Psychological impact of the pandemic (H1N1) 2009 on general hospital workers in Kobe

Authors


Kunitaka Matsuishi, MD, PhD, Department of Psychiatry, Kobe City Medical Center General Hospital, 2-1-1 Minatojima Minamimachi, Chuo-ku, Kobe, Hyogo 650-0047, Japan. Email: matuisi@kcho.jp

Abstract

Aims:  In order for hospitals to work efficiently in a pandemic, it is important to know how a pandemic affects the hospital staff. The aim of the present study was to investigate the psychological impact of the pandemic (H1N1) 2009 on hospital workers and how it was affected by the characteristics of the hospital, gender, age, job and work environment.

Methods:  In late June 2009, soon after the pandemic had ended in Kobe city, Japan, a questionnaire was distributed consisting of questions on sociodemographic characteristics, 19 stress-related questions and the Impact of Event Scale (IES) to all 3635 employees at three core general hospitals in Kobe. Exploratory factor analysis was applied to the 19 stress-related questions, and this produced four factors for evaluation (anxiety about infection, exhaustion, workload, and feeling of being protected). Multiple regression models were used to evaluate the association of personal characteristics with each score of the four factors and the IES.

Results:  Valid answers were received from 1625 employees. Workers at a hospital with intense liaison psychiatric services felt less psychological impact. Workers at a hospital that provided staff with information about the pandemic less frequently, felt unprotected. Workers in work environments that had a high risk of infection felt more anxious and more exhausted. The total IES score was higher in workers in high-risk work environments.

Conclusions:  It is important for hospitals to protect hospital workers during a pandemic and to rapidly share information about the pandemic. Liaison psychiatric services can help to reduce the impact of the pandemic on hospital workers.

IN APRIL 2009, H1N1 influenza was confirmed to pass from human to human in Mexico. On 11 June 2009, the World Health Organization declared the H1N1 influenza infection a pandemic. In Japan, the H1N1 influenza pandemic plunged the public into confusion and anxiety in May 2009. Every day, the media reported the number of infected patients in each prefecture.

On 16 May, Kobe City Medical Center General Hospital (hospital X) admitted the first patient to have been domestically infected with the H1N1 influenza virus in Japan. This attracted a great deal of public attention. On that day, hospital X, Kobe City Medical Center West hospital (hospital Y) and Nishi-Kobe Medical Center (hospital Z) started fever consultation centers that attended to patients suspected of having H1N1 influenza. In the following 2 weeks, 1831 people who suspected that they had H1N1 influenza infection came to these hospitals. Among them, 1687 were released as outpatients. The remaining 144 patients were suspected of having H1N1 and were admitted. Forty-nine of these patients were later diagnosed as having an H1N1 influenza infection. Although the number of outpatients was similar among the three hospitals, hospital X had 122 admissions in which H1N1 positivity was suspected (confirmed in 31), and hospital Z had 22 admissions (confirmed in 18). The peak occurred on 17 May when 211 patients came to the fever consultation centers. On 27 May, the mayor of Kobe City declared that the emergency had subsided. On 3 June, the fever consultation centers were closed, and by 8 June 2009, the hospitals had returned to normal practice. During these 2 weeks, the hospital workers would have experienced huge physical and psychological stress.

To clarify the impact of the influenza pandemic on hospital workers, we distributed questionnaires to hospital workers at the three hospitals (hospitals X, Y and Z) that compose the Kobe Municipal Hospital Group. Several studies have examined the stress that hospital workers experienced in the HIN1 and in other pandemics.1–9 But to our knowledge, no study has investigated the effects of sociodemographic characteristics including place of employment on the stress of hospital workers in a pandemic. There is a possibility that hospital workers' psychological response to pandemic is associated with work environment or characteristics of the hospital. In this study we investigated the psychological impact of the H1N1 influenza pandemic on hospital workers and how the impact was affected by the characteristics of the hospital, gender, age, job and work environment.

METHODS

Subjects

We conducted the study at hospitals X, Y and Z, which are tertiary teaching hospitals in Kobe City. All three hospitals accepted H1N1 influenza patients from 16 March 2009. The characteristics of the three hospitals are given in Table 1. Hospital X was the largest hospital and hospital Y was the smallest. Hospital X and hospital Z received both outpatients and inpatients of H1N1, while hospital Y received only outpatients. The number of outpatients was almost the same among the three hospitals, but the number of inpatients was much greater at hospital X than at hospital Z.

Table 1.  Hospital characteristics
 Hospital XHospital YHospital Z
  1. Information about the pandemic was provided frequently by email to all staff and via bulletin boards from the Infectious Control Team in hospital X and hospital Z, whereas in hospital Y, it was infrequently provided via hardcopy report to every section. Because there were staff in both the fever consultation center and the ward for H1N1 infection, the number of staff differed from that in the high-risk area.

No. beds912358500
No. staff16257751235
No. doctors26174124
Ward for infectious patientsYesNoYes
Type of service in the pandemic (H1N1) in 2009Hospitalization and ambulant careAmbulant careHospitalization and ambulant care
Total no. outpatients of the fever consultation center715564552
Total no. inpatients suspected of being H1N1 positive122022
Total no. staff in the fever consultation center23171112
Total no. staff in the ward for H1N1 infection85052
Support for staff within each hospitalFrequent provision of informationInfrequent provision of informationFrequent provision of information and liaison psychiatric service

We distributed a questionnaire to all 3635 staff at the three hospitals and received a total of 1995 valid responses (54.9%). Of these responses, 370 were excluded because they were missing at least one answer, leaving 1625 questionnaires (44.7%) for analysis. The number of valid questionnaires received at each of the three hospitals was as follows: hospital X, 894 (55.0%); hospital Y, 239 (14.7%); hospital Z, 492 (30.3%). The characteristics of the staff and the respondents are listed in Table 2.

Table 2.  Staff characteristics
 Total staff (n)Valid respondents (n)
  1. Jobs classified as medical doctor; nurse; or others (radiological technologists, clinical laboratory technicians, pharmacists, dieticians, social workers, physical therapists, occupational therapists and speech therapists, office workers, clinical clerks, guards and janitors).

Gender  
 Male955397
 Female26801228
Age groups (years)  
 20–291529605
 30–39993453
 40–49566311
 50–59457222
 60–699034
Job  
 Medical doctor495218
 Nurse1839864
 Others1301543

Procedure

Paper-based, self-administered anonymous questionnaires were personally handed to all employees or placed in their mail boxes starting on 22 June 2009 and were collected from collection boxes in each hospital until 31 July 2009, which was approximately 1 month after the peak of the H1N1 outbreak in Kobe City.

Content of questionnaire

The questionnaire explained the purpose, which was to examine the stress that hospital workers experienced during the pandemic (H1N1) 2009 and stated that the results would be published, and respondents would remain anonymous. The first question asked for permission to use the responses in the survey. Answers without this approval were omitted from the analysis. The questionnaire consisted of questions on sociodemographic characteristics, stress-related questions associated with the H1N1 event and the Impact of Event Scale (IES).

The personal characteristics included gender, age group, job, the hospital and the work environment during the H1N1 pandemic. Work environment was categorized into high risk (the ward for H1N1 infection and the fever consultation center) and low risk (all other places). We were unable to determine how many workers in high-risk work environments actually came into contact with H1N1 patients, but all such workers could have come into contact with H1N1 patients and they recognized this.

The stress-related questions consisted of 19 items (Table 3).10,11 The respondents used a 4-point Likert scale (0, never; 1, rarely; 2, sometimes; 3, always) to describe how often they experienced the 19 items during the pandemic. Sixteen of the 19 items were based on the items in studies on severe acute respiratory syndrome (SARS)12,13 and hypothetical influenza pandemics.1–4 We added three original items about incentive to work during the pandemic (Q12, Q18 and Q19).

Table 3.  Factor analysis of the 19 stress-related questions
QuestionsF1F2F3F4h2
  1. Cronbach α was computed without excluded items.

  2. Bold, factor loading ≥0.40.

Factor 1: Anxiety about infection (Cronbach α = 0.81)
Q1Anxiety about being infected0.76−0.03−0.020.010.56
Q2Anxiety about infecting family0.70−0.010.030.040.50
Q5Anxiety of being infected during commuting0.68−0.01−0.110.050.42
Q7Lack of knowledge about infectivity and virulence0.67−0.030.010.020.44
Q6Lack of knowledge about prevention and protection from infection0.63−0.060.090.000.40
Q11Anxiety about compensation0.440.120.05−0.090.29
Q12Hesitation to work0.360.26−0.10−0.190.26
Q8Feeling of being avoided by others0.320.130.030.000.17
Q19Feeling of having no choice but to work due to obligation0.190.130.130.100.15
Factor 2: Exhaustion (Cronbach α = 0.80)
Q16Physical exhaustion−0.100.990.00−0.030.88
Q17Mental exhaustion0.000.880.020.000.78
Q15Insomnia0.090.50−0.040.070.28
Q14Elevated mood0.040.410.020.150.23
Q13Feeling of being isolated0.220.350.03−0.020.26
Factor 3: Workload (Cronbach α = 0.83)
Q4Burden of change of quality of work0.02−0.061.00−0.030.93
Q3Burden of increase quantity of work−0.030.230.610.000.58
Factor4: Feeling of being protected (Cronbach α = 0.69)
Q10Feeling of being protected by hospital−0.020.010.020.780.62
Q9Feeling of being protected by national and local governments0.030.01−0.080.660.44
Q18Motivation to work0.000.200.040.380.21
Eigenvalue3.723.802.791.31 
Variance Explained (%)   44.09 
Between factor correlation     
F20.44    
F30.330.58   
F40.040.050.05  

The IES was developed by Horowitz et al. as a measure of psychological stress reactions after trauma,14 and has been widely used in related research for >20 years.15,16 The IES consists of 15 items that evaluate the psychological response to a stressful event, and respondents use a 4-point Likert scale (0, not at all; 1, rarely; 3, sometimes; 5, often). The validity and reliability of the Japanese version of the IES-Revised (IES-R), which included all Japanese items of IES used in this study, have been reported.17 Because the IES was found to be useful for evaluating nurses' anxiety about becoming infected with SARS,15 we believed it would be effective for determining the psychological impact of the H1N1 pandemic on hospital workers.

The Cronbach alpha coefficient of the 16 stress-related questions without three items (Q12, Q18 and Q19) was α = 0.84 (n = 1625), indicating good internal consistency and acceptable reliability. Validity of the 16 items was confirmed because there was a modest correlation between the total score of the 16 items and the IES score (r = 0.35, r2 = 0.12, P < 0.001, n = 1625, Pearson product-moment correlation).

The survey was approved by the Kobe City Medical Center General Hospital Ethical Review Board and participation was voluntary.

Statistical analysis

We applied factor analysis to the 19 stress-related questions with the maximum likelihood method and promax rotation because factor structures had not yet been formulated according to previous studies. The number of factors was determined by the size of the eigenvalue and the relative size of the values according to different factor models.

For each of the factors, the total score of the stress-related questions was calculated. Multiple regression models with dummy coded demographic variables (gender, age group, job, hospital and work environment) were used to compute standardized partial regression coefficients (β) to evaluate the effects of personal characteristics on the total score of each factor and the IES (statistical significance: two-tailed P < 0.05 in the present study).

Statistical analysis was carried out using spss (17.0 J: SPSS, Tokyo, Japan).

RESULTS

Factor analysis

We applied exploratory factor analysis to the 19 stress-related questions. Four factors were extracted from 14 items having factor loadings of ≥0.40 (Table 3). Factor 1 included six items of the 14 items and was labeled ‘anxiety about infection’. Factor 2 included four items and was labeled ‘exhaustion’. Factor 3 included two items and was labeled ‘workload’. Factor 4 included two items and was labeled ‘feeling of being protected’.

Multiple regression analysis

Table 4 lists the estimated associations between the sociodemographic characteristics with the total score in each of the four factors and the IES. All regression models were significant.

Table 4.  Total scores and multiple regression analysis
Variablesn (%)Factor 1: Anxiety about infectionFactor 2: ExhaustionFactor 3: Workload§Factor 4: Feeling of being protectedIES††
Total scoreβTotal scoreβTotal scoreβTotal scoreβTotal scoreβ
(mean ± SD)(mean ± SD)(mean ± SD)(mean ± SD)(mean ± SD)
  1. Adjusted for gender, age group, job, hospital, work environment.

  2. *P < 0.05; **P < 0.01; ***P < 0.001.

  3. R2 = 0.07, adjusted R2 = 0.06; R2 = 0.07, adjusted R2 = 0.07; §R2 = 0.14, adjusted R2 = 0.13; R2 = 0.03, adjusted R2 = 0.03; ††R2 = 0.02, adjusted R2 = 0.01.

  4. β, standardized partial regression coefficient; IES, Impact of Event Scale; MD, medical doctor.

Gender
 Male397 (24.4)7.7 ± 3.43.2 ± 2.62.7 ± 1.71.6 ± 1.22.6 ± 7.2
 Female1228 (75.6)8.6 ± 3.40.043.3 ± 2.50.042.4 ± 1.6−0.06*1.7 ± 1.10.042.4 ± 6.5−0.02
Age group (years)
 20–29605 (37.2)8.8 ± 3.63.2 ± 2.52.6 ± 1.71.5 ± 1.02.4 ± 6.2
 30–39453 (27.9)8.7 ± 3.3−0.013.3 ± 2.60.052.4 ± 1.60.001.6 ± 1.00.042.2 ± 6.2−0.01
 40–49311 (19.1)7.7 ± 3.1−0.13***3.3 ± 2.60.032.4 ± 1.7−0.021.8 ± 1.10.10***2.5 ± 6.30.01
 50–59222 (13.7)7.5 ± 3.1−0.14***3.6 ± 2.40.08**2.6 ± 1.60.042.0 ± 1.10.14***3.5 ± 8090.05
 60–7034 (2.1)7.4 ± 2.6−0.05*2.9 ± 2.60.012.5 ± 1.50.011.9 ± 1.30.06*1.9 ± 2.7−0.01
Job
 MD218 (13.4)7.0 ± 3.43.2 ± 2.62.7 ± 1.81.4 ± 1.21.6 ± 4.5
 Nurse864 (53.2)8.7 ± 3.30.27***3.4 ± 2.50.14**2.6 ± 1.60.21***1.6 ± 1.00.10*2.6 ± 6.80.14*
 Others543 (33.4)8.3 ± 3.30.27***3.1 ± 2.40.082.5 ± 1.50.071.8 ± 1.10.13**2.7 ± 7.00.13**
Hospital
 X894 (55.0)8.3 ± 3.53.5 ± 2.52.6 ± 1.71.7 ± 1.12.8 ± 7.2
 Y239 (14.7)8.6 ± 3.20.042.8 ± 2.5−0.12***2.2 ± 1.6−0.09***1.5 ± 1.0−0.06*2.4 ± 7.0−0.04
 Z492 (30.3)8.3 ± 3.30.013.1 ± 2.5−0.07**2.4 ± 1.60.001.7 ± 1.1−0.012.0 ± 5.2−0.06*
Work environment
 Low risk1105 (68.0)8.2 ± 3.42.9 ± 2.32.1 ± 1.51.7 ± 1.12.2 ± 5.9
 High risk520 (32.0)8.6 ± 3.40.13***4.0 ± 2.60.25***3.2 ± 1.60.36***1.6 ± 1.10.033.1 ± 8.00.09**

For factor 1, ‘anxiety about infection’, workers in their 20s had more anxiety than workers in their 40s, 50s and 60s (40s: partial regression coefficient (B) = −1.08, SE = 0.24, β = −0.13, P < 0.001; 50s: B = −1.42, SE = 0.27, β = −0.14, P < 0.001; 60s: B = −1.28, SE = 0.60, β = −0.05, P = 0.031). In regard to job, nurses and others had stronger anxiety about infection than medical doctors (MDs; nurses: B = 0.16, SE = 0.16, β = 0.27, P < 0.001; others: B = 0.65, SE = 0.10, β = 0.27, P < 0.001). Workers in high-risk work environments had stronger anxiety than workers in low-risk work environments (B = 0.94, SE = 0.19, β = 0.13, P < 0.001).

For factor 2, ‘exhaustion’, workers in their 50s felt more exhaustion as compared with workers in their 20s (B = 0.60, SE = 0.20, β = 0.08, P = 0.003). In regard to job, nurses felt more exhaustion than MDs (B = 0.34, SE = 0.12, β = 0.14, P = 0.004). Workers at hospital Y and hospital Z had less exhaustion than workers at hospital X (hospital Y: B = −0.81, SE = 0.18, β = −0.12, P < 0.001; hospital Z: B = −0.37, SE = 0.14, β = −0.07, P = 0.008). Workers in high-risk work environments felt more exhaustion than workers in low-risk work environments (B = 1.35, SE = 0.14, β = 0.25, P < 0.001).

For factor 3, ‘workload’, nurses answered that they had a higher workload than MDs (B = 0.34, SE = 0.07, β = 0.21, P < 0.001). Workers at hospital Y felt that they had a lower workload than workers at hospital X (B = −0.39, SE = 0.11, β = −0.09, P < 0.001). Workers in high-risk work environments felt that they had a higher workload than workers in low-risk work environments (B = 1.24, SE = 0.09, β = 0.36, P < 0.001).

For factor 4, ‘feeling of being protected’, workers in their 40s, 50s and 60s had a stronger feeling of being protected than workers in their 20s (40s: B = 0.27, SE = 0.08, β = 0.10, P < 0.001; 50s: B = 0.44, SE = 0.09, β = 0.14, P < 0.001; 60s: B = 0.45, SE = 0.20, β = 0.06, P = 0.021). In regard to job, nurses and others had a stronger feeling of being protected than MDs (nurses: B = 0.11, SE = 0.05, β = 0.10, P = 0.043; others: B = 0.10, SE = 0.03, β = 0.13, P = 0.002). Workers at hospital Y had a weaker feeling of being protected than workers at hospital X (B = −0.18, SE = 0.08, β = −0.06, P = 0.028).

Total IES score was affected by job, hospital and work environment. In regard to job, the total IES score of nurses and others was higher than that of MDs (nurses: B = 0.90, SE = 0.32, β = 0.14, P = 0.005; others: B = 0.60, SE = 0.20, β = 0.13, P = 0.002). The total IES score of workers at hospital Z was lower than that of workers at hospital X (B = −0.82, SE = 0.39, β = −0.06, P = 0.035). Workers in high-risk work environments had higher total IES scores than did workers in low-risk work environments (B = 1.24, SE = 0.38, β = 0.09, P = 0.001). The mean total IES score was 2.49 ± 6.63 and ranged from 0 to 73.

DISCUSSION

In the present study, none of the four stress-related factors except workload was significantly different between men and women. Also the IES scores were not significantly different between genders. Similar results were reported for studies of a hypothetical influenza pandemic in the USA1 and a hypothetical SARS pandemic in Singapore.18 These studies reported that stress in the face of a pandemic was not significantly affected by age, whereas in the present study ‘anxiety about infection’ was stronger among workers in their 20s than among those in their 40s, 50s and 60s. This may be because it was commonly known that younger people were more likely to be infected in pandemic (H1N1) 2009.19 The fact that hospital workers in their 50s were more exhausted than those in their 20s could be due to declining physical strength with age. Because middle-aged workers had gained various experience and a strong sense of belonging to the organization, which is a characteristic in Japanese workers, ‘feeling of being protected’ appeared to be stronger in middle-aged workers (those in their 40s and 50s) than in workers in their 20s. Most of the workers in their 60s would be hospital managers and would be responsible for providing protection for their employees.

With regard to job, nurses and others were significantly more anxious about infection. ‘Exhaustion’ and ‘workload’ were significantly stronger in nurses than in MDs. Moreover the total IES score was significantly higher in nurses and others than in MDs. Similar results were reported in a study of the 2003 SARS outbreak20 and a study of pandemic (H1N1) 2009 in Greece.21 The amount of time spent with infectious patients may underlie the difference in the effect of job. Although information on the H1N1 pandemic was sent to all of the hospital staff via email at both hospital X and Z, this information was posted on the bulletin board only in the doctors' lounges. At hospital Y, the hardcopy information about infection was distributed to only the MDs and the managers of each section. Thus at each of the three hospital, nurses and others might have been informed about infection less successfully than MDs. These results suggest that sharing exact information about infection as rapidly as possible is essential to reduce the stress and the impact and to provide a favorable work environment.

During the pandemic (H1N1) 2009, the three hospitals had approximately equal numbers of outpatients, but the number of inpatients was considerably larger in hospital X than in hospital Z, while hospital Y had no inpatients. The total number of staff in the fever consultation center was 231 at hospital X, 71 at hospital Y and 112 at hospital Z (Table 1). The number of outpatients per staff member was 3.1 at hospital X, 7.9 at hospital Y and 4.9 at hospital Z. The total number of staff in the ward for H1N1 influenza infection was 85 at hospital X and 52 at hospital Z. The number of inpatients per staff member was 1.4 at hospital X and 0.42 at hospital Z. Therefore the work for inpatients but not for outpatients may induce a significant difference in ‘exhaustion’ and ‘workload’ among the hospitals. Hospital X workers were probably more stressed and impacted on than workers at the other hospitals. The mean score of ‘feeling of being protected’ was lower in hospital Y than in hospital X, possibly because hospital X provided its workers with more information frequently about the pandemic. The staff at hospital X and hospital Z might have acquired information about H1N1 more easily than the staff at hospital Y. The total IES score was significantly lower at hospital Z than at hospital X. At hospitals X and Y, liaison psychiatric services were not provided for the staff in high-risk working environments, while at hospital Z, psychiatrists and a clinical psychologist visited the infectious inpatient ward every day to provide moral support to the staff. This action, which was recommended in a study of the 2003 SARS outbreak,16 could explain the lower total IES score at hospital Z.

Hospital workers in high-risk work environments felt significantly more ‘anxiety about infection’, ‘exhaustion’ and ‘workload’ and had significantly higher total IES scores than workers in low-risk work environments. A study of 1557 Toronto hospital workers in high- and low-risk areas in the 2003 SARS outbreak found a similar difference in IES scores.12 The mean total IES score of 2.49 in the present study, however, was much lower than the score of 16.84 in the Toronto study. This is probably due to the lower toxicity of the virus in the pandemic (H1N1) 2009. The H1N1 influenza pandemic in Japan had run its course within 1 month without heavy mortality. This may be why the impact on the hospital workers who attended H1N1 infectious patients was not so severe as the impact of the SARS outbreak. But irrespective of the viral toxicity, the fact that total IES score was significantly higher in workers in high-risk areas than in low-risk areas shows that workers who possibly come into contact with infectious patients may feel a psychological impact to some extent.

The ‘feeling of being protected’ was not significantly different between the work environments. The total score in ‘feeling of being protected’ was low for workers in both high- and low-risk areas. This could be because during the pandemic (H1N1) 2009 in Japan, especially in the latter half of May, national government and local governments tried to carry out infection control actively, but they did not provide hospital workers with information about protection against or compensation for H1N1 infection acquired while performing hospital duties. In preparation for a pandemic, some studies have emphasized that communities or employers should take all reasonable precautions to prevent illness among health-care workers and should provide them with reliable compensation if they become ill while carrying out their duties.21–24

The fact that there is no validated Japanese version of the stress-related questionnaire for SARS; the low return rate of 54.9%; and the fact that there were only valid answers for 44.7% of respondents are likely substantial limitations of this study. Because the adjusted R2 in Table 4 are low, future studies should look for other variables that better explain the psychological impact of a pandemic.

Conclusions

This study examined the stresses experienced by hospital workers responding to pandemic (H1N1) 2009 in Kobe City, Japan.

In order for hospital workers, especially those in high-risk areas, to manage a pandemic in a favorable work environment with minimal stress, it is essential for hospitals and governments to run public campaigns to protect their workers. In addition, frequent providing of information about the pandemic and liaison psychiatric services could help to reduce the stress and psychological impact of a pandemic like the pandemic (H1N1) 2009 on hospital workers.

ACKNOWLEDGMENTS

We thank the three hospitals for their assistance in this study. The authors declare that they have no competing interests and did not receive any financial support.

Ancillary