Examkin

Question

Difficulty: MediumInferential Statistical Tests

A gastroenterology research group is conducting a randomized controlled trial to evaluate a new monoclonal antibody against standard therapy for moderate-to-severe Crohn's disease. Four hundred patients are randomized equally into two independent treatment arms. The primary endpoint compares the proportion of patients achieving clinical remission at week 12, categorized as either 'achieved' or 'not achieved', between the two groups. A secondary endpoint compares the absolute reduction in the Crohn's Disease Activity Index (CDAI), a continuous numerical score, between the two treatment groups. Preliminary data analysis reveals that the CDAI reduction scores in both treatment arms are highly skewed and significantly violate the assumption of normality.

Which of the following statistical tests are appropriate for analyzing the endpoints in this study? (Select all that apply)

  1. Pearson's chi-square test to analyze the difference in clinical remission rates between the two groupsAnswer
  2. Mann-Whitney U test to analyze the difference in absolute CDAI reduction between the two groupsAnswer
  3. C
    Unpaired Student t-test to analyze the difference in absolute CDAI reduction between the two groups
  4. D
    McNemar test to analyze the difference in clinical remission rates between the two groups
  5. E
    Wilcoxon signed-rank test to analyze the difference in absolute CDAI reduction between the two groups

Answer

The appropriate statistical tests are Pearson's chi-square test for the primary endpoint and the Mann-Whitney U test for the secondary endpoint.
The study design involves two independent groups: patients receiving the novel monoclonal antibody and those receiving standard therapy. For the primary endpoint, the researchers are comparing the proportions of a categorical outcome (remission achieved versus not achieved). Pearson's chi-square test is the appropriate test for comparing proportions between two or more independent categorical variables when the sample sizes and expected frequencies are sufficiently large. For the secondary endpoint, the researchers are comparing a continuous variable (the absolute reduction in CDAI score). Usually, an unpaired Student t-test would be used for this; however, because the data is highly skewed and significantly violates the assumption of normality, a non-parametric equivalent must be chosen. The Mann-Whitney U test (also called the Wilcoxon rank-sum test) is the correct non-parametric test for comparing continuous or ordinal variables between two independent groups.

Step-by-Step Solution

1
Identify the data type and study design for the primary endpoint.
The primary endpoint evaluates categorical data (clinical remission achieved vs. not achieved) across two independent groups.
Determining variable type and group independence is the first step in selecting the correct statistical test.
2
Select the appropriate statistical test for the primary endpoint.
Pearson's chi-square test should be used.
Chi-square is the standard test for comparing categorical variables between two or more independent samples with large expected counts.
3
Identify the data type, distribution, and study design for the secondary endpoint.
The secondary endpoint evaluates a continuous numerical score (CDAI reduction) between two independent groups, but the data is highly skewed (non-normal distribution).
Evaluating for normality is crucial before applying tests for continuous variables.
4
Select the appropriate statistical test for the secondary endpoint.
The Mann-Whitney U test should be used.
Because the continuous data violates the assumption of normality, a non-parametric test is required. The Mann-Whitney U test (Wilcoxon rank-sum test) is the correct non-parametric alternative to the unpaired t-test for two independent groups.

Key Concept

Selection of statistical tests based on data type (categorical vs. continuous), study design (independent vs. paired groups), and distribution (normal vs. skewed).
Rate this question

Topics

Inferential Statistical Tests

More questions from this topic