The analysis of factors affecting municipal employees’ willingness to report to work during an influenza pandemic by means of the extended parallel process model (EPPM)

A total of 1.566 employees answered the survey. The overall response rate was 26 %
of the total staff members. However, 44 (2.8 %) participants were excluded from further
analyses because of missing demographic data. The characteristics of the remaining
sample are presented in Table 2. On average the respondents were 45 years old (sd?=?9.6 years) and 763 (50.1 %) were
male. 1.035 (68 %) of the participating employees had a high school degree or above
and 595 (39.45 %) had an executive position. 595 (39.33 %) had a civil servant status.
Of those participants with a civil servant status 299 (50.5 %) had also an executive
position. 673 (44.33 %) of the respondents stated that they had daily contact with
citizens as part of their work.

Table 2. Study sample

Psychometric properties of the summary scales

As indicated by the Spearman-Brown reliability coefficients (Table 3) all EPPM summary scales have at least a sufficient internal consistency.

Table 3. Spearman-Brown reliability of the EPPM summary scales

Willingness to report to work

1358 participants answered the questions about their willingness to report to work
if required and 1.393 answered the question about their willingness to report to work
if asked but not required. As indicated by Table 4 about 11 % of the municipal administration workers stated that they would not, or
would rather not, report to work during an IP, even if they were required by their
department. If they were only asked, but not required to report to work, the percentage
of those who indicated that they would not report to work increased to 20 %.

Table 4. Willingness to report to work

Path model 1: Willingness to report to work if required by the department

Figure 1 shows the standardised path coefficients of the restricted path model for
the willingness to report to work if required by the agency.

Factors affecting the perceived danger of pandemic

As indicated by the standardised path coefficient (ß?=?0.053; 95 % CI?=??0.104 to
?0.002; p???0.050) staff members with daily customer contact assessed the danger of
an IP higher than those who had only occasional contact with customers. The R
²
reveals that 0.3 % of the variance in the assessment of danger was explained by the
model.

Factors affecting perceived personal risk

Staff members with daily customer contact assessed their own risk of becoming infected
or carrying the infection into the family higher than those with a low frequency of
contact (ß?=?0.223; 95 % CI?=?0.181 to 0.264; p???0.001). In addition, the perception
of risk increased with the assessment of the danger that a pandemic will break out
(ß?=?0.294, 95 % CI?=?0.252 to 0.336; p???0.001), but decreased with the estimation
of self-efficacy (ß?=??0.341, 95 % CI?=??0.382 to ?0.300; p???0.001) and with the
participants’ age (ß?=??0.147; 95 % CI?=??0.190 to ?0.104; p???0.001). Decomposition
of total effects (not shown in figure 1) indicated a significant negative indirect
effect from competence over efficacy to perceived risk (ß?=??0.261; 95 % CI?=??0.330
to ?0.192; p???0.001). As indicated by the R
²
the model explained 28.4 % of the variance in the perception of risk.

Factors affecting perceived role competences

Standardised path coefficients revealed that participants in an executive position
(ß?=?0.169; 95 % CI?=?0.119 to 0.218; p???0.001) and participants with a civil servant
status (ß?=?0.086; 95 % CI?=?0.038 to 0.135; p???0.001) assessed their competence
to fulfil their role requirements during a pandemic higher than those in lower positions
or those without a civil servant status. Male participants assessed their competence
higher than their female counterparts (ß?=?0.080; 95 % CI?=?0.031 to 0.130, p???0.001).
As indicated by the R
²
, 5.7 % of the variance in the assessment of role competence was explained by the
model.

Factors affecting self-efficacy expectations

As indicated by the standardised path coefficient (ß?=?0.790; 95 % CI?=?0.609 to 0.971;
p???0.001) self-efficacy expectations regarding the capability of performing effectively
in a pandemic situation were only affected by the assessment of role competence, which
explained 22.6 % of the variance.

Factors affecting perceived role importance

The assessment of the importance of participants’ professional role increased with
the estimation of their role competence (ß?=?0.329; 95 % CI?=?0.285 to 0.373; p???0.001).
Staff members with daily customer contacts assessed their role during a pandemic as
more important than those with occasional contact (ß?=?0.167; 95 % CI?=?0.122 to 0.212;
p???0.001). Participants in executive positions (ß?=?0.163; 95 % CI?=?0.116 to 0.211;
p???0.001) and those with a civil servant status (ß?=?0.108; 95 % CI?=?0.062 to 0.154,
p???0.001) assessed their role importance higher than those in lower positions and
those with an employee status. Older staff members assessed their professional role
as less important than their younger colleagues (ß?=??0.077; 95 % CI?=??0.124 to ?0.031,
p???0.001). 21.6 % of the variance in the assessment of role importance was explained
by the model.

Factors affecting the sense of duty

With increasing self-efficacy (ß?=?0.293; 95 % CI?=?0.249 to 0.337; p???0.001) and
an increasing estimation of their role importance (ß?=?0.351; 95 %?=?CI 0.307 to 0.394,
p???0.001), staff members also felt more responsible to fulfil their professional
tasks during a pandemic. Participants with a civil servant status (ß?=?0.083; 95 %?=?CI
0.039 to 0.127; p???0.001) and those at a higher age (ß?=?0.081; 95 % CI?=?0.037 to
0.125; p???0.001) felt more responsible to fulfil their tasks during a pandemic than
ordinary employees and younger persons. Decomposition of total effects (not shown
in figure 1) revealed indirect positive effects from competence over role importance
(ß?=?0.115; 95 % CI?=?0.094 to 0.137; p???0.001) and from competence over efficacy
(ß?=?0.231; 95 % CI?=?0.168 to 0.295; p???0.001) to sense of duty.

The model explained 28.1 % of the variance in the sense of duty.

Factors affecting the willingness to report to work

Staff members’ willingness to report to work if required increased with increasing
self-efficacy (ß?=?0.144; 95 % CI?=?0.087 to 0.201; p???0.001), sense of duty (ß?=?0.228;
95 % CI?=?0.172 to 0.283; p???0.001), and with increasing assessment of role importance
(ß?=?0.187; 95 % CI?=?0.133 to 0.240; p???0.001). However, it decreased with an increasing
perception of the risk to become infected or transmit the infection to a family member
(ß?=??0.110; 95 % CI?=??0.162 to ?0.059; p???0.001). While willingness to report to
work increased with the level of education (ß?=?0.092; 95 % CI?=?0.044 to 0.193; p???0.001),
it decreased with increasing age (ß?=??0.093; 95 % CI?=??0.141 to ?0.045; p???0.001).
Decomposition of total effects (not shown in figure 1) indicated negative indirect
effects of perceived danger over perceived risk (ß?=??0.032; 95 % CI?=??0.017 to ?0.048;
p???0.001). Positive indirect effects of competence on willingness to report to work
were identified over importance (ß?=?0.061; 95 %?=?CI 0.042 to 0.081; p???0.001),
efficacy (ß?=?0.114; 95 % CI?=?0.062 to 0.166; p???0.001), efficacy and perceived
risk (ß?=?0.030; 95 % CI?=?0.014 to 0.046; p???0.001), role importance and sense of
duty (ß?=?0.026; 95 % CI?=?0.018 to 0.034; p???0.001), and efficacy and sense of duty
(ß?=?0.053; 95 % CI?=?0.033 to 0.072; p???0.001). Significant indirect effects on
willingness to report to work were identified from age over perceived risk (ß?=?0.016;
95 % CI?=?0.007 to 0.025; p???0.001), role importance (ß?=??0.014; 95 % CI?=??0.024
to ?0.005; p?=?0.003), sense of duty (ß?=?0.018; 95 % CI?=?0.007 to 0.029; p???0.001),
and over importance and sense of duty (ß?=??0.006; 95 % CI?=??0.010 to ?0.002; p???0.010).
Minor significant indirect effects on willingness to report to work were also obtained
from gender over competence and role importance (ß?=?0.005; 95 % CI?=?0.002 to 0.008;
p???0.010), competence and efficacy (ß?=?0.009; 95 % CI 0.002 to 0.016; p???0.010),
competence, efficacy, and perceived risk (ß?=?0.002; 95 % CI 0.000 to 0.004, p???0.050),
competence, role importance, and sense of duty (ß?=?0.002; 95 % CI 0.001 to 0.004;
p???0.010), and competence, efficacy and sense of duty (ß?=?0.004; 95 % CI 0.001 to
0.007, p???0.010).

In total the model explained 22.6 % of the variance in participants’ willingness to
report to work if required by the employing department.

Path model 2: Willingness to report to work if requested but not required by the department

With few exceptions the path coefficients and the R
²
values of model 2 (figure 2) were largely identical to those of path model 1. As one
major difference the effect of perceived role competence on perceived self efficacy
was smaller in model 2 (ß?=?0.538; 95 % CI?=?0.503 to 0.573; p???0.001). Furthermore,
in contrast to the willingness to report to work if required by the department, the
willingness to report to work if only asked but not required was not affected by the
participants’ educational level and age. Instead the willingness to report to work
voluntarily was found to be more strongly negatively related to the perception of
the risk to get infected or to pass on the infection to a family member (ß?=??0.225;
95 % CI?=??0.271 to ?0.179 p???0.001). While the positive effect of the assessment
of the importance of the professional role (ß?=?0.100; 95 % CI?=?0.049 to 0.151; p???0.001)
was found to be lower than in model 1, the positive effects of competence (ß?=?0.073;
95 % CI 0.019 to 0.126; p???0.010), sense of duty (ß?=?0.278; 95 % CI?=?0.228 to 0.328;
p???0.001), and self-efficacy (ß?=?0.192; 95 % CI?=?0.134 to 0.250; p???0.001) were
stronger.

Decomposition of total effects (not shown in figure 2) revealed an indirect negative
effect from perceived danger to willingness to report to work over perceived risk
(ß?=??0.066; 95 % CI?=??0.083 to ?0.049; p???0.001). From competence indirect effects
on willingness to report to work were identified over role importance (ß?=?0.033;
95 % CI?=?0.016 to 0.051; p???0.001), efficacy (ß?=?0.103; 95 % CI?=?0.071 to 0.136;
p???0.001), efficacy and perceived risk (ß?=?0.041; 95 % CI?=?0.031 to 0.052; p???0.001),
role importance and sense of duty (ß?=?0.032; 95 % CI?=?0.024 to 0.041 p???0.001),
and efficacy and sense of duty (ß?=?0.044; 95 % CI 0.033 to 0.055 p???0.001). Indirect
effects of age were identified over perceived risk (ß?=?0.034; 95 % CI?=?0.022 to
0.046; p???0.001), role importance (ß?=??0.008; 95 % CI?=??0.014 to ?0.002; p???0.010),
sense of duty (ß?=?0.023; 95 % CI?=?0.010 to 0.035; p???0.001), and role importance
and sense of duty (ß?=??0.008; 95 % CI?=??0.012 to ?0.003; p???0.010). Significant
indirect effects of gender were found over role competence (ß?=?0.006; 95 % CI?=?0.000
to 0.012; p???0.050), role competence and role importance (ß?=?0.003; 95 % CI?=?0.000
to 0.005; p???0.050), role competence and efficacy (ß?=?0.008; 95 % CI?=?0.002 to
0.014; p???0.010), role competence, efficacy and perceived risk (ß?=?0.003; 95 % CI?=?0.001
to 0.006; p???0.010), role competence, role importance, and sense of duty (ß?=?0.003;
95 % CI?=?0.001 to 0.004; p???0.010), and role competence, efficacy, and sense of
duty (ß?=?0.004; 95 % CI?=?0.001 to 0.006; p???0.010).

The R
²
value indicated that with 33.8 % a larger proportion of the variance in the willingness
to report to work voluntarily was explained by the model than in the variance of the
willingness to report to work if required by the department.

Overall model fit

The parameters for the goodness of fit of the path models are presented in table 5. The results of the Chi
²
tests indicated a significant deviation of the empirical covariance structure from
the structure hypothezised by the models. However, due to the large sample size the
Chi
²
is not a good indicator of the model fit and the global fit parameters CFI and TLI
as well as the RMSEA and the SRMR are within the range of an acceptable fit for both
models. The comparison of the information criteria AIC and BIC indicated that there
is no difference in the fit of both models.

Table 5. Fit indices of the path models