Note: The risk of bias by domain corresponds to the highest risk of bias among outcomes by domain.
The overall risk of bias corresponds to the overall highest risk of bias assessed among outcomes.
| Bias | Author's judgement | Support for judgement |
| Confounding |
Serious |
Matched on age (5-year band), sex, date (within 3 days), municipality, race, and previous symptomatic covid; adjusted for comorbidities and age (continuous). No SES, but municipality and race thought to be reasonable proxies in this setting. Analysis of data 0-13 days after vaccination did suggest some benefit in this period for severe disease (hospitalisation) and mortality, particularly the former, suggesting some unmeasured/residual confounding |
| Selection of participants into the study |
Serious |
Test negative design and selection by non-missing race, which could potentially be associated with outcome and differential with respect to vaccination status. No information presented on whether this was the case. For severe disease (hospitalisation and mortality) authors planned a time to event analysis among test-positive individuals but did not present results from this. |
| Clasification of interventions |
Low |
No concerns in this domain |
| Deviations from intervention |
Low |
No concerns in this domain |
| Missing outcome data |
Low |
No concerns in this domain |
| Measurement of the outcome |
Low |
No concerns in this domain |
| Selection of the reported results |
Low |
Protocol was published and available |
| Overall risk of bias |
Serious |
|
| Overall comment | Main concern was over possible selection bias due to the fact that those with non-missing race were excluded (and this could possibly be influenced by both vaccination status and case/control status); for hospitalisation and mortality a planned sensitivity analysis to address potential selection bias was not presented. Also some concerns over residual confounding for hospitalisation and mortality. |