INATE - INvestigators Against ThromboEmbolism
Study design
Clinical and epidemiological studies involve the collection of what is sometimes huge amounts of data. In order to be able to draw conclusions from these data, the principles of proper study design should be adhered to.


Introduction

If applied in a correct manner, the principles of study design will guarantee that the results of the study are valid, unbiased and generalizable to other patients. The principles are:

  • that randomization is the best guarantee that groups will be comparable
  • that blinding will ensure comparability of information obtained
  • that the incidence is a more appropriate measure for the occurrence of a disease than the prevalence (In most prospective clinical trials related to VTE the outcome is reported by the incidence of, for example, newly developed DVT, bleeding or death. Most cohort studies also use incidence to report outcomes but, conversely, cross-sectional studies generally use prevalence.)
  • that a crossover design may be applied when studying a chronic disease but not when studying an acute disease
  • that within a clinical study, the internal validity should be high.

Anyone who wants to practise EBM must be able to critically appraise the results of clinical studies. This is not possible without knowledge of the principles of study design. Therefore, in this chapter, the first principles of study design are discussed and illustrated with examples.

We conclude this introduction with a brief overview of the most important designs for randomized and epidemiological studies.

The two most frequently employed designs for randomized clinical trials are:

  • the parallel-group study
    in which patients are randomized to only one of the treatments that are being compared, i.e. to different treatment groups which are observed in parallel
  • the crossover study
    in which patients are sequentially exposed to two or more treatments, to investigate the differences between those treatments.

The most frequently reported types of observational (nonrandomized) studies are:

  • the case-control study
    in which the frequency of exposure is compared between a group of people with a disease (the cases) and a group of people free of the disease (the controls)
  • the longitudinal or cohort study
    in which a group of people is observed over time
  • the cross-sectional study or survey
    in which all the information on a group of participants is collected at the same time, without any follow-up.


 
Home Print this page Send to a friend Add to favorites
  Search help 
      
 
Sign in to INATE
User name:

Password:




Is this your first time here? Register online
I forgot my password

Select a term below
Reference materials