Table 3

The quality of evidence per GRADE methodology and is to be used in conjunction with table 2

StudyOutcomeRisk of biasInconsistencyIndirectnessImprecisionPublication biasQuality
Lee et al
22 Lab		PF vs NaCl particles of varying sizesNot seriousNot seriousSeriousNot seriousUndetected⊕
VERY LOW
Relationship between fit factor and PFNot seriousNot seriousSeriousNot seriousUndetected⊕
VERY LOW
Gawn et al
6 Lab		PFNot seriousNot seriousSeriousNot seriousUndetected⊕⊕
LOW
Fit factorNot seriousNot seriousNot seriousNot seriousUndetected⊕⊕
LOW
He et al
23 Lab		Total inward leakageNot seriousNot seriousSeriousNot seriousUndetected⊕⊕
LOW
Radonovich et al
24 RCT		Incidence of laboratory-confirmed influenza Not seriousNot seriousNot seriousSeriousUndetected⊕⊕
LOW
Incidence of viral respiratory illnessNot seriousNot seriousNot seriousSeriousUndetected⊕⊕
LOW
Loeb et al
25 Cohort		Incidence of laboratory-confirmed influenza SeriousNot seriousSeriousNot seriousUndetected⊕⊕
LOW
Incidence of non-influenza virusesSeriousNot seriousSeriousNot seriousUndetected⊕⊕
LOW
MacIntyre et al
26 RCT		Clinical respiratory illnessNot seriousNot seriousNot seriousNot seriousUndetected⊕⊕⊕
MODERATE
Incidence of laboratory-confirmed pathogensNot seriousNot seriousNot seriousSeriousUndetected⊕⊕
LOW
MacIntyre et al
27 RCT		Incidence of symptomatic laboratory-confirmed bacterial respiratory tract infectionsNot seriousNot seriousNot seriousSeriousUndetected⊕⊕⊕
MODERATE
Co-colonisation with bacterial/viral infectionNot seriousNot seriousNot seriousSeriousUndetected⊕⊕⊕
MODERATE
Ng et al
28 Case report		Laboratory-confirmed SARS-CoV-2SeriousNot seriousVery seriousSeriousUndetected⊕
VERY LOW
Loeb et al
29 Retrospective		Confirmed SARS infectionSeriousNot seriousSeriousSeriousUndetected⊕
VERY LOW

  • PF, protection factor; RCT, randomised controlled trial.