Evaluation of a Covariate-Constrained Randomization Procedure in Stepped Wedge Cluster Randomized Trials

Join us as this presenter discusses this poster live on Tuesday, August 11, 2020 | Track A at 4:35 PM Mountain

PRESENTER
ERIN LEISTER CHAUSSEE
PhD Candidate, Colorado School of Public Health
BACKGROUND
In stepped wedge (SW) designs, differing cluster-level characteristics or individual-level covariate distributions that differ by cluster can lead to imbalance by treatment arm and potential confounding of the treatment effect.
SETTING
SW cluster randomized trials
METHODS
Adapting a method used in cluster-randomized trials, we propose a covariate-constrained randomization method to be used in SW designs. First, we define a balance metric to be calculated for all possible randomizations of cluster order for a given SW design (denoted BSW). The resulting distribution of this balance metric across all possible randomizations is used to select a candidate set of randomizations with acceptable covariate balance, for example, the best tenth percentile (P10) of the BSW distribution. One cluster order is selected at random from this candidate set to be used as the cluster order for treatment implementation. In a simulation study, we implement the covariate-constrained randomization procedure and computed treatment effect estimates and average absolute bias, and estimates of type I error and power. We used these outcomes to evaluate differing cutoffs of the BSW distribution used to define the candidate set and various analysis methods, under varying SW design and confounding settings.
RESULTS
We observed optimal statistical properties when the balance metric was used to exclude a small set of potential randomizations with the highest level of imbalance, and when analysis methods were adjusted for the potential confounders (see Figure for average absolute bias by BSW cutoff results). The covariate-constrained randomization was most beneficial in settings with a small number of clusters, low intra-class correlation, a low number of participants per cluster, and in the presence of cluster-level confounding variables.
CONCLUSIONS
We recommend researchers using the SW design carefully consider potential confounders, both cluster-level and individual-level, prior to cluster order randomization and consider a covariate-constrained randomization if appropriate. Treatment effect estimation should be adjusted for these potential confounders, and other covariates associated with the outcome of interest.
POSTER

Loader Loading...
EAD Logo Taking too long?

Reload Reload document
| Open Open in new tab

VIDEO


Posted in Poster Session, Study Design & Analysis.

Leave a Reply