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 29 | MultiplicityHow does multiplicity occur in the setting of clinical trials?Multiple endpointsIn contemporary practice, trials compare two treatments in terms of severaloutcomes. This might be because we are keen to appreciate the impact of a newtreatment on several outcomes (such as mortality, disease incidence, and qualityof life) or because for some diseases there might not be a single dominant measureof outcome. This is common in trials addressing chronic disease conditions such asarthritis or neuropathy in which pain or joint mobility are the outcomes of interest.An example of this problem can be found in a recently published trial in childrenwith osteogenesis imperfecta [2]. This disease is characterized by fragile bones,which break easily. The trial evaluated olpadronate (a bisphosphonate) withrespect to its skeletal effects in children. There were three primary outcomes:• incidence of fractures of long bones• changes in bone mineral density and content• functionalityIt is clinically sensible to address this series of outcomes together, as no singleendpoint is sufficient to address the primary objective of the study. However,caution is needed in the interpretation of the results since many separatestatistical tests may be needed.Multiple endpoints also arise in studies of diseases that have major morbidityoutcome measures, such as the occurrence of a stroke, heart attack, or death;outcomes commonly used in cardiovascular trials.Multiple treatmentsThe decision to evaluate multiple treatments, combinations of treatments, ordifferent doses of the same treatment within a trial with several parallel armsalmost always implies some form of multiple comparison study. An example mightbe the simultaneous investigation of two treatments, both individually and incombination, on the disease-free survival of postmenopausal women with earlybreast cancer [3]. The number of possible treatment combinations increases rapidlywith the number of treatments, and so do the number of statistical tests performed.Repeated measurementsIn many clinical trials, an outcome or marker measurement is recorded atdifferent time points for each subject, generating repeated measurement data.To analyze such data, a time-by-time analysis is sometimes used: group meansare calculated at each separate time point, and separate statistical analyses are332

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

Saved successfully!

Ooh no, something went wrong!