I think that where the intervention partly automates something that would otherwise be done by clinicians, a comparison with usual care may reveal a clinically-useful benefit, while using a heavily monitored & protocolised comparison group may obscure that benefit. I wouldn’t say it’s a criticism.