Ascertaining and isolating case-patients, combined with rapid identification and management of contacts, were highly effective in interrupting transmission in several countries (1–6).² For example, a study in Singapore demonstrated a correlation between rapidly isolating patients after onset of symptoms and a decreased number of secondary cases among their contacts (4) (Figure). Contacts in these countries were placed in various forms of quarantine or, less commonly, monitored for symptoms without confinement and isolated if and when symptoms emerged. The location of quarantine was usually at home but was sometimes at a designated residential facility (e.g., for travelers, persons who did not wish to remain at home for fear of exposing their families, homeless persons, and noncompliant persons). In some cases, quarantined persons were allowed to leave the quarantine site with the permission of local health authorities if they wore masks and did not use public transportation or visit crowded public places. In at least one area, these restrictions were applied to essential workers and termed "work quarantine."

Several respondents emphasized that the modern concept of quarantine differs greatly from quarantine in past centuries. Quarantine is most acceptable and arguably most effective when protecting the health and rights of quarantined persons is emphasized. In previous centuries, sick and exposed persons were often locked up together and received limited medical care. Moreover, quarantine was sometimes applied in an arbitrary and discriminatory fashion, targeting lower socioeconomic classes and racial minorities. The modern concept emphasizes science-based interventions with attention to the medical, material, and mental health needs of quarantined persons and protecting fundamental human rights. Exposed persons who are not sick should be separated from symptomatic patients, monitored for the minimum time necessary (e.g., one maximum incubation period), and provided appropriate medical care at the first sign of illness during the monitoring period. Quarantine may be applied to individual persons, to small groups, or, in extreme cases, to entire neighborhoods or other geographic districts ("cordon sanitaire") (7,8).

In the SARS epidemic, persons under quarantine were mostly confined at home and actively monitored for symptoms. In several countries, quarantine was legally mandated and monitored by neighborhood support groups, police and other workers, or video cameras in homes. In other areas, compliance was "requested," but court orders were issued for a small percentage of noncompliant persons. Reports indicate that SARS was diagnosed in 0.22% of quarantined contacts in China-Taiwan, 2.7% in China-Hong Kong Special Administrative Region (SAR), and 3.8%–6.3% in China-Beijing. These different rates were partly due to different criteria for placing persons in quarantine. Contacts at highest risk (aside from healthcare workers with certain unprotected patient care exposures) had been exposed to ill family members (6,9–11).

Quarantine led to financial and psychosocial stresses, risk communication, compensation, and workforce staffing issues for persons, families, employers, and governments. Legal appeals and defiance of quarantine orders were rare (2,6,8–13).

The optimal management of contacts, stratified according to risk of becoming ill, remains under discussion in several countries, e.g., whether confinement is always needed or close monitoring of health status without confinement would suffice. Reports from Canada indicate that the insidious onset of symptoms sometimes posed challenges for clinicians and public health officials. "Timely diagnosis and isolation of cases were sometimes hindered by delays in patient recognition of symptoms, obtaining medical evaluation, and/or physician recognition of the significance of symptoms, which occasionally waxed and waned early in illness" (A. McGeer and D. Low, Mount Sinai Hospital Toronto, pers. comm.). "In Toronto, some healthcare workers continued to work without recognizing that they were ill, perhaps confusing their symptoms with fatigue, despite daily screening and repeated messages not to come to work if ill. This resulted in transmission to patients and staff" (B. Henry, Toronto Public Health, pers. comm.).

Measures to increase social distance, e.g., canceling mass gatherings; closing schools, theaters, and public facilities; and requiring masks for all persons using public transport, working in restaurants, or entering hospitals, were implemented in areas where extensive unlinked community transmission of SARS coronavirus (SARS-CoV) was suspected. Many persons in these areas also chose to wear masks outside their homes. These measures were often applied simultaneously with other measures, including enhanced contact tracing, which makes their independent effectiveness difficult to assess. However the simultaneous introduction of a variety of measures was temporally associated with dramatic declines in new SARS cases. A case-control study in Beijing found that wearing a mask more frequently in public places may have been associated with increasing protection (15). Another case-control study in China-Hong Kong found that using a mask "frequently" in public places, washing one's hands >10 times per day, and "disinfecting living quarters thoroughly" appeared to be protective (16). The types of masks used were not specified. With the exception of the Amoy Gardens cluster in which SARS-CoV was apparently transmitted through accidentally produced aerosols of sewage (17), SARS transmission in the community from aerosols or in social settings appeared to be rare.

In some areas, disinfectants were applied inside the homes and vehicles of persons with SARS, ambulance tires, and pedestrian walking zones. Little information exists on the effectiveness of disinfectant use in reducing community or hospital transmission. In Hong Kong, disinfecting living quarters thoroughly (not otherwise defined and reported retrospectively by telephone) appeared to be protective (16).

Travel advisories (e.g., advice to postpone nonessential travel) were issued by WHO and various governments. Air travel to areas affected by the advisories decreased dramatically during the epidemic (M.A. Hinayon and D. Gamper, Airports Council International, communication to WHO), although the impact of advisories compared with other sources of information to travelers, such as news media reports of SARS cases, is difficult to assess.

Passive and active methods were used to provide information and screen entering and exiting travelers. These methods included signs, videos, public address announcements, distributing health alert notices, administering questionnaires to assess symptoms and possible exposure, visual inspection to detect symptoms, and thermal scanning.

Few data exist on the relative effectiveness of methods of providing information to travelers. Available data on the effectiveness of screening and other measures directed to travelers are sometimes difficult to interpret because they may not distinguish between entry and exit screening, specify how many entering travelers were from affected countries, distinguish the epidemic period from subsequent, or include the number of SARS cases detected.

Combined data from Canada, China (mainland, Hong Kong SAR, and Taiwan), France, Singapore, Switzerland, Thailand, and the United States indicate that approximately 31 million travelers entering these countries received health alert notices. Of these, approximately 1.8 million were reported as arriving from affected areas; this estimate is likely low given the difficulties in tracking travelers and the fact that many airline passengers change planes en route. Inadequate data exist to evaluate the effect of distribution of most of these notices. China-mainland reported distributing 450,000 notices and detecting four SARS cases that may have been linked to the notices (M. Song, China Dept of Health and Quarantine Supervision and Management, communication to WHO). Thailand reported having printed 1 million notices and detecting 113 cases of illness directly linked to the notices (108 at airports, 1 at a seaport, and 4 at land crossings). Twenty-four cases were suspected or probable SARS: all of which were detected at airports (S. Warintrawat, Ministry of Public Health, Thailand, communication to WHO).

Preliminary data from a worldwide survey indicate that among 72 patients with imported probable or confirmed SARS cases, 30 (42%) had onset of symptoms before or on the same day as entry into the country and symptoms developed in 42 patients (58%) after entry (J. Jones, United Kingdom Health Protection Agency, communication to WHO). A small percentage of persons completing entry health declaration questionnaires in affected areas during the SARS epidemic were diagnosed with SARS (Table 1).

Results combined from Canada, China (including the mainland and Hong Kong SAR), and Singapore indicate that no cases of SARS were detected by thermal scanning among >35 million international travelers scanned at entry during the SARS epidemic (Table 2; data for Hong Kong SAR include travelers arriving from China-mainland). Temperature screening of 13,839,500 travelers entering or leaving Beijing by air, train, or automobile identified 5,097 patients with fever, of whom 12 had probable SARS. These 12 included 10 of 952,200 domestic airline passengers and 2 of 5,246,100 train passengers. None of 275,600 international travelers who underwent temperature screening had SARS (5).

In China-Taiwan, incoming travelers from affected areas were quarantined; probable or suspected SARS was diagnosed in 21 (0.03%) of 80,813. None of these 21 was detected by thermal scanning when they entered Taiwan (9) (S.K. Lai, Taiwan Center for Disease Control, pers. comm.).

After WHO recommended exit screening on March 27, 2003 (18), no additional cases from airline travel were documented from countries with screening. Combined data from China (Hong Kong SAR and Taiwan) indicate that among 1.8 million people who completed health questionnaires at exit, 1 probable case of SARS was detected. Combined data from Canada, China (Hong Kong SAR and Taiwan), and Singapore indicate that no cases of SARS were detected among >7 million people who underwent thermal scanning at exit (Table 3) (S. Courage, Health Canada, S.K. Lai, China-Taiwan Center for Disease Control; P.L. Ma, Hong Kong SAR China Dept of Health; and B.K.W. Koh, Singapore Ministry of Health, communications to WHO). In some areas, "stop lists" were used at borders to prevent persons on isolation or quarantine lists from exiting. Anecdotes suggest that exit screening may have helped dissuade ill persons from traveling by air but may have been more successful in dissuading local residents from traveling abroad than in dissuading ill travelers from attempting to return home.

Five commercial international flights were associated with transmission of SARS from patients with symptomatic probable cases to passengers and crew (1). Notification of exposed passengers and studies of transmission risk were greatly hampered by difficulties in identifying and tracing passenger contacts (19–23). In the most comprehensive investigation, involving three flights with extensive passenger tracing and laboratory confirmation of index and secondary cases, a wide range of risk was noted (Table 4). For flight 2, in which the secondary attack rate was 18.3%, the risk of infection was increased for persons seated close to the index patient, but most passengers who became infected were seated farther away, even though their individual risk was lower (19). In another study, one person with SARS, who had difficulty breathing but was not coughing, infected two other passengers. One of these sat in the row in front of the index patient but the other passenger sat four rows, plus a passageway, behind and on the opposite side of the plane (20). On nine flights arriving in Singapore, the incidence of transmission from passengers with SARS who had respiratory symptoms was estimated at 1 in 156 persons (21). A fourth study found no transmission to passengers seated near a patient who took multiple flights (22). In comparison, an influenzalike illness developed within 3 days in 72% of passengers in a plane containing a person with symptomatic influenza and grounded for 3 hours without ventilation (24). The risk for transmission of tuberculosis during a long flight was also increased among, but not limited to, passengers seated close to a highly infectious index patient (25).