I think it's going to depend very much on the distribution of patients with the different WHO Covid Ordinal Scale categories within the cohorts. If there were

I think it's going to depend very much on the distribution of patients with the different WHO Covid Ordinal Scale categories within the cohorts. If there were a lot more 5s (severely ill hospitalised patient on non-invasive ventilation or high-flow oxygen) vs. 3s (hospitalised patients not needing oxygen therapy), it would presumably make a large difference.