I don't think it's obvious that meta-review is better than nothing. It can lend credence to incorrect/fraudulent results, making actual replication appear less important.