accessible during the study and are found to be frequently missing during analyses (15). A study design and study conduct that prevents missing information is always preferable to statistical handling of missing information at the analysis stage (16). Such methods often have unverifiable assumptions. Sensitivity analyses to evaluate the impact of the handling of missing data and associated assumptions are essential for the interpretation of study results. The importance of choosing an appropriate estimand during the planning stage of a study, so that attempts to prevent missing data can be tailored to that choice, and appropriate estimation methods can be specified, must be stressed (17). It is essential to detect and adequately handle intercurrent events, such as use of alternative treatment, discontinuation of treatment, and treatment switches, in the analyses. Otherwise, these types of events may lead to invalid conclusions regarding treatment effects (17). These two fundamental considerations in the study design, how to prevent missing information and defining the appropriate estimand, may be a greater concern for pragmatic study designs.