Design and Sample-Size Issues for Cluster-Randomized Trials: