Bias | Author's judgement | Support for judgement |
Randomization |
Some concerns |
Quote: "An independent biostatistician from the Indian Council
Medical Research-National Institute of Epidemiology,
Chennai, India, generated the randomisation
sequence using the RALLOC module in STATA v.14
(College Station, TX). A stratified block randomisation
strategy was used to allocate participants in a 1:1 ratio
to receive either convalescent plasma with the best
standard of care (intervention arm) or best standard
of care alone (control arm). Stratification was by
sites; block randomisation was done with unequal
block sizes. After written, informed consent had been
obtained from eligible patients, the site investigators
screened the participants for recruitment and
contacted a member of the central trial coordinating
team to receive the randomisation sequence, ensuring
concealment of allocation" Comment: There were baseline imbalances in diabetes, however, logistic regression done for diabetes. |
Deviations from intervention |
Low |
Comment: Unblinded study.
3 participants in the control group received intervention (Convalescent Plasma). 7 participants of the intervention group did not receive allocated treatment. Information reported on administration of co-interventions of interest: antivirals, corticosteroids, biologics. Co-interventions were balanced across arms Data were analyzed using intention-to-treat analysis. |
Missing outcome data |
High |
Comment: 464 participants analyzed, 464 analyzed (ITT), 451 analyzed (per protocol)
Mortality was analyzed via ITT analysis. Viral negative conversion via per-protocol analysis. 1 participant was lost to follow-up in each group. Per-protocol analysis: 7 participants were excluded from intervention group, 3 excluded from control group. Reason: did not receive allocated intervention Data available for > 95% of all participants Risk assessed to be low for outcomes: Mortality Data available for <95% of all participants for outcome viral negative conversion No reason given for missingness No information on whether missingness could or is likely to depend on its true value. Risk assessed to be high for outcomes: Viral negative conversion |
Measurement of the outcome |
Low |
Comment: Unblinded study (outcome assessor)
Mortality and viral negative conversion are observer-reported outcomes, not involving judgement. Risk assessed to be low for the outcomes: Mortality. Viral negative conversion. |
Selection of the reported results |
Some concerns |
Comment: Neither the protocol or statistical analysis plan were available. Registry was available.
The outcome mortality was registered and reported in the publication with appropriate time points. Result was not selected from multiple outcome measurements nor multiple analyses of the data. Trial analyzed as pre specified. Risk assessed to be low for outcome: Mortality. Viral negative conversion was not specified in the registry, but was reported in the publication. Result was not selected from multiple outcome measurements nor multiple analyses of the data. Trial not analyzed as pre specified. Risk assessed to be some concerns: Viral negative conversion. |
Overall risk of bias |
High |