Use the following figure to answer questions 12-14: Figure 3b (British J Derm 2008) comes from a randomized trial of DHA versus a placebo pill for treating eczema. The figure shows boxplots of the eczema severity score, SCORAD, at baseline and week 8 for each group. SCORAD score is not normally distributed and the sample size is small.
The image shows 4 box plots. Box Plot A (DHA, Week 0) Median 32, lower quartile 29 and upper quartile 44 (minimum 18 and maximum 49); Box Plot B (DHA, Week 8) Median 28, lower quartile 23 and upper quartile 34 (minimum 18 and maximum 52); Box Plot C (Control, Week 0) Median 36, lower quartile 30 and upper quartile 45 (minimum 18 and maximum 65); Box Plot D (Control, Week 8) Median 34, lower quartile 25 and upper quartile 38 (minimum 13 and maximum 55)
Reproduced with permission from: Koch C, Dölle S, Metzger M, et al. Docosahexaenoic acid (DHA) supplementation in atopic eczema: a randomized, double-blind, controlled trial. Br J Dermatol 2008;158:786-792.
QUESTION 12
12. What test should be used to determine whether DHA is a better treatment than control for treating eczema (SCORAD is not normally distributed)?
A t-test that compares A and B. A Wilcoxon sign-rank test that compares B and D. Wilcoxon sign-rank tests comparing A with B and C with D. A Wilcoxon sum-rank test that compares the difference between A and B with the difference between C and D. A chi-square test that compares B and C
A Wilcoxon sum-rank test that compares the difference between A and B with the difference between C and D.
To determine whether DHA is a better treatment than control for treating eczema, we can use nonparametric statistical tests since the SCORAD score is not normally distributed. One commonly used nonparametric test is the Wilcoxon sign-rank test.
Options A, B, and D involve using the Wilcoxon sign-rank test. The Wilcoxon sign-rank test compares the paired observations between two groups. In this case, we can compare the SCORAD scores before and after treatment within the DHA group by comparing Box Plot A (DHA, Week 0) with Box Plot B (DHA, Week 8). Therefore, comparing Box Plot B with Box Plot D would not be appropriate for determining whether DHA is a better treatment than control.
Option C suggests using two separate Wilcoxon sign-rank tests to compare Box Plot A with Box Plot B and Box Plot C with Box Plot D. This approach is reasonable since it allows for independent comparisons between the treatment and control groups.
Option E suggests using a chi-square test to compare Box Plot B (DHA, Week 8) with Box Plot C (Control, Week 0). However, a chi-square test is typically used for categorical data, not continuous variables like the SCORAD score. Therefore, this option would not be appropriate for this scenario.
Therefore, the correct answer is option C: Wilcoxon sign-rank tests comparing A with B and C with D.