https://www.infirmiere-canadienne.com/blogs/ic-contenu/2014/11/01/terminologie-101-dissimulation-de-laffectation-des
Nov 01, 2014, By: Maher M. El-Masri, RN, PhD
Allocation concealment in randomized controlled trials:The process of shielding those involved in the trial from knowing the upcoming participant group assignment
Source: DiCenso, A., Guyatt, G., & Ciliska, D. (2005). Evidence-Based Nursing: A Guide to Clinical Practice. St Louis: Elsevier Mosby
Elimination of bias is the defining characteristic of randomized controlled trials (RCTs) because it allows researchers to determine causality between the intervention and the outcome. Randomization, which refers to the random assignment of participants to study groups in an RCT, is the most important procedure to eliminate bias. Proper randomization depends on the implementation of adequate allocation concealment.
In a well-conducted RCT, the sequence by which participants will be allocated to the study groups is decided before the first participant is recruited. Allocation concealment is the practice of keeping researchers and participants unaware of the sequence, with the goal of preventing the researchers from (unconsciously or consciously) influencing the group assignment of study participants.
To illustrate the concept of allocation concealment, consider an RCT in which the intervention and the control groups are coded as A and B. After the researchers have determined the sample size, a person who is at arm’s length from the study is asked to generate the random sequence of participant allocation (e.g., A, B, B, A, A, A, B, …). Often this process is completed electronically, using computer-based or Internet-based random sequence generation programs. This person will then prepare sequentially numbered opaque envelopes, each of which will contain a card that will be marked with the group assignment (A or B) according to the randomly generated sequence. When the recruitment process starts and a consenting participant is deemed eligible to participate in the RCT, the next sealed envelope in the sequence is opened and the participant is assigned to the study group indicated on the card.
Subversion of allocation concealment presents a major challenge to the validity of the findings of an RCT. RCTs that do not provide evidence of allocation concealment have been reported to be associated with an inflated intervention effect (i.e., an overestimation of the effect of the intervention). In fact, one should also be concerned when researchers report allocation concealment but fail to provide details about the processes associated with the creation and administration of the concealed allocation sequence. The report should clearly indicate who created the sequence, who assigned the participants to study groups, and when and how these steps were accomplished.
In conclusion, the elimination of bias is the key strength of RCTs. Implementing allocation concealment in RCTs ensures that participants are allocated to study groups in an unpredictable, random way. Without allocation concealment, selection bias is a possibility; certain participants could be either excluded from or assigned to particular study groups on the basis of their prognosis or general condition. It is important that researchers clearly describe the steps they have taken to conceal allocation so that readers can judge whether the study has been subject to bias. If bias is present, the validity of the results has to be questioned.
NursesONE.ca resources on this topic
EBSCOhost
- Clark, L., Schmidt, U., Tharmanathan, P., Adamson, J., Hewitt, C., & Torgerson, D. (2013). Allocation concealment: A methodological review. Journal of Evaluation in Clinical Practice, 19(4), 708-712. doi:10.1111/jep.12032
- Clark, L., Schmidt, U., Tharmanathan, P., Adamson, J., Hewitt, C., & Torgerson, D. (2013). Poor reporting quality of key Randomization and Allocation Concealment details is still prevalent among published RCTs in 2011: A review. Journal of Evaluation in Clinical Practice, 19(4), 703-707. doi:10.1111/jep.12031
- Odgaard-Jensen, J., Vist, G. E., Timmer, A., Kunz, R., Akl, E. A., Schünemann, H.,…Oxman, A. D. (2011). Randomisation to protect against selection bias in healthcare trials. Cochrane Database of Systematic Reviews, 2011(4). doi:10.1002/14651858.MR000012.pub3
Maher M. El-Masri, RN, PhD, is a full professor and research chair in the faculty of nursing, University of Windsor, in Windsor, Ont.
#research#nursing-research