Subject-level matching for imbalance in cluster randomized trials with a small number of clusters.

Subject-level matching for imbalance in cluster randomized trials with a small number of clusters.