Original Research

Immediate and Sustained Psychological Impact of an

Emerging Infectious Disease Outbreak on
Health Care Workers

Grainne M McAlonan, MBBS PhD1, Antoinette M Lee, PhD1, Vinci Cheung, MPhil2,
Charlton Cheung, BSc2, Kenneth WT Tsang, FRCP3, Pak C Sham, MRCPsych, PhD4,
Siew E Chua, MRCPsych1, Josephine GWS Wong, MRCPsych1

Objective: To assess the immediate and sustained psychological health of health care workers who were at high
risk of exposure during the severe acute respiratory syndrome (SARS) outbreak.
Methods: At the peak of the 2003 SARS outbreak, we assessed health care workers in 2 acute care Hong Kong
general hospitals with the Perceived Stress Scale (PSS-10). One year later, we reassessed these health care
workers with the PSS-10, the 21-Item Depression and Anxiety Scale (DASS-21), and the Impact of Events
Scale-Revised (IES-R). We recruited high-risk health care workers who practised respiratory medicine and
compared them with nonrespiratory medicine workers, who formed the low-risk health care worker control
group.
Results: In 2003, high-risk health care workers had elevated stress levels (PSS-10 score = 17.0) that were not
significantly different from levels in low-risk health care worker control subjects (PSS-10 score = 15.9). More
high-risk health care workers reported fatigue, poor sleep, worry about health, and fear of social contact, despite
their confidence in infection-control measures. By 2004, however, stress levels in the high-risk group were not
only higher (PSS-10 score = 18.6) but also significantly higher than scores among low-risk health care worker
control subjects (PSS-10 score = 14.8, P < 0.05). In 2004, the perceived stress levels in the high-risk group were
associated with higher depression, anxiety, and posttraumatic stress scores (P < 0.001). Posttraumatic stress
scores were a partial mediator of the relation between the high risk of exposure to SARS and higher perceived
stress.
Conclusions: Health care workers who were at high risk of contracting SARS appear not only to have chronic
stress but also higher levels of depression and anxiety. Front-line staff could benefit from stress management as
part of preparation for future outbreaks.
(Can J Psychiatry 2007;52:241–247)
Information on funding and support and author affiliations appears at the end of the article.

Clinical Implications
· One year after the outbreak, health care workers who were at high risk of contracting SARS had elevated levels of
stress, depression, and anxiety.
· Mental health measures should be available to identify and limit psychological morbidity in this high-risk group.
· This information should guide preparedness planning for future infectious disease outbreaks.

Limitations
· Data were obtained by self-report because health care workers were in isolation wards during the first study
assessment and because there were issues about stigmatization and legal compensation at the time of the second
assessment.
· We could not formally evaluate the specific psychometric properties of our questionnaire on the psychological effects
of SARS.
· Owing to the anomymous nature of the data collection, we could not pair individuals at the 2 time intervals .

The Canadian Journal of Psychiatry, Vol 52, No 4, April 2007 W 241

Original Research

Key Words: severe acute respiratory syndrome, outbreak, perceived stress levels in 2004. Such information could help
health care workers, stress, Hong Kong alleviate or prevent future psychiatric morbidity in health care
workers and could contribute to an overall strategy of
ffective containment of infectious disease outbreaks has
E become a global health imperative.1 Cocirculation of ani-
mal and human viruses may allow the exchange of genetic
advanced preparedness for future pandemics.

Methods
material and create a new virus with potential to trigger a seri- During the peak period of hospital admissions for SARS
ous pandemic.2,3 The 1918 Spanish influenza virus killed over (mid-April to mid-May 2003), we distributed questionnaires
20 million individuals and could recur today with increased containing the PSS-1011 to health care workers (that is, doc-
virulence.4 In April 2005, unlabelled samples of the H2N2 tors, nurses, and health care assistants) in 2 acute care general
virus, responsible for more than 1 million deaths in 1957, were hospitals in Hong Kong. Health care workers from respiratory
inadvertently couriered to 5000 laboratories in 19 countries medicine departments were considered high-risk because
and subsequently needed swift destruction.5 The experience they were front-line workers for patients with novel infectious
of SARS, more cases of avian influenza in Asia and Europe, respiratory diseases. Health care workers who did not work in
and the threat of person-to-person transmission of influenza A respiratory medicine departments were recruited as low-risk
(H5N1),6 in combination with contemporary air travel, mean health care worker control subjects. The questionnaires were
that the threat of another pandemic is serious.7 In such an self-administered and brief because research personnel could
event, health care workers will no doubt be affected and possi- not interview high-risk participants who were in isolation
bly infected as they care for their patients. Health care workers wards along with all other suspected SARS patients. To facili-
comprised almost 50% of SARS cases in 2003, making the tate systematic recording of psychological responses, study
SARS outbreak a rehearsal for the next great influenza pan- subjects could select any of 40 possible psychological
demic.8 “History tells us that protection against a viral pan- responses (16 positive and 24 negative) and rate their confi-
demic will always depend on the availability of a local cadre dence in infection-control measures. This portion of the
of skilled public-health workers.” 9, p 1372 instrument was deemed to have face validity by an expert
The courage of health care workers during the 2003 SARS panel of 5 psychiatrists and 5 physicians.10 Questionnaires
outbreak did not make them immune to anxiety or stress, were anonymous to guarantee confidentiality. The study was
despite the fact that, at the time of the outbreak, the stress lev- approved by the local institutional review board. Question-
els of high-risk workers did not appear to differ from those of naires from isolation wards were quarantined for 72 hours and
the community.10 However, possible longer-term, residual scrutinized for contamination before data entry.
psychological impacts and morbidity have not been previ-
In 2004, 1 year after the outbreak, health care workers from
ously evaluated. We therefore undertook the first study to
the same clinical teams in the wards from the 2 hospitals were
investigate the sustained psychological impact of the outbreak
again invited to characterize their psychological state by com-
on health care workers in 2 acute care general hospitals. We
pleting the PSS-10, the DASS-21,12 and the IES-R.13 Once
sought to examine changes in perceived stress among
again, all questionnaires were brief, self-adminstered, and
high-risk health care workers over time, from the height of the
anonymous.10 This feature of our study meant that, although
outbreak in 2003 to 1 year after. In 2004, we extended our
the groups were made up of essentially the same individuals,
study to more formally characterize the psychological health
we were unable to follow up change in individual participants
of health care workers. We sought to determine whether
from the first assessment to the second. This was unavoidable
posttraumatic stress associated with risk of exposure to SARS
because we needed to ensure absolute anomymity in the face
among high-risk health care workers contributed to their
of stigma and ongoing compensation lawsuits.

Statistical Analysis
Group comparisons of nominal variables were analyzed by
chi-square test, ordinal variables by Mann–Whitney U tests,
Abbreviations used in this article
and interval variables by t tests. Bonferroni’s correction was
ANOVA analysis of variance
made for multiple comparisons. We used 2-way ANOVA to
DASS-21 21-item Depression Anxiety Stress Scales compare groups on differences in perceived stress across
IES-R Impact of Events Scale-Revised time. We calculated Spearman’s rho for correlations with the
PSS-10 10-item Perceived Stress Scale PSS-10 score or for psychological responses. We conducted
SARS severe acute respiratory syndrome hierarchical multiple regression analyses based on Baron and
SD standard deviation Kenny’s criteria14 to test the mediation model (Figure 1), in
which posttraumatic stress associated with risk of exposure to

242 W La Revue canadienne de psychiatrie, vol 52, no 4, avril 2007

 Immediate and Sustained Psychological Impact of an Emerging Infectious Disease Outbreak on Health Care Workers

Figure 1 Mediation model of risk of exposure, posttruamatic stress owing

to the impact of SARS outbreak, and perceived stress

b = 0.3, t = 4.1a b = 0.5, t = 8.9a

P < 0.0001, R2 = 0.12 P < 0.0001, R2 = 0.36

High- Posttraumatic
compared with stress owing to Perceived
Low-risk impact of SARS stress
SARS outbreak
exposure
b = 0.3, t = 4.3, P < 0.0001, R2 = 0.23a

b = 0.2, t = 2.7, P < 0.01, R2 = 0.11b

a
Regression analysis with age and education level as covariates
b
Regression analysis with the impact of SARS outbreak, age, and education level as covariates
There was a significant effect of mediation by the impact of SARS outbreak;
Sobel test (z = 3.6, P < 0.0001)

SARS was hypothesized to mediate the effect of perceived In 2004, we again invited health care workers to participate.
stress levels in 2004. According to Baron and Kenny’s14 crite- As was the case with the first survey, we excluded incomplete
ria for testing the mediation effect, mediation is evidenced or unclear questionnaires, which left 184 of a total of 280
when a predictor is significantly related to the mediator, when (66%) valid questionaires. This follow-up sample comprised
the mediator is significantly related to the dependent variable, 71 high-risk health care workers and 113 low-risk health care
and when the relation between the predictor and the depend- workers. The 2 groups were matched on sex and education
ent variable is significantly reduced when the mediator is level but differed significantly in age (÷2 = 9.80, P < 0.05);
included in the analysis. In the present study, we tested a therefore, we controlled for age in the group comparisons of
mediation model (Figure 1) in which difference in perceived the 2004 data (see Table 1). As Table 1 shows, in 2003 and
stress between the high- and low-risk health care workers was 2004, the groups were unevenly distributed in terms of age
mediated by the IES-R13 scores measuring posttraumatic (÷2 = 73.70, P < 0.001) and education level (÷2 = 15.87,
stress owing to the impact of the SARS outbreak. As the pre- P < 0.05) but not in terms of sex (÷2 = 4.21, P not significant).
dictor variable, we entered risk group (high-risk, compared Hence, when we compared high-risk and low-risk health care
with low-risk, health care workers); as the dependent variable, workers across the time point, we controlled for age and edu-
we entered the perceived stress as measured by PSS-10. cation level as possible confounding variables

Results General Stress Levels

Subjects In 2003, both high-risk and low-risk health care workers had
In 2003, we discarded incomplete or unclear questionnaires, equally high perceived stress levels as measured by the
leaving 210 (79%) of a total 266 valid. We also excluded 30 PSS-10 (t164 =–1.36, P = 0.176). Their mean PSS-10 scores
health care workers who had a suspected or confirmed diag- (17.0, SD 5.66, for high-risk subjects and mean 15.9, SD 4.68,
nosis of SARS and 4 who had a history of psychiatric illness, for low-risk subjects) were higher than the normative value of
leaving 176 health care workers. The final sample included 13 in a US community sample (for reference, the community
106 high-risk health care workers (who worked in SARS iso- normative PSS-10 values for life events included unemploy-
lation units) and 70 low-risk health care workers (who worked ment = 16.5, separation = 16.6, and work disability = 19.911).
in psychiatric inpatient units). We matched groups on age, For high-risk workers only, PSS-10 ratings were significantly
sex, and education level (see Table 1). correlated with age (Spearman’s rho =–0.26, P < 0.01,

The Canadian Journal of Psychiatry, Vol 52, No 4, April 2007 W 243

Original Research

Table 1 Comparison of high- and low-risk health care workers in 2003 and 2004
2003 2004

High-risk (n =106) Low-risk (n = 70) High-risk (n = 71) Low-risk (n = 113)

Variable n (%) n (%) n (%) n (%)

Age (years)
30 to 40a 60 (56.6) 31 (44.3) 29 (40.8) 35 (31.0)
41 to 50 39 (36.8) 31 (44.3) 13 (18.3) 42 (37.2)
Female sex 75 (70.8) 53 (75.7) 47 (66.2) 71 (62.8)
Education(at least secondary level)b 100 (94.4) 64 (91.4) 71 (100.0) 111 (98.2)
a
P < 0.05 for 2004 health care workers only
b
P < 0.05 for 2003 and 2004 health care workers

Table 2 Stress levels in 2003 and 2004

2003 2004

High-risk Low-risk High-risk Low-risk

Instrument Mean (SD) Mean (SD) Mean (SD) Mean (SD)

PSS-10 total 17.0 (5.7) 15.9 (4.7) 18.6 (4.9) 14.8 (5.0)
DASS-21 total — — 16.4 (12.7) 8.3 (7.5)
DASS-21 Depression — — 4.9 (4.8) 2.2 (2.6)
DASS-21 Anxiety — — 4.9 (4.3) 2.1 (2.5)
DASS-21 Stress — — 7.2 (4.8) 4.1 (3.2)
IES-R total — — 17.2 (14.2) 9.3 (11.2)

Bonferroni-corrected); they did not interact with education infection-control practices, compared with only 64% of
level for either group. low-risk health care workers (÷2 = 8.2, P < 0.05).

Psychological Effects of SARS in 2003 Difference in Stress Levels Between 2003 and 2004
(Table 2)
In 2003, total negative psychological responses correlated
Counterintuitively, perceived stress remained the same from
with PSS-10 ratings in both high-risk (Spearman’s rho = 0.5,
2003 to 2004 for both high-risk health care workers (F1,170 =
P < 0.05) and low-risk (Spearman’s rho = 0.6, P < 0.05) health
1.35, P not significant) and their low-risk colleagues (F1,164 =
care workers, but overall, both groups selected a significantly
2.84, P not significant). However, further analysis with 2-way
higher percentage of positive (35.8%) responses (such as
ANOVA showed that, although there was no significant time
“unity with others,” “civic-mindedness,” and “bravery”),
main effect for either group, a significant time × risk level
compared with negative (14.6%) responses (t174 = 12.5, P <
interaction effect was found. This indicated that the change in
0.05). However, high-risk health care workers selected a sig-
perceived stress from 2003 to 2004 was significantly different
nificantly higher percentage of negative responses (17.4%),
for the 2 groups (F1,336 = 4.61, P < 0.05), with a general trend
compared with low-risk subjects (10.4%) (t173 =–2.8, P <
toward a decrease over time for low-risk health care workers
0.005). A higher percentage of high-risk health care workers
and an increase over time for high-risk health care workers.
reported fatigue (70.3%, compared with 22.1% of low-risk
workers; ÷2 = 37.9, P < 0.05), poor sleep (30.2%, compared High-Risk Health Care Workers Remained Highly
with 7.4%; ÷2 = 12.7, P < 0.05), worry about health (57.3%, Stressed After the Outbreak
compared with 41.2%; ÷2 = 4.1, P < 0.05), and fear of social In 2004, perceived stress levels were significantly higher
contact. (41.7%, compared with 23.5%; ÷2 = 5.8, P < 0.05). In among high-risk health care workers (mean score 18.56, SD
2003, 84% of high-risk health care workers felt confident in 4.91) than among low-risk control subjects (mean score

244 W La Revue canadienne de psychiatrie, vol 52, no 4, avril 2007

 Immediate and Sustained Psychological Impact of an Emerging Infectious Disease Outbreak on Health Care Workers

14.81, SD 5.02) (F1,175 = 18.33, P < 0.001) . Among high-risk mediation effect was significant (z = 3.61, P < 0.001). All 3 of
health care workers, perceived stress was not associated with Baron and Kenny’s criteria were satisfied, indicating that
age or education level but was higher among men (mean score posttraumatic stress owing to the impact of the SARS out-
20.39, SD 5.62) than women (mean score 17.63, SD 4.75) break could account for the difference in perceived stress
(t67 = 2.25, P < 0.05). In low-risk health care workers, per- between high-risk and low-risk health care workers. How-
ceived stress was not associated with sex or education level ever, since risk group was still a significant predictor of per-
but was significantly correlated with age (Spearman’s ceived stress even after the IES-R scores were entered into the
rho = –0.26, P < 0.01, Bonferroni-corrected). Among both the final regression, posttraumatic stress owing to the impact of
high- and low-risk groups, there were no differences in per- the SARS outbreak was only a partial mediator; the difference
ceived stress levels among doctors, nurses, and other health in perceived stress between high-risk and low-risk health care
care workers. workers could have been influenced by other factors that were
In 2004, the stress response of the high-risk health care work- not included in the present study.
ers was characterized in more detail. This revealed substan-
tially higher DASS-21 Depression subscale scores (mean Discussion
4.89, SD 4.75) and DASS-21 Anxiety subscale scores (mean This is the first study to report that, despite similar perceived
4.85, SD 4.25) for high-risk, compared with low-risk, health stress levels between high-risk and low-risk health care work-
care workers (mean DASS-21 Depression subscale score ers in 2003, high-risk health care workers remained highly
2.17, SD 2.57; mean DASS-21 Anxiety subscale score 2.07, stressed 1 year later. There was a significant relation between
SD 2.50) (F1,178 = 21.31, P < 0.001 for the Depression the latter and higher DASS-21 Depression scores, DASS-21
subscale; F1,176 = 21.54, P < 0.001 for the Anxiety subscale). Anxiety scores, and IES-R score for posttraumatic stress asso-
Generally, PSS-10 scores were significantly and positively ciated with risk of exposure to SARS.
correlated with the DASS-21 Depression subscale
(Spearman’s rho = 0.67, P < 0.001), the DASS-21 Anxiety The SARS outbreak in 2003 imposed particular constraints on
subscale (Spearman’s rho = 0.62, P < 0.001), and the this type of research. Subjects worked in isolation wards; thus,
DASS-21 total scores (Spearman’s rho = 0.72, P < 0.001, all the questionnaire had to be brief, self-administered, and anon-
Bonferroni-corrected); thus, perceived stress was associated ymous.10 This design encouraged participation and mini-
with higher levels of depression, anxiety, and general mized stigma. 1 6 We did not formally evaluate the
psychological distress. psychometric properties of psychological responses to the
outbreak10 because it was soon contained by mass quarantine.
Is Higher Stress Partly Mediated by Contact With SARS? Health care workers reported more positive than negative
In an attempt to explain the persistently elevated stress levels responses to the outbreak, but we do not believe that this is
in high-risk health care workers, we examined the possible likely to be a result of questionnaire artifact. Although, com-
contribution of contact with SARS. We tested a mediation pared with low-risk control subjects, significantly more
model (see Figure 1) in which the difference in perceived high-risk health care workers reported fatigue, sleep depriva-
stress between the high- and low-risk health care workers was tion, worry about health, and fear of social contact, they
mediated by the IES-R13 scores measuring posttraumatic expressed more confidence in infection-control practices. We
stress owing to the impact of the SARS outbreak. We per- believe this was instrumental in their coping with a novel
formed a series of hierarchical multiple regression analyses pathogen because, during the outbreak, high- and low-risk
using Baron and Kenny’s14 criteria. In each of the analyses, health care workers had equivalent stress levels that were also
we entered age and education level into Block 1 as covariates. comparable to those of the local community.10
In the first regression analysis, IES-R scores were regressed
onto risk group. Risk group was found to be a significant pre- One year later, the persistence of elevated stress levels for
dictor of IES-R scores (â = 0.31, t = 4.08, P < 0.001, R2 = high-risk health care workers was unexpected. One explana-
0.116). PSS-10 scores were then regressed onto IES-R scores. tion is that the health care workers sampled in 2003 and 2004
IES-R was found to be a significant predictor of perceived were not exactly the same group. Although the anonymity of
stress (â = 0.54, t = 8.94, P < 0.001, R2 = 0.361). Finally, the the questionnaire made this kind of sampling unavoidable, the
effect of risk group on perceived stress was reduced when rate of questionnaire return was fair and suggests considerable
both risk group and IES-R were entered into the regression overlap between the groups sampled at each time point. There
analyses (from â = 0.31, t = 4.32, P < 0.0001 when only risk were some differences in education level and age, which we
group was entered into the regression equation to â = 0.20, t = attempted to control for in our analysis. However we did not
2.67, P < 0.01 when both risk group and IES-R scores were previously find any relation between education level and per-
entered). The result of the Sobel test showed that the ceived stress among health care workers.10 We therefore

The Canadian Journal of Psychiatry, Vol 52, No 4, April 2007 W 245

Original Research

believe that the persistence of elevated stress levels among individuals could benefit from routine psychological training,
high-risk health care workers is real. scheduled rest periods, flexible staffing resources, and even
pandemic rehearsal.
Compared with the low-risk health care workers, the elevated
stress levels among high-risk health care workers in 2004 was
Conclusion
consistent with a substantially higher DASS-21 Anxiety
Our data suggest that stress persists 1 year after an epidemic is
score, DASS-21 Depression score, and IES-R posttraumatic
over. This indicates that stress management for front-line
contact with SARS score. Our study included a control health
health care workers is integral to a protocol for outbreak pre-
care worker group, making it unlikely that the elevated stress
paredness. This would enhance infection-control measures
was attributable to the caregiver role.17 Alternative explana-
and patient care in the face of a future pandemic.
tions for this pattern of results include health care workers’
ongoing adverse experiences during SARS, both locally18 and
Funding and Support
in Toronto, Ontario.19 Our study suggests that the experience
This study was supported by a small project fund from the
of working in proximity with SARS did confer a persistent University of Hong Kong.
detrimental effect, as posttraumatic stress associated with
high risk of exposure to SARS was found to be at least par- Acknowledgment
tially responsible for the persistence of elevated stress levels We thank our colleagues at Queen Mary Hospital and United
in 2004. We noted that high- and low-risk health care workers Christian Hospital for their assistance. We also thank Professor
had similar stress levels in 2003 but that, 1 year later, the stress Peter Lee for his advice on rendering support to health care
workers.
levels of high-risk health care workers significantly exceeded
those of their low-risk counterparts. One possible explanation
is that high-risk health care workers might have displayed References
“denial” during the outbreak, causing their psychological dis- 1. World Health Organisation. Communicable disease surveillance and response.
tress to be minimized during that time. This is plausible, given Avian influenza A(H5N1) in humans and poultry in Viet Nam. 2004. [cited 2004
Jan 13]. Available from: http://www.who.int/csr/don/2004_01_14/en/.
their high rate of confidence in infection-control practices, 2. Shortridge KF, Peiris JS, Guan Y. The next influenza pandemic: lessons from
compared with that of their low-risk colleagues. Although it is Hong Kong. J Appl Microbiol. 2003;94(Suppl):70S–79S.
3. Peiris JS, Guan Y, Markwell D, et al. Cocirculation of avian H9N2 and
true that they received more training in infection control, their contemporary “human” H3N2 influenza A viruses in pigs in southeastern China:
confidence was arguably disproportionate to the real threat potential for genetic reassortment? J Virol. 2001;75(20):9679–9686.
4. Kobasa D, Takada A, Shinya K, et al. Enhanced virulence of influenza A viruses
that they were facing, to the efficiency with which SARS is with the haemagglutinin of the 1918 pandemic virus. Nature.
2004;431(7009):703–707.
transmitted, and to the large number of health care workers
5. Manning A, Schmit J. Samples of pandemic flu virus found in 3 foreign
infected while caring for SARS patients. In other words, con- warehouses of shipper. USA Today. 2005;Apr 17.
6. Ungchusak K, Auewarakul P, Dowell SF, et al. Probable person-to-person
fidence in infection control possibly diminished the collective transmission of avian influenza A (H5N1). N Engl J Med. 2005;352(4):333–340.
sense of threat and thus conferred a protective function and 7. Grais RF, Ellis JH, Glass GE. Assessing the impact of airline travel on the
geographic spread of pandemic influenza. Eur J Epidemiol. 2003;18:1065–1072.
minimized stress levels. A year after the outbreak, stress lev- 8. Zambon M, Nicholson KG. Sudden acute respiratory syndrome may be a
els decreased in low-risk health care workers but increased in rehearsal for the next influenza pandemic. BMJ. 2003;326:669– 670.
9. Mack TM. The ghost of pandemics past. Lancet. 2005;365(9468):1370 –1372.
high-risk health care workers. This decompensation in 10. Chua SE, Cheung V, Cheung C, et al. Psychological effects of the SARS
high-risk health care workers could be a rebound effect as the outbreak in Hong Kong on high-risk health care workers. Can J Psychiatry.
2004;49(6):391–393.
imminent threat ended and denial was lessened. This is also 11. Cohen S, Williamson G. Perceived stress in a probability sample of the United
possibly the result of a sense of frustration arising from the States. In Spacapam S and Oskamp S, editors. The social psychology of health:
Claremont Symposium on applied social psychology. Newbury Park.(CA): Sage;
high-risk group’s not having received particular recognition 1988. Chapter 3. p. 31–67.
for their contribution, or it could even be the result of 12. Taouk M, Lovibond PF, Laube R. Psychometric properties of a Chinese version
of the short Depression Anxiety Stress Scales (DASS21). Sydney (AU): Report
anticipating new outbreaks to combat after SARS. for New South Wales Transcultural Mental Health Centre, Cumberland Hospital,
Sydney; 2001.
Whatever the reason for the reported high stress levels among 13. Horowitz MJ, Wilner NR, Alvarez W. Impact of Event Scale. A measure of
subjective stress. Psychosom Med. 1979;41:209–218.
high-risk health care workers in 2004, our findings are a cause 14. Baron RM, Kenny DA. The moderator-mediator variable distinction in social
for concern. Chronic stress may erode immunologic media- psychological research: Conceptual, strategic, and statistical considerations.
J Personality Soc Psy. 1986;51:1173–1180.
tors,17 increase risk of influenza and cytokine production,20 15. Chua SE, Cheung V, McAlonan GM, et al. Stress and psychological impact on
and decrease antibodies.21 We do not know to what extent SARS patients during the outbreak. Can J Psychiatry. 2004;49(6):385–390.
16. Bai Y, Lin CC, Lin CY, et al. Survey of stress reactions among health care
such physical ramifications affected our health care workers. workers involved with the SARS outbreak. Psychiatr Serv.
Stress can also impair cognitive functioning and task perfor- 2004;55(9):1055–1057.
17. Segerstrom SC, Miller GE. Psychological stress and the human immune system:
mance.22 Whether the chronic stress endured by front-line a meta-analytic study of 30 years of inquiry. Psychol Bull. 2004;130(4):601–630.
health care workers might impair their effectiveness in a 18. Tam CW, Pang EP, Lam LC, et al. Severe acute respiratory syndrome (SARS) in
Hong Kong in 2003: stress and psychological impact among frontline health care
future pandemic is of concern. It is speculated that such workers. Psychol Med. 2004;34(7):1197–1204.

246 W La Revue canadienne de psychiatrie, vol 52, no 4, avril 2007

 Immediate and Sustained Psychological Impact of an Emerging Infectious Disease Outbreak on Health Care Workers

19. Maunder R. The experience of the 2003 SARS outbreak as a traumatic stress
among frontline health care workers in Toronto: lessons learned. Philos Trans R
Soc Lond B Biol Sci. 2004;359(1447):1117–1125. Manuscript received August 2006, revised, and accepted October 2006.
1
Assistant Professor, Department of Psychiatry, University of Hong Kong,
Queen Mary Hospital, Pokfulam, Hong Kong SAR, China.
20. Cohen S, Doyle WJ, Skoner DP. Psychological stress, cytokine production, and 2
severity of upper respiratory illness. Psychosom Med. 1999;61(2):175–180.
PhD Candidate, Department of Psychiatry, University of Hong Kong
Queen Mary Hospital, Pokfulam, Hong Kong SAR, China.
3
Honorary Clinical Associate Professor, Department of Medicine,
21. Miller GE, Cohen S, Pressman S, et al. Psychological stress and antibody
response to influenza vaccination: when is the critical period for stress, and how
University of Hong Kong, Queen Mary Hospital, Pokfulam, Hong Kong
does it get inside the body? Psychosom Med. 2004;66(2):215–223. SAR, China.
4
Chair and Professor in Psychiatric Genomics, Department of Psychiatry,
University of Hong Kong, Queen Mary Hospital, Pokfulam, Hong Kong
22. Janis IL. Decision making under stress. In: Goldberger L and Breznitz S, editors.
Handbook of stress: theoretical and clinical aspects. New York (NY): Free Press; SAR, China.
1993. p 56–74. Address for correspondence: Antionette Lee, Department of Psychiatry,
University of Hong Kong, Queen Mary Hospital, Pokfulam, Hong Kong
SAR, China; [email protected]

Résumé : L’effet psychologique immédiat et prolongé du début d’épidémie d’une

maladie infectieuse sur les travailleurs de la santé
Objectif : Évaluer la santé psychologique immédiate et prolongée des travailleurs de la santé qui
étaient à haut risque d’exposition durant l’épidémie du syndrome respiratoire aigu sévère (SRAS).
Méthodes : Au plus fort de l’épidémie du SRAS en 2003, nous avons évalué les travailleurs de la
santé de 2 hôpitaux généraux de soins actifs de Hong Kong, à l’aide de l’échelle de stress perçu en
10 items (PSS-10). Un an plus tard, nous avons réévalué ces travailleurs avec la PSS-10, l’échelle
de dépression et d’anxiété en 21 items (DASS-21) et l’échelle des répercussions d’événements
révisée (IES-R). Nous avons recruté des travailleurs de la santé à risque élevé qui pratiquaient la
médecine respiratoire et nous les avons comparés avec des travailleurs de la médecine non
respiratoire, qui formaient le groupe témoin de travailleurs de la santé à faible risque.
Résultats : En 2003, les travailleurs de la santé à risque élevé avaient des niveaux de stress élevés
(score à la PSS-10 = 17,0), lesquels ne différaient pas significativement des niveaux du groupe
témoin de travailleurs de la santé à faible risque (score à la PSS-10 = 15,9). Un plus grand nombre
de travailleurs de la santé à risque élevé déclaraient de la fatigue, un mauvais sommeil, des
inquiétudes pour leur santé, et la crainte des contacts sociaux, malgré leur confiance dans les
mesures de contrôle de l’infection. En 2004, cependant, les niveaux de stress du groupe à risque
élevé étaient non seulement plus hauts (score à la PSS-10 = 18,6) mais aussi significativement plus
élevés que les scores du groupe témoin de travailleurs de la santé à faible risque (score à la
PSS-10 = 14,8, P < 0,05). En 2004, les niveaux de stress perçu du groupe à risque élevé étaient
associés avec des scores de stress plus élevés de dépression, d’anxiété et post-traumatique
(P < 0,001). Les scores de stress post-traumatique étaient un médiateur partiel de la relation entre le
risque élevé d’exposition au SRAS et un stress perçu élevé.
Conclusions : Les travailleurs de la santé qui étaient à risque élevé de contracter le SRAS semblent
avoir non seulement un stress chronique mais aussi des niveaux élevés de dépression et d’anxiété.
Les employés de première ligne pourraient bénéficier de gestion du stress dans le cadre de la
préparation à d’autres épidémies.

The Canadian Journal of Psychiatry, Vol 52, No 4, April 2007 W 247