11.07.2015 Views

Clinical Trials

Clinical Trials

Clinical Trials

SHOW MORE
SHOW LESS
  • No tags were found...

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

❘❙❚■ Chapter 22 | Intention-to-Treat AnalysisFigure 2. Trial profile for the vitamin A supplementation trial described in Example 2 [5].225 villages randomized12,094 childrenassigned to vitamin A11,588 childrenassigned to placebo9,675 childrencompliant2,419 childrennoncompliant12 children died 34 children died74 children diedand subsequently withdraw from a study, so excluding such patients will result ina greater treatment effect being shown than that likely to be seen in the generalpopulation. In both ITT and PP analyses, a judgment call may need to be made abouthow to deal with subjects for whom data are incomplete, and by doing this there isa risk of biasing the effectiveness of a drug. Therefore, when it is inappropriate to useeither method alone, statistical methods beyond both ITT and PP can add essentialinsight. This point is illustrated in the following example.Example 2Subjects with vitamin A deficiency are more susceptible to infections due toimpaired immunity. Therefore, a community-based randomized trial wasconducted in rural Indonesia, where infant mortality due to infections is high.The aim of the study was to estimate whether two high doses of vitamin A giventwice over an 8-month period would reduce mortality among preschool children.In total, 23,682 children were randomized to vitamin A or placebo. During thestudy, 20% of children in the treatment group did not receive vitamin A becausedistribution to their villages was a significant problem. Figure 2 plots the trialprofile and Table 2 summarizes the results of the trial and analyses.On completion of the study, there were 46 deaths out of the 12,094 children(0.38%) who were randomized to vitamin A supplementation, and 74 out of11,588 (0.64%) in the no-supplementation group. Using an ITT analysis, the oddsratio for death was estimated to be 0.59 (95% CI 0.40, 0.87; P = 0.005), suggestinga 41% lower rate of death in the group who were supposed to receive vitamin A.The PP analysis calculated an odds ratio of 0.19 (95% CI 0.10, 0.36; P = 0.001),260

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!