Group sequential design is a type of clinical trial design in which data are analyzed at multiple points during the course of the trial, rather than waiting until the end of the trial to analyze all the data.
An unblinded sample size re-estimation is done by increasing the sample size in response to an interim effect size that is ‘promising’. The uncertainty that you’re taking from the initial sample size determination will be around the effect size.
Video Tutorial
Two sample mean blinded sample size re-estimation is a statistical method used when the outcome of interest is continuous (i.e., a numerical measurement) and the trial is designed to compare the means of two treatments.
Two sample binomial blinded sample size re-estimation is a statistical method used in clinical trials to adjust the sample size of a study based on the results of an interim analysis.
Binary Data is data where subjects are categorized into one of two categories (e.g. disease/no disease) as the outcome of interest. This data is extremely common across various clinical areas with common statistical methods for binary data including chi-squared tests and logistic regression.
Ordered categorical (also known as ordinal) data is where subjects are categorized into one of the available categories and in which each category is ordered from best to worst. Scales for areas such as pain or disability are a common example in clinical trials with common statistical methods including the proportional odds model and the Wilcoxon test.
Contingency table data (also known as cross-tab) shows the frequency between and within two or more categorical outcomes and would be common in areas such as epidemiology. Tests for independence such as the chi-squared and Fisher’s exact tests are commonly used for this type of data.
Bayesian assurance refers to the degree of confidence or belief that one has in the correctness of a statistical model or hypothesis based on available data.
Bayesian assurance refers to the degree of confidence or belief that one has in the correctness of a statistical model or hypothesis based on available data.
Mixed Bayesian likelihood combines information from different sources of data with different levels of uncertainty.
Consensus-based criteria are criteria or standards that are developed through agreement or consensus among a group of experts or stakeholders.
Linear regression is used to evaluate the relationship between two continuous variables. The goal of linear regression is often to model the relationship between an independent variable (such as a treatment or a risk factor) and a dependent variable (such as a patient outcome or a biomarker.
Logistic regression is used to evaluate the effectiveness of a treatment or intervention by analyzing the probability of a patient experiencing a particular outcome based on their demographic, clinical, or other relevant characteristics.
Cox regression (proportional hazards model) is used to compare the survival or time-to-event outcomes between two or more treatment groups. The Cox regression model estimates the hazard ratio, which is a measure of the relative risk of an event occurring in one treatment group compared to another.
Hierarchical models are an example of regression models used in the analysis of Cluster randomized trials (CRTs). CRTs can be more efficient and cost-effective than individually randomized trials, leading to smaller sample sizes and reduced variance in outcomes.
A practical tutorial on how software can be used to generate randomization lists in a quick and flexible manner.
(Block Randomization, Complete Randomization, Efron’s biased coin randomization).
This type of design allows for more efficient and flexible study conduct, as it enables the trial to be modified in response to emerging results or changing circumstances. The primary goal of adaptive trial design is to improve the efficiency and accuracy of clinical trials, by enabling researchers to modify the trial design in real-time. This approach can help to minimize the number of participants needed, reduce the overall cost of the study, and speed up the time to reach the study endpoint.
A clinical prediction model is used to estimate the probability of a particular health outcome or disease in an individual patient. It combines various patient characteristics (such as age, gender, medical history, and test results), to provide an individualized estimate of the likelihood of a particular outcome.
Sample size for a normal endpoint helps to determine the minimum sample size required to detect a statistically significant difference between two groups in a clinical trial, where the endpoint of interest is a continuous variable that follows a normal (Gaussian) distribution.
Binomial proportion parallel is used in clinical trials to compare the proportions of success rates between two independent groups.
Survival parallel arm is used in clinical trials to compare the time to an event (such as death, disease progression, or recurrence) between two or more independent treatment groups.
Count parallel arm is used in clinical trials to compare the count or frequency of a particular event or outcome between two or more independent treatment groups.
