Is this another observational study that can't distinguish between sicker people (those who end up staying on steroids) and less sick people (those who don't require steroids)???