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 |
Moderate |
Estimate was adjusted for local virus circulation, location, and ethnicity and weighted for probability of being vaccinated using: age, sex, chronic conditions (none vs 1+), use of PPE, contact with infected patients, local virus circulation, marital status, occupation, household characteristics, health status and behaviours (smoking, history of influenza vaccination), so covered all the pre-specified domains, including time varying risk of exposure to infection. |
| Selection of participants into the study |
Low |
No particular concerns in this domain. |
| Clasification of interventions |
Low |
No particular concerns in this domain - vaccination status is likely to be adequately measured. |
| Deviations from intervention |
Low |
No concerns in this domain - this study was observational. |
| Missing outcome data |
Moderate |
Out of the 17,891 screened for eligibility, 4,985 were eligible but refused consent and a further 1,046 withdrew or were lost to follow-up. Participants in the latter group were quite different from those included in the analysis; no information is known about the former group. There is the potential for the outcome to be different among those who refused or were lost to follow up compared to those included in the analysis, even after taking account of the measured covariates. |
| Measurement of the outcome |
Low |
No concerns in the domain - participants were tested weekly, regardless of whether or not they had symptoms. |
| Selection of the reported results |
Low |
No concerns in this domain - there was a published protocol. |
| Overall risk of bias |
Moderate |
|
| Overall comment | Some concerns over the possibility of uncontrolled confounding given the observational nature of the data and the fact that some confounders were self-reported, and uncertainty in whether those who were eligible but either refused to participate or withdrew/were lost to follow up might have differed with respect to the outcome compared to those included in the analysis. |