| Study |
Bias |
| Randomisation process |
Deviations from intended interventions |
Missing outcome data |
Measurement of the outcome |
Selection of the reported results |
Overall |
| Authors' judgement |
Support for judgement |
Authors' judgement |
Support for judgement |
Authors' judgement |
Support for judgement |
Authors' judgement |
Support for judgement |
Authors' judgement |
Support for judgement |
Authors' judgement |
Support for judgement |
| Berger 2022 |
Low risk of bias |
Participants were appropriate randomized via computer generated random numbering in a 1:1 ratio to receive standard treatment and the allocazion sequence was concealed. There are no baseline differences that would suggest a problem with randomization. |
Low risk of bias |
An appropriate analysis to assess an effect of assignment were conducted on all randomised participants, therefore any deviations from interventions are unlikely to impact the results. |
Low risk of bias |
Data for this outcome was available for all participants ramodmized. |
Low risk of bias |
No differences in the measurement or ascertainment of the outcome between intervention groups were noticed or indicated. |
Low risk of bias |
Data was analysed as specified in a previously published protocol. The outcome was pre‐specified by the review authors. |
Low risk of bias |
For this outcome, there is a low risk of bias for all the domains. |
| Horby 2021 (RECOVERY) |
Low risk of bias |
A web‐based randomisation with allocation concealment was used and the baseline characteristics were balanced between groups. |
High risk of bias |
In the intervention group, 10% never obtained the treatment and 31% did not receive it consistently. There were even 210 (3%) participants in the control group receiving the treatment. |
Low risk of bias |
Data for this outcome were available for 99% of the included participants. ITT analysis conducted. |
Low risk of bias |
The measurement of the outcome was appropriate. We suppose, that the assessment of an objective outcome as death could not be influenced by the knowledge of the intervention. |
Low risk of bias |
The SAP was finalized before unblinded outcome data were available, there were no multiple elegible outcome measurements/data for this outcome. |
High risk of bias |
Overall, we rate the high risk of bias to be high due to deviations from the intended intervention. |
| REMAP‐CAP 2022 |
Low risk of bias |
Due to platform study design, allocation concealment for randomisation between Cox‐inhibitor and P2Y12 inhibitor was not given, but as we analyzed both interventions together, the impact seems negligible. Any baseline differences are probably due to chance. |
Some concerns |
We rate some concerns due to missing information whether there were crossovers or whether the assigned medication was really taken. The rate of withdrewn consent was low among all strata. |
Low risk of bias |
The outcome data was available for nearly all participants randomised. |
Some concerns |
The outcome was measured making phone calls and due to the unblinded design, we rate some concerns in this domain. |
Low risk of bias |
The outcome was assessed and analyzed as planned. |
Some concerns |
Overall rated some concerns due to missing information about how many participants recieved their assigned intervention and due to the unblinded design. |
