The study population was composed of Chinese-speaking residents of Hong Kong (22) who were 18–60 years old. We conducted 2 independent cross-sectional telephone surveys to achieve the 2 aforementioned objectives. Survey 1 was conducted from December 30, 2003, to January 17, 2004, and survey 2 from May 27, 2003, to June 1, 2003. Telephone numbers were randomly selected from up-to-date residential phone directories. Nearly 100% of the Hong Kong residents have telephones at home (Hong Kong Office of the Telecommunications Authority, pers. commun.), and other local studies have used telephone surveys for data collection (21).

Interviewers called between 6:00 p.m. and 10:30 p.m. to avoid undersampling workers and students. If a telephone call was not answered, at least 2 follow-up calls were made at different hours on weekdays. An eligible household member, whose birthday was closest to the date of the interview, was invited to participate in the study. Verbal informed consent was obtained from participants. Ethical approval was obtained from the Chinese University of Hong Kong. The response rate, defined as the number of participants completing the survey divided by the number of valid households contacted, was 65% for survey 1 and 57.7% for survey 2. Relevant sociodemographic characteristics of respondents are shown in Table 1.

Respondents in survey 1 were asked about SARS-related perceptions, different public health measures currently practiced, and one's anticipated public health and emotional responses if SARS were to return to Hong Kong. Respondents in survey 2 were asked about psychological effects of SARS. These included whether respondents felt horrified, helpless, or apprehensive; had sleeping and psychosomatic problems; had increased smoking and alcohol consumption; or had perceived stress. The Chinese version of the Impact of Event Scale (IES) (23,24) and the mental health (5 items) and the vitality (4 items) subscales of the Medical Outcomes Study 36-Item Short Form Health Survey (SF-36) (25–27) were also used in survey 2.

In survey 1, multivariate logistic regression analysis, using univariately significant responses as input variables, was performed. In survey 2, 2-staged stepwise linear and logistic regression models were used. SPSS for Windows Release 11.0.1 (SPSS Inc., Chicago, IL, USA) was used and p values < 0.05 were considered significant.

The general public in Hong Kong did not perceive the possibility of a resurgence of SARS. The degree of vigilance was high when respondents were asked about current preventive behaviors and hypothetical situations of having a few new SARS cases reported in Hong Kong. The entire city was expected to react strongly to a resurgence of SARS. However, some precautions may be unwarranted and could have a negative economic effect (29). Approximately 20% of respondents believed that they would be in a state of panic, 37% were still emotionally disturbed by SARS, and 4% had family members in need of psychological or psychiatric services. Thus, the general public needs to be better prepared psychologically to be able to avoid possible panic and emotional disturbances in a resurgence of SARS.

More than 90% of respondents perceived that mask use, frequent hand washing, and disinfection of living quarters are efficacious means of SARS prevention. Although the droplet theory of transmission has been widely accepted by the scientific community, other theories involving fomites (30), aerosols (5), sewage (31), rats (7), and wild animals (32) remain controversial. No conclusions have been reached regarding these topics. Information provided by health workers has also shown marked variations (33). In the absence of confirmed "top-down" official information, the general public has apparently been forming their own attitudes in a "bottom-up" manner. Similarly, another study claimed that laypersons in Hong Kong, Taiwan, and Toronto used "naive knowledge models" that were either incomplete or faulty in conceptualizing the symptoms, threat, spread, and prevention of SARS (34). Another study also reported substantial misinformation and false beliefs related to the existence of SARS in the general public (20). Therefore, it is important to understand how perceptions were formed during a newly emerging epidemic.

If one compares the results of this study with those obtained in March 2003, SARS-related perceptions and behaviors changed sharply over time (21,35). The results of several studies show that most of the general public had always believed that SARS could be transmitted through droplets, and increasingly more people believed that SARS is transmittable through fomites, but opinions about aerosol transmission of SARS remained split (20,21,35). Different studies had similar conclusions that perceptions such as perceived efficacy and perceived susceptibility were predictive of the use of preventive measures and emotional responses (20,21,35).

In survey 2, the prevalences of avoiding hospitals and China were 66.7% and 68.6%, respectively, which are comparable with the results obtained in another study conducted in May 2003 (21). More than 80% of respondents in this study would use a mask if they had influenza while traveling, while another study conducted from April 22 to April 29, 2003, documented that ≈70% would do so (36). A third study reported that ≈50% of the general public practiced at least 5 of 7 studied types of preventive measures (20). Preventive behaviors were thus prevalent throughout different phases of the epidemic.

A study conducted on approximately April 1, 2003 (20), reported that 12.6% of the respondents were quite or very anxious. Our survey 2, which was conducted at the ending phase of the epidemic, showed that ≈16% of the respondents had moderate or severe posttraumatic stress symptoms. Another study conducted from April 11 to May 19, 2003 (37), documented that ≈68% of healthy control participants experienced negative SARS-related effects. Our study showed that ≈48% assessed their mental health as being worse because of SARS. Also, 20% of the respondents worried about finances, whereas ≈27% of the respondents had financial stress. Emotional disturbance (our survey 2) and anxiety level (20) were associated with use of preventive measures. Psychological stress was prevalent throughout different phases of the epidemic.

Sex differences in perceptions and responses were observed. Men and women may have reacted differently to the incomplete evidence available when forming their views about the spread and control of SARS. Women were more likely than men to believe that SARS could be transmitted through different modes or that different methods could effectively prevent SARS.

A sizable proportion of the population felt horrified, apprehensive, or helpless because of the SARS epidemic in Hong Kong. Approximately 40%–50% of the respondents reported that their mental health status had been worsened, and 40% felt that their levels of work- and family-related stress had increased during the epidemic. The SARS epidemic exerted adverse effects on multiple aspects of social, family, sexual, and occupational domains. Those who smoked and drank in Hong Kong also increased their frequency of smoking and drinking. Thus, the mental health effect was prevalent and pervasive. Longitudinal studies are therefore required to understand the long-term mental health effects of SARS. Similar effects had been documented in studies conducted after the September 11, 2001, terrorist attack in the United States (15,38,39). Some similarities may exist in the community responses of different large-scale disasters.

Married persons tended to have a worsened mental health status because of SARS. Married people usually have a lower prevalence of psychological problems and a better support system compared with single people. However, ≈25% of respondents reported that their family members were emotionally affected by the epidemic, and ≈40% reported increased family stress. When an infectious epidemic is being faced, the worries of cross-infection and the well-being of family members are critical in determining the mental health effects of the epidemic on a person. Mental health services should take into account mutual influences among family members. Increased work-related stress was another predictor of mental health effects. Business activity decreased sharply, and the job security of many people was threatened. Similarly, social life was reported as worse among 40% of respondents. The effect of SARS was not confined to physical and psychological aspects, but it also affected socioeconomic and social aspects, which in turn determined the psychological well-being of persons.

This study had several limitations. First, data were self-reported and are subject to reporting biases. However, the interviews were anonymous. Second, some questions were asked about behavior in response to a potential resurgence of SARS, rather than measuring actual behavior because we were investigating how the general public would respond to a possible resurgence of SARS. Third, the response rates of the studies were moderate (≈58% in survey 1 and 65% in survey 2), and no data were available from nonresponders. The response rates were comparable with those of other survey studies in Hong Kong (40,41), and the age composition of the 2 samples was comparable with those of the Hong Kong census figures. Furthermore, we were not able to ascertain the previous psychological conditions of the respondents. However, results of the study should reflect the direct effect of SARS, rather than the general psychological status of the respondents. Some important factors, such as intensity of media exposure, were not measured in the study. However, many variables in this study (e.g., perceived reaction to resurgence and some psychological responses variables) have not been reported elsewhere.

SARS may return to some parts of the world, and preparative work is warranted. Up-to-date SARS-related knowledge should be collated and disseminated to the general public to promote effective public health measures and avoid unnecessary panic in case of a resurgence. Sex differences and concerns for family members and work need to be considered by relevant information campaigns. The perception of the general public changes rapidly over time and needs to be monitored closely. Bioterrorism may be similar to SARS in many ways. The results of this study predict that, in cases of bioterrorism, the general public would form their perceptions based on weak evidence, and the effect on mental health would also be evident. Modifying perceptions of the public would facilitate control of the disaster and alleviate panic among the general population. Further studies on the process of perception formation and its consequences on psychological responses in newly emerged epidemics are warranted.

This study was supported by the Chinese University of Hong Kong.