Detecting Anomalies Among Practice Sites Within Multicenter Trials: An Application of Transportability Methods to the TOPCAT and ACCORD BP Trials