Study phase 2: incidence of new SARS-CoV-2 infections
after implementation of the vaccination programme
The incidence of SARS-CoV-2 infections was determined for vaccinated
(partially and fully) and unvaccinated HCWs. HCWs with unknown
vaccination status, previous SARS-CoV-2 infection, and/or without
follow-up data were excluded from the incidence analysis (Fig.3; study flowchart). After applying the exclusion criteria, the
incidence analysis included 1,461 HCWs with a median follow-up of 79
(IQR: 70–86) days.
In total, 93 of 1,461 (6.4%) HCWs tested positive for SARS-CoV-2 by the
end of August, 2021.
There was no significant difference between infected and uninfected HCWs
across demographic and occupational categories (Table S5 ). The
cumulative incidence of SARS-CoV-2 infection was significantly higher
among unvaccinated HCWs than in vaccinated HCWs (40.0% vs 3.2%; p
< 0.001). When vaccination groups were considered, the
cumulative incidence of new SARS-CoV-2 cases was 12.2% and 2.9% in
partially and fully vaccinated HCWs, respectively (Fig. 4 ). The
median interval from the first vaccine dose to SARS-CoV-2 infection was
28 (IQR: 21–31, range: 18–73) days in partially vaccinated HCWs, while
in the fully vaccinated group, the median interval between the second
vaccine dose and infection was 62 (IQR: 49–76, range: 25–200) days.
In a multivariate logistic regression model with confirmed SARS-CoV-2
cases as a response variable, vaccination status was the only
significant predictor for SARS-CoV-2 infection after adjusting for age,
gender, and number of PCR tests per HCWs (Table S7 ).
Vaccine effectiveness in preventing any SARS-CoV-2 infection was 79%
(95% CI 46–92%) and 95% (95% CI 91–97%) in partially and fully
vaccinated HCWs, respectively.
Vaccine effectiveness (VE) by age group and professional categories is
presented in Table 4 . VE for fully vaccinated HCWs did not vary
by age (age group < 50 years: adjusted VE (aVE) = 95%, 95%
CI 90–97%; 50+ years: aVE = 95%, 95% CI 87–98%) and was slightly
lower for the other without direct patient contact category (aVE = 89%,
95% CI 73–96%) when compared with physicians (aVE = 96%, 95% CI
84–99%) and nurses (aVE = 98%, 95% CI 92–99%).
DISCUSSION
By using longitudinal data, our study provides robust data on the
incidence of new SARS-CoV-2 infection among HCWs in a paediatric
hospital during the pre- and post-vaccinations periods, encompassing the
second and third wave of the COVID-19 pandemic in Poland.
During the ten months of the study period, approximately one fifth
(19.4%) of the HCWs susceptible to primary infection became infected
with SARS-CoV-2.
It is worth noting that almost half of the HCWs who had laboratory
confirmed SARS-CoV-2 infection in our study did not report symptoms,
suggesting that without the implementation of proactive universal
screening a significant proportion of infection among HCWs would have
remained undetected and that they would likely have continued working
while unaware of their status, therefore presenting a risk of
transmission to patients and co-workers. Although this study was not
designed to address the issue of nosocomial transmission reduction
through asymptomatic testing, we assume that the universal screening
programme of HCWs applied in our hospital (as part of a bundle of
intensified infection control measures) reduced the number of SARS-CoV-2
infections by early identification and isolation of SARS-CoV-2–positive
individuals. There are also some issues associated with massive
asymptomatic testing that must be underlined. First, it requires
additional operational and testing costs and a large laboratory capacity
to provide rapid turnaround for testing. Second, the results of the
PCR-based screening are valid only for the day of the test, which can
cause a false sense of confidence13. Furthermore, the
resources required to identify a single asymptomatic case are
substantial and may not be cost-effective during a low prevalence
period14. Thus, testing strategies should be
implemented after careful consideration of resources, infrastructure
capacity, and logistical issues15. In addition, the
testing strategy must be guided by the local epidemiology, vaccination
coverage, and efficacy among HCWs and in the community.
In line with some other studies conducted after the first wave of the
pandemic, the changes in the incidence rate of SARS-CoV-2 infection
among HCWs in our cohort were closely followed by community infection
rates8,16,17. Especially in the period before
vaccination (study phase 1), both the dynamic and the magnitude of new
SARS-CoV-2 infections were similar. On the other hand, after deployment
of the HCW vaccination programme, a reduction in the weekly incidence of
new cases among HCWs was observed. This downward trend initially was in
parallel with the incidence decline in the community but persisted up to
epi week 11, 2021, while in the general population a rapid increase in
SARS-CoV-2 incidence was observed attributed to the SARS-CoV-2 Alpha
variant domination (corresponding to the third pandemic wave). The
incidence rate among HCWs also increased afterwards, but the observed
peak was relatively lower. This finding is consistent with the
anticipated protective effect of the COVID-19 vaccine.
When considering the two phases of the present study, the cumulative
incidence of new SARS-CoV-2 infections decreased from 16.6% before
vaccination to 6.3% after the implementation of the vaccination
programme at the CMHI (study phase 2).
Due to the observational design of our study, we were unable to assess
the extent to which HCW vaccination contributed to the observed
reduction in the incidence of HCW SARS-CoV-2 infections. However, as the
uptake of vaccination among HCWs at the CMHI was high (88.9%), we
hypothesise that this impact could be significant, especially when the
short-term effect is considered.
In addition, this hypothesis was supported by the high vaccine
effectiveness observed in our cohort. Our findings indicate that the
effectiveness of the vaccine was 95% and 79% in fully and partially
vaccinated HCWs, respectively. These results are comparable to previous
reports including those from clinical trials and other real-world
studies6,7,18. In line with a previous study conducted
among HCWs, we did not observe significant differences in vaccine
effectiveness by age or occupational groups7,18,
although we observed slightly lower vaccine efficacy among HCWs in the
‘other without direct patient contact’ group. This finding could reflect
lower awareness, behaviour change, and a misbelief that vaccination
allows infection control measures to be relaxed.
However, despite high vaccine efficacy and high vaccination coverage, we
still observed new SARS-CoV-2 infections among HCWs at a rate from 2.9%
in the fully vaccinated to 39.7% in those unvaccinated.
These numbers are higher than those reported previously. In a systematic
review of eighteen studies, the pooled proportion of SARS-CoV-2
infection was 1.3%, 2.3%, and 10.1% among fully vaccinated, partially
vaccinated, and unvaccinated HCWs, respectively19.
Variability in demographic and occupational factors, as well as case
definitions, could explain this difference. Furthermore, the fact that
we excluded HCWs with a previous SARS-CoV-2 infection from our analysis
may have affected the higher infection rates observed in our cohort (the
absence of the protective effect of post-infection immunity).
Although rare, these infections might diminish HCWs’ belief in
vaccination effectiveness, especially in those becoming infected with
SARS-CoV-2 after initial vaccination. It is worth noting that, at the
time of manuscript preparation, we faced a low rate (approximately 5%)
of second booster dose uptake among HCWs at the CMHI. This led to
concerns when considering waning vaccine immunity over time and the
emergence of new variants of concern (VOC).
A systematic review by Biswas et al. revealed that vaccine acceptance
varied widely between countries and ranged from 4.3% to
72%20. Individuals who were men, older, physicians,
or well educated had a lower hesitancy to receive the COVID-19 vaccine.
Other factors associated with higher vaccine acceptance were: higher
income, medical risk, chronic disease history, not being infected with
SARS-CoV-2 in the past, knowledge of COVID-19, and a belief that
vaccines may protect friends, family, and community
members20. Although the present study was not designed
to assess the attitude to vaccination, these findings are consistent
with our study. We observed that older individuals, physicians, those
working in clinical settings, and those previously uninfected were more
likely to be vaccinated. These observations might inform tailored
communication strategies to be implemented to increase the uptake rate
of COVID-19 vaccines among HCWs.
Reliable data on risk factors, SARS-CoV-2 incidence, and the proportion
of HCWs who remain naïve (had no history of SARS-CoV-2 infection and/or
unvaccinated) are crucial to inform infection control strategies.
Previous studies reported widely varying estimates of incidence and risk
factors for infection among HCW21, and only a few
studies assessed HCW infection and risk
longitudinally17.
When considering potential risk factors of SARS-CoV-2 infection among
HCWs in our study, we did not find a significant difference in the
infection rates among HCWs working in wards involved and not involved in
the treatment of COVID-19 patients, clinical or non-clinical areas, or
across different wards and these findings applied to both phases of the
study. Our findings are in contrast to some previous studies, especially
those studies performed during the early pandemic
stage22, which reported higher infection rates among
HCWs working in COVID-19 wards with direct patient contact, but our
findings are in line with the emerging literature pointing out that,
except for breaches in PPE, the main risks to HCWs come from outside of
work factors (in the community and
household)17,23–26. In our cohort, before
implementation of the vaccination programme (study phase 1), nurses had
the highest adjusted likelihood of being infected, while physicians had
a lower likelihood. Surprisingly, in the post-vaccination study period,
we did not find any association between risk of SARS-CoV-2 infection and
professional category. The only factor significantly associated with
SARS-CoV-2 infection in the multivariate analysis was vaccination
status. The increased risk of SARS-CoV-2 infection for nurses was
previously reported in many observational
studies27–29. The higher risk among nursing staff has
been consistently explained by their more frequent contact with and
longer contact times with COVID-19 patients when compared with
physicians28,30,31. However, in our study we did not
observe an increase in the number of cases on the wards involved in the
treatment of patients with COVID-19. Furthermore, only a small
percentage of HCWs (with available data on the possible source of
exposure) reported exposure at work, and even fewer had high-risk
contact with a COVID-19 patient. A recent case-control study from
Ireland, investigating the impact of demographic and work-related
factors on the risk of SARS-CoV-2 infection after in-work exposure to a
confirmed case of COVID-19, revealed that male sex, Eastern European
nationality, exposure location, PPE use, and vaccination status all
impact the likelihood of SARS-CoV-2 infection throughout the first,
second, and third pandemic waves32. In this study, no
individual job role was determined to have a consistently higher risk of
infection after documented nosocomial exposure. It is likely that the
increased risk of SARS-CoV-2 infection among nurses observed in our and
other studies may be an indication of the cumulative risk of certain
nursing staff roles, which have an increased intensity of contact with
patients and other HCWs over time, and this increases the probability of
infection. In addition, nurses make up a significant proportion of the
total hospital staff (approximately one-third in our cohort), thus
resulting in over-representation, which may influence the interpretation
of risk32.
Together, these observations highlight the need for awareness of
non-patient care exposure risk that contributes to infection in HCWs and
should be considered in infection control measures (i.e. universal
masking, reinforcement of hand hygiene, and distancing).
The limitations of our data should be borne in mind. First, the
retrospective and observational design of the study is subject to
missing and incomplete data and thus to unmeasured confounding bias.
Second, although our study cohort was stratified by different
occupational categories, there may still be different exposure risks
within the categories. Third, we were unable to adjust our analysis for
other potential confounders that may affect the risk of infection (such
as comorbidities, contact time with COVID-19 cases, movement of HCWs
between high- and low-risk wards, compliance with PPE, household size)
due to data limitations, and therefore the results may not generalise to
other settings with different characteristics. Fourth, the clinical data
regarding symptoms for each HCW were retrieved from an unstandardized
database generated by the ICP team for epidemiological purposes and were
not consistently available. In addition, we relied on self-reported
symptoms and community/household exposures. Exposures in the workplace
were revealed through epidemiological investigations, including contact
tracing, but are still subject to recall bias. Data on exposure was
missing for almost 40% of the HCWs, therefore it was unknown whether
SARS-CoV-2 PCR positive cases resulted from infections in the workplace
or were acquired in the community or household. On the other hand, the
exact source and direction of infection may only be inferred from
epidemiological data combined with viral genomic data, otherwise
assessing the source is often subjective. Fifth, the lack of detailed
information on SARS-CoV-2 exposure and the use of PPE, prevented us from
analysing the association of adherence to PPE over time and across
different occupational categories with the risk of infection. Six, we
attempted to include only new cases of SARS-CoV-2 infection, excluding
HCWs with records of a positive test by PCR or anti-nucleocapsid IgG.
However, as systematic testing by PCR was not performed before the study
period and not all HCWs had participated in a previous serology
screening programme, we cannot rule out that we included some previous
positive cases in our incidence analysis. Seventh, testing in study
phase 2 was less dense as the universal screening programme ended on May
6, 2021, thus HCWs with asymptomatic SARS-CoV-2 infection might be
underrepresented. Eight, although our estimated vaccine effectiveness
aligns with the results provided in other reports, this estimate is
based on a relatively short follow-up period. Finally, our study took
place prior to the emergence of the Delta variant of SARS-CoV-2, thus it
might not be generalizable to an epidemic situation with a domination of
another VOC (variant of concern), however, the evidence from the early
period of the pandemic may inform the infection control policy in a
vulnerable population with regard to other respiratory viruses.
To our knowledge, this is the only study of the SARS-CoV-2 infection
among HCWs in Poland and one of the few studies based on robust data
obtained mainly from proactive universal screening PCR testing. The size
and diversity of the cohort (all HCWs, including nurses, physicians, and
other personnel providing direct and indirect patient care, as well as
non-clinical staff) together with longitudinal data collected in the
period before and after vaccination implementation in HCWs, meant that
it was possible to examine risk factors of the new SARS-CoV-2 infection
by demographic, occupational, and vaccination status (which was
determined based on vaccination records instead of self-reporting, to
mitigate recall bias).
In summary, our analysis provides information on the incidence of
SARS-CoV-2 infection in the pre- and post-vaccination periods. We found
that after the implementation of vaccination, the risk of infection
among HCWs remained relatively high despite the high coverage of
vaccination and the high effectiveness of the vaccine. Although rare,
breakthrough infections are challenging, as they may pose a risk to the
vulnerable population. Furthermore, there are concerns about the
decrease in vaccine effectiveness over time and during the emergence of
new VOCs.
Thus, continued efforts to promote infection control measures and
vaccination (including booster vaccine doses) and distancing remain
necessary during the ongoing COVID-19 pandemic.