n Journal of Minimum Intervention in Dentistry - System research note on : assessing attrition bias risk
|Article Title||System research note on : assessing attrition bias risk|
|© Publisher:||Midentistry CC|
|Journal||Journal of Minimum Intervention in Dentistry|
|Affiliations||1 University of the Witwatersrand|
|Publication Date||Jan 2012|
|Pages||23 - 28|
|Keyword(s)||Attrition bias, Loss-to-follow up and Systematic review|
Context : Systematic reviews of clinical trials need to assess the risk of attrition bias as part of its appraisal of the currently available evidence to a particular review question.
Problem : Notwithstanding the possible merits of different approaches to estimate the potential intervention outcomes of lost trial participants as the main reason for attrition bias, most remain arbitrary.
Suggested solution : Assuming a worst- and best-case scenario of intervention outcomes provides the certainty that neither lower nor higher values beyond these scenarios, respectively, are possible. Thus, worst- and best-case scenarios provide extreme outcome values that have the same probability to correspond with the true intervention outcome as any other possible scenario in between these extremes. Worst- and best-case scenarios can be calculated for dichotomous and continuous data, if the number of lost trial participants per intervention group is known. The results may then be compared to the intervention outcomes computed for participants available to follow-up and on this basis conclusions concerning attrition bias risk been drawn: i.e. risk of attrition bias may be assumed if the computed outcomes between worst- and best-case scenario and the intervention outcomes computed for participants available to follow-up differ significantly. Care needs to be taken not to accept the results of either worst- or best case-scenario as evidence for clinical considerations. They only provide evidence, if they differ significantly, for reasonable doubt concerning the validity of trial results in light of potential attrition bias risk.
Article metrics loading...