Quasi-experimental designs as alternatives to RCT for providing high quality evidence