Summary of findings 3. Quarantine.

Disease: COVID‐19 Interventions: implementing quarantine; implementing a highly stringent quarantine Comparators: no measure; implementing an alternative measure (e.g. screening); implementing a less stringent quarantine
Outcome	Number of studies	Summary of findings	Certainty of evidence
Outcome category: cases avoided due to the measure
Number or proportion of cases in the community	3 modelling studies	All studies reported reductions in the number or proportion of cases. These positive effects ranged from 450 fewer cases to 64028 fewer cases during the first wave of the pandemic. The variation in the magnitude of effect might be explained by differences in the population group targeted by the measure.	Very lowa,b,c ⨁◯◯◯
Proportion of imported cases	1 modelling study	The study reported that quarantining all incoming travellers would reduce the proportion of imported cases by 55% for a 7‐day quarantine period and by 91% for a 14‐day quarantine period.	Very lowb,d,e,f ⨁◯◯◯
Number or proportion of cases seeded by imported cases	3 modelling studies	All studies reported reductions in the number or proportion of cases seeded by imported cases as a result of quarantine of travellers. These positive effects ranged from a 26% (95% CI 19% to 37%) reduction to a 100% reduction. The variation in the magnitude of effect might be explained by enforcement of the quarantine, age, and the length of the quarantine period.	Very low c,g,h ⨁◯◯◯
Probability of an imported case not infecting anyone	1 modelling study	The study reported that a 14‐day quarantine of all international arrivals in New Zealand would lead to a 4% increase in probability in adults and a 14% in the elderly that an imported case would not infect anyone among adults and the elderly. The increase in the probably would be larger when a 14‐day government‐mandated quarantine is required (31% and 36% among adults and the elderly, respectively).	Very low e,f,i ⨁◯◯◯
Outcome category: shift in epidemic development
Time to outbreak	1 modelling study	The study reported that increasing the effectiveness of quarantine to 80% and 90% from the base case of 75% effectiveness would delay the peak in active cases and deaths by 3.5 and 5.5 days, respectively.	Lowe,b ⨁⨁◯◯
Outcome category: cases detected due to the measure
Days at risk of transmitting the infection into the community	2 modelling studies	Both studies reported reductions in the numbers of days that travellers, upon release, remain at risk of transmitting the infection into the community. These positive effects ranged from 0.1 fewer days to 2.1 fewer days at risk of transmission. The variation in the magnitude of effect might be explained by the length of quarantine.	Lowf,h ⨁⨁◯◯
Proportion of cases detected	1 modelling study	The study reported that requiring travellers to quarantine upon arrival in the UK would lead to detecting different proportions of cases, with the magnitude increasing with the number of days in quarantine (7‐day quarantine: 51% (95% CI 47% to 56%); 14‐day quarantine: 78% (95% CI 74% to 82%)). These proportions are higher than those for screening alone (with either thermal imaging scanners or health checks detecting 0.78% and 1.13% of cases, respectively).	Very low a,e,f ⨁◯◯◯
Probability of releasing an infected individual into the community	3 modelling studies	All studies reported reductions in the risk or probability of releasing an infected individual into the community. These positive effects included a risk ratio ranging from 0.00 (95% CI 0.00 to 0.01) to 0.59 (95% CI 0.28 to 0.85) and probabilities of releasing an infected individual ranging from 0% to 85%. The variation in the magnitude of effect might be explained by the length of the quarantine period and the risk of transmission within quarantine settings.	Very lowf,h,i ⨁◯◯◯

^aDowngraded ‐1 for risk of bias, due to major quality concerns in some studies related to the adequacy of assessment of the model’s uncertainty and incomplete technical documentation.

^bDowngraded ‐1 for imprecision, due to insufficient data reported to enable assessment of precision.

^cDowngraded ‐1 for indirectness, due to no reporting of external validation in some studies and concerns with reporting of external validation in others.

^dDowngraded ‐1 for risk of bias, due to major quality concerns in some studies related to the appropriateness of the models’ structural assumptions and adequacy of assessment of the model’s uncertainty.

^eDowngraded ‐1 for imprecision, due to only one contributing study.

^fDowngraded ‐1 for indirectness, due to no reporting of external validation in included studies.

^gDowngraded ‐1 for risk of bias, due to major quality concerns in some studies related to the appropriateness of the models’ structural assumptions, the input parameters and the adequacy of assessment of the model’s uncertainty.

^hDowngraded ‐1 for imprecision, due to a wide range of plausible effects.

^IDowngraded ‐1 for risk of bias, due to major quality concerns in some studies related to the adequacy of assessment of the model’s uncertainty.