Table 4.

 Characteristics of included retrospective cohort studies

Study	Participants	Interventions v comparisons	Outcomes	Risk of bias (notes)
Doherty 1998w40	Children aged <2 years with differential diagnosis of respiratory syncytial virus infection	Diagnosis of respiratory syncytial virus infection and cohortingv normal care	“RSV infection reduced” (but data did not support conclusion)	High (poor descriptions)
Isaacs 1991w41	Children aged <2 years with differential diagnosis of respiratory syncytial virus infection	Isolation and handwashing with alcohol based hand rubs (Amphisept 80; GoldsCHmidt) v normal care	Respiratory syncytial virus infection reduced by “up to 60%”	High (poor descriptions)
Ou 2003w42 w43	171 cases of SARS and 1210 people quarantined from selected districts in China	Quarantine at home or hospital for 14 days after exposure: comparisons between reductions of incidence (95% CIs) of SARS for carers 31% (20% to 44%), visitors 9% (3% to 22%), and cohabiting contacts 5% (2% to 9%)	SARS attack rates reduced for all groups except non-cohabitants living in same building; carers of cases during incubation period (quarantine therefore not necessary)	High (non-random basis for sample, selection bias of sample and responders, recall bias of responders, and absence of laboratory confirmed diagnosis may have affected conclusion. Overall, insufficient denominator data, or data on non-exposed people, precluded data extraction or calculation of odds ratios)
Yen 2006w44	One intervention military hospital (459 healthcare workers) and 86 control hospitals in Taiwan	Integrated infection control strategy: triage and barrier traffic flow into hospital, zoning of risk, negative pressure areas of isolation, personal hygiene, and barrier interventions v normal isolation procedures	Only two healthcare workers infected with SARS, compared with 50 probable cases and 43 suspected cases in control hospitals	High (sketchily reported study with missing denominators and data on exposure to SARS. Not clear how intervention differed from high risk isolation procedures)