Experimental Designs and Randomization Schemes
Section 7
Concealment and Blinding – ARCHIVED
Whenever feasible, randomized trials should incorporate mechanisms for safeguarding the identity of the treatment assignment from investigators, study staff, and participants. In traditional individually randomized trials, the assignment is concealed from the investigators before randomization to protect against treatment selection bias. For example, in an individually randomized trial, a provider may be biased to believe that one of the “treatments” would not be good for elderly patients. If the provider is able to predict the next treatment assignment, the provider may hold the elderly patient back from being randomized or might only “randomize” such patients when the other treatment is likely.
After randomization, blinding (also called "masking") is used to guard against a placebo effect and/or biased outcome assessment. It is most important for the primary outcome to be measured and recorded objectively, without knowledge of the actual treatment assignment. Trials may be single-blind, in which the participant does not know whether they are receiving the experimental therapy or the comparator; double-blind, in which the physician and staff also do not have access to this information but statistical analysis personnel do; or triple-blind, in which statisticians are also unaware of which participants have been assigned to which treatments.
With CRTs, concealment is not usually an issue because all of the clusters are identified ahead of time and are randomized at the same time. Hence, unlike individually randomized trials in which participants enter the study over time, there is no opportunity to “predict” the assignment and alter behavior accordingly. However, it is important to obtain written assurances before randomization that each cluster will comply with the assigned strategy for the duration of the study.
Blinding, on the other hand, is usually either not possible or not practical in CRTs. Most interventions that call for cluster randomization need to be disclosed to those who are implementing them. However, it is important to maintain as much objectivity as possible in recording the outcome assessments. For example, in the previously mentioned TiME trial, objective measures such as hospitalizations and mortality should not be collected differentially for the 2 arms of the trial. When evaluating quality of life in such a trial, care needs to be taken so that the measures are not elicited by referring to the conditions of the trial, such as starting a questionnaire with "Did you find that your dialysis time influenced how you feel on a daily basis?" Likewise, for a handwashing campaign, the assessment of nosocomial infection should be conducted according to objective criteria applied in the same way in both arms.
SECTIONS
sections
- Introduction – ARCHIVED
- Statistical Design Considerations – ARCHIVED
- Cluster Randomized Trials – ARCHIVED
- Randomization Methods – ARCHIVED
- Choosing Between Cluster and Individual Randomization – ARCHIVED
- Alternative Cluster Randomized Designs – ARCHIVED
- Concealment and Blinding – ARCHIVED
- Designing to Avoid Identification Bias – ARCHIVED
- Additional Resources – ARCHIVED
current section : Concealment and Blinding – ARCHIVED
- Introduction – ARCHIVED
- Statistical Design Considerations – ARCHIVED
- Cluster Randomized Trials – ARCHIVED
- Randomization Methods – ARCHIVED
- Choosing Between Cluster and Individual Randomization – ARCHIVED
- Alternative Cluster Randomized Designs – ARCHIVED
- Concealment and Blinding – ARCHIVED
- Designing to Avoid Identification Bias – ARCHIVED
- Additional Resources – ARCHIVED