Psychosocial effects of SARS on hospital staff: survey of a large tertiary care institution

Methods

We distributed a self-administered questionnaire to employees of SWC, a large teaching hospital, between Apr. 10 and 22, 2003. The hospital comprises 3 campuses: the Sunnybrook site, the Women's College site and the Orthopaedic and Arthritic site. During the first phase of Toronto's SARS outbreak most patients with SARS were admitted to the SARS unit of the Sunnybrook site. The Women's College site was the location for Toronto's first SARS screening and assessment clinic.

Questionnaires were distributed to all willing employees (both health care and non-health-care workers) entering the hospital between 5:45 am and 7:15 pm (covering all major shift changes) over a 3-day period. Access at each campus was limited, making it possible to ensure that most people working that day were handed a questionnaire and explanatory letter. Completed questionnaires were deposited in drop boxes located at specified hospital locations over the following 2 weeks. Reminders to staff were sent out by email. Announcements were made at 2 staff fora on SARS held during the study period, and questionnaires were handed out to those who, when asked, stated that they had not received a questionnaire previously.

The questionnaire consisted of 5 main sections: occupation and work history; closed and open-ended questions about the respondent's concerns about SARS; closed and open-ended questions about the use and effects of SARS precautionary measures; standard sociodemographic characteristics; and the 12-item version of the General Health Questionnaire (GHQ-12),11 a frequently used and well-standardized measure of recent emotional distress. We pilot-tested the survey instrument among clinicians and health care services researchers to assess its face validity. Owing to resource constraints, the distribution of the GHQ-12 was limited to a subsample of staff made up of only those who received the questionnaire on day 2 of data collection. A copy of the questionnaire is available from the authors on request.

We coded the response categories on the GHQ-12 items using the GHQ-12 scoring method as recommended by Goldberg and Williams11 and calculated a total score. Following the example of numerous authors,12^,13^,14^,15 we used a threshold score of greater than 3 to identify the presence of emotional distress manifested as a break from normal functioning (e.g., loss of sleep, loss of self-confidence or inability to make decisions). The α reliability coefficient for the GHQ-12 in this study was 0.85.

We analyzed the data using SPSS (version 11.1, SPSS Inc., Chicago) and included bivariate and multivariate techniques. Bivariate analysis was used to identify potential explanatory variables for outcomes of interest. Multivariate logistic regression analysis was used to assess the association between outcome variables and potential predictors, while adjusting for other identified explanatory variables.

We constructed 2 logistic regression models for the dependent variables: GHQ-12 threshold score and concern for the health of oneself or one's family. We ran models using a backward stepwise selection algorithm. Variables were determined to contribute to the model if the significance level for the Wald inclusion test statistic was less than 0.05. Owing to its a priori importance, sex was forced into every model, regardless of its contribution. The analysis of the data for all open-ended questions involved grouping like responses into categories and then analyzing the categories quantitatively using bivariate and multivariate techniques, as described in the preceding paragraph.

Ethics approval for the study was obtained from the SWC Research Ethics Board.

Results

We distributed 4283 questionnaires across the 3 hospital sites, of which 2001 (47%) were returned, representing 27% of the total SWC employee population of 7474. Of the 2001 returned questionnaires, 510 contained the GHQ-12. With the exception of occupation, no significant differences were found between the subsample that completed the GHQ-12 and the full sample. There was a higher proportion of nurses and a lower proportion of allied health care professionals in the full sample than in the subsample.

The demographic characteristics of the respondents are shown in Table 1. Of the 1983 respondents who indicated their sex, 1563 (78.8%) were women. Of the 1857 who indicated their occupation, 615 (33.1%) were allied health care professionals, 593 (31.9%) worked in areas other than patient care (e.g., administration, food services, maintenance or research), 476 (25.6%) were nurses, and 173 (9.3%) were doctors; these proportions are representative of the hospital staff population as a whole. Half of the respondents stated that they had worked 10 or more years in their current occupation, and 83.7% (1645/1966) reported full-time status.

In all, 64.7% of the respondents reported concerns about their own health during the SARS outbreak, and almost the same proportion (62.7%) reported concerns about their family's health (Table 2). Among the various occupations, nurses were most frequently concerned (363 [76.3%]), and doctors were least frequently concerned (104 [60.1%]). Almost all (93.8%) of those who reported concerns felt that they had friends, family or other people to talk to about those concerns (data not shown).

Of the 510 respondents who completed the GHQ-12, 148 (29.0%) scored above the threshold of greater than 3, indicating probable emotional distress (Table 2). Statistically significant differences (p < 0.001) were found between the occupational groups in the proportion with a score greater than 3: 45.1% (37/82) of nurses, 33.3% (66/198) of allied health care professionals, 17.4% (8/46) of doctors and 18.9% (28/148) of staff not working in patient care (data not shown).

Wearing a mask was the precaution most frequently cited as most bothersome (Table 3). The most commonly cited difficulty with the mask was physical discomfort (92.9% [1588/1710] of respondents). Three-quarters of the respondents reported that the control measures were sufficient to prevent contracting SARS.

Reported negative effects of the SARS outbreak included financial losses, being treated differently by people because of working in a hospital and changes to personal and familial lifestyle (Table 4). Frequently mentioned lifestyle changes included avoiding public spaces (e.g., restaurants and shopping centres) and avoiding interaction with family or friends. Over half of the respondents (1161 [58.0%]) also reported at least 1 positive effect (Table 4). A total of 493 (41.1%) felt there was an increased awareness of disease control, 317 (26.4%) found the SARS outbreak to be a learning experience, and 285 (23.8%) felt an increased sense of togetherness and cooperation. Other positive aspects included being less busy than usual and feeling a greater appreciation of life and work.

Logistic regression analysis identified 4 factors as being significantly associated with increased levels of concern for personal or family health: perception of a greater risk of death from SARS (adjusted odds ratio [OR] 5.0, 95% confidence interval [CI] 2.6–9.6), living with children (adjusted OR 1.8, 95% CI 1.5–2.3), personal or family lifestyle affected by the SARS outbreak (adjusted OR 3.3, 95% CI 2.5–4.3) and being treated differently by people because of working in a hospital (adjusted OR 1.6, 95% CI 1.2–2.1) (Table 5). Three factors were found to be significantly associated with decreased levels of concern: working in a management or supervisory position (adjusted OR 0.6, 95% CI 0.4–0.8), belief that the precautionary measures in the workplace were sufficient (adjusted OR 0.4, 95% CI 0.3–0.5) and age 50 years or more (adjusted OR 0.6, 95% CI 0.4–0.9).

Regression analysis identified 4 factors as being significantly associated with the presence of emotional distress, as identified with the GHQ-12: being a nurse (adjusted OR 2.8, 95% CI 1.5–5.5), part-time employment status (adjusted OR 2.6, 95% CI 1.2–5.4), lifestyle affected by the SARS outbreak (adjusted OR 2.2, 95% CI 1.4–3.5) and ability to do one's job affected by the precautionary measures (adjusted OR 2.9, 95% CI 1.9–4.6) (Table 6).

Interpretation

During the peak of the first phase of the SARS outbreak in Toronto, two-thirds of SWC staff experienced concern for their own or their family's health. This finding is consistent with a recent study by Maunder and colleagues.1 Those who felt that they were being treated differently by people because they worked in a hospital were more likely to report health concerns than were those who did not feel they were being treated differently. This association between stigmatization and concern is in keeping with an earlier finding in a study of an outbreak of vancomycin-resistant enterococci in an Australian hospital.7 In our study, being in a management or supervisory position reduced the likelihood of reporting concerns. This is consistent with previous research indicating that having some level of control (real or perceived) over a situation reduces the risk of psychosocial effects.1^,16^,17^,18

A total of 29% of the respondents experienced emotional distress, as evidenced by their score on the GHQ-12. This rate is more than double that found in a recent study of the general adult population in Canada.19 Other North American and British studies have typically shown prevalence rates of emotional distress in general populations of 10% to 24%.20^,21^,22 In a study of nurses working under normal circumstances in 3 Singapore hospitals, 14.8% scored above the threshold point on the GHQ,23 as compared with 45% of the nurses in our study. We found that part-time staff were more likely than full-time staff to experience emotional distress; we attribute this to part-time staff's receiving less current information, having a reduced sense of involvement in the hospital's response to the SARS situation and having a weaker social support network within their work environment.

Positive aspects of the SARS outbreak were also reported. For instance, staff noted an increased awareness of infection control, which may continue to benefit the hospital community in the future. Some respondents found that the situation provided a positive learning experience and an increased sense of togetherness and cooperation among the hospital staff, which may provide an opportunity to continue to build stronger relationships.

Our study has several potential limitations. First, the fact that the response rate was relatively low (47%) adds to the possibility of response bias. Second, staff who were in quarantine or away because of illness or vacation were not included; some of these people may have been among those most concerned and most affected by the outbreak. Third, the use of the GHQ-12 was limited to a portion of the sample, which may have affected the results. Fourth, because the study was conducted in a large, urban teaching hospital, the results may not be generalizable to other hospital or community settings. Finally, the reported financial effects of the SARS outbreak were likely underestimated, as the data were collected in the initial weeks after the infection control measures were implemented.

Our findings demonstrate a significant psychosocial impact of the SARS outbreak in Toronto on hospital staff. The effects on the family and lifestyle of staff were substantial, highlighting the need for greater personal and family support for employees during such crises in addition to educational interventions to address psychosocial distress and concern. Although the hospital provided numerous staff fora and daily SARS updates to keep staff informed and to address questions, there may be a need for other approaches for some staff. Development of less-restrictive infection control equipment that is more acceptable to staff also appears to be an important priority.

Our study was conducted during the peak of the initial phase of the SARS outbreak in Toronto, when knowledge of the disease was limited and information was rapidly changing. As the disease evolves and subsequent provincial and institutional policies and practices are developed and implemented, staff perceptions and experience may change. Future follow-up investigations, using both qualitative and quantitative approaches, will be necessary to understand the psychosocial effects of SARS on hospital staff over time.

𝛃 See related article page 811