n Journal of Minimum Intervention in Dentistry - SYSTEM Research note on: the association of lack of adequate allocation concealment with effect-size overestimation
|Article Title||SYSTEM Research note on: the association of lack of adequate allocation concealment with effect-size overestimation|
|© Publisher:||Midentistry CC|
|Journal||Journal of Minimum Intervention in Dentistry|
|Affiliations||1 University of the Witwatersrand|
|Publication Date||Jan 2013|
|Pages||1 - 14|
|Keyword(s)||Allocation concealment, Dental, Empirical evidence, Meta-epidemiological study, Randomisation and Selection bias|
Objective :To test the null hypothesis that clinical dental trials with adequate random sequence generation together with adequate allocation concealment do not differ in their effect size from trials lacking adequate concealment.
Study design and setting : The oral health section of the Cochrane database was searched online until 20 July 2011. All listed entries, except reviews in protocol stage and duplications, were included. Included systematic reviews were accepted for data extraction if they included bias assessment for random sequence generation and allocation concealment, as well as an analysis report of dichotomous data. From the accepted systematic reviews, datasets including number of observed events and total number of evaluated units for test- and control group; assessment results for selection bias concerning random sequence generation and allocation concealment were extracted. The extracted datasets were statistically analysed.
Results : A 28% effect size overestimation in association with the lack of adequate allocation concealment in selected dental randomised control trials was established and the null hypothesis was rejected.
Conclusion : Adequate random sequence allocation without adequate concealment of such allocation appears to be insufficient to prevent exaggeration of results due to selection bias in clinical dental trials.
Article metrics loading...