ACRP Biostatistics and Data Analysis Flashcards

What is the difference between descriptive and inferential statistics in clinical research?

Answer: Descriptive statistics summarize data; inferential statistics draw conclusions about populations from sample data

Descriptive statistics (mean, median, frequency) characterize the study sample, while inferential statistics use sample data to make generalizations about the broader patient population.

What is an interim analysis in a clinical trial and what risk does it pose?

Answer: A pre-planned analysis conducted before study completion that risks inflating the Type I error rate if not adjusted

Interim analyses allow early stopping for efficacy or futility but require alpha-spending adjustments (e.g., O'Brien-Fleming) to maintain overall Type I error control.

What is a primary endpoint in a clinical trial?

Answer: The pre-specified main outcome measure used to determine the trial's success or failure

The primary endpoint is the single most important outcome that the trial is designed and powered to assess, forming the basis of regulatory decision-making.

What is the purpose of a sensitivity analysis in clinical trials?

Answer: To assess how robust the primary results are under different assumptions or analytical approaches

Sensitivity analyses repeat the primary analysis under alternative assumptions (e.g., different imputation methods) to evaluate the stability of the main conclusions.

What is missing data imputation in the context of clinical trial analysis?

Answer: Using statistical methods to estimate and replace missing values to reduce bias in the analysis

Imputation techniques such as multiple imputation or last observation carried forward (LOCF) address missing data to minimize bias and preserve sample size.

Which regulatory guidance document provides statistical principles for clinical trials and is widely referenced in the US?

Answer: ICH E9 — Statistical Principles for Clinical Trials

ICH E9 provides foundational guidance on statistical methodology, including endpoint selection, randomization, blinding, sample size, and analysis populations.