Knowledge, attitudes, and perceptions of COVID-19 vaccine and refusal to receive COVID-19 vaccine among healthcare workers in northeastern Ethiopia

Background Major efforts are being made to control the spread and impacts of the coronavirus pandemic using vaccines. Ethiopia began on March 13, 2021, to vaccinate healthcare workers (HCWs) for COVID-19 with the AstraZeneca vaccine. However, willingness to be vaccinated depends to a large extent on factors beyond the availability of vaccines. This study aimed to determine the rate of intention to refuse COVID-19 vaccination and associated factors among HCWs in northeastern Ethiopia. northeastern, Ethiopia. Method An institution-based cross-sectional study was employed among 404 HCWs in Dessie City, northeastern Ethiopia in May, 2021. Data were collected, checked, coded, entered into EpiData Version 4.6 and exported to Statistical Package of Social Sciences (SPSS) Version 25.0 for cleaning and analysis. The dependent variable was refuse to receive COVID-19 vaccination and the independent variables included socio-demographic factors, knowledge, attitudes and perceptions. A Binary logistic regression model was used to determine the association between vaccine refusal and the independent variables. From bivariate analysis, variables with p-values < 0.25 were retained for multivariable analysis. From multivariable analysis, variables with adjusted odds ratio (AOR), p-values <0.05 at 95% confidence interval (CI) were declared as factors significantly associated with refusal to be vaccinated among HCWs in Dessie City, northeastern Ethiopia. Results The proportion of HCWs with overall good knowledge, good perception, and positive attitudes about COVID-19 vaccination were 62.5%, 60.5%, and 52.3%, respectively; 64.0% of the HCWs wanted to be vaccinated while 36.0% said that they would refuse to do so. Multivariable analysis identified negative attitudes (AOR: 3.057; 95%CI [1.860 - 5.026]) and poor perceptions (AOR: 4.73; 95%CI [2.911 - 7.684]) about COVID-19 vaccines were significantly associated with refusal to be vaccinated for COVID-19. Nearly half (46.9%) of the HCWs stated that vaccines could worsen any pre-existing medical conditions and 39.5% of them thought that vaccines could cause COVID-19 infections. Conclusion The willingness of HCWs to get vaccinated against COVID-19 was relatively high among HCWs. Negative attitudes and poor perceptions towards the anticipated COVID-19 vaccination were significant factors to refuse to be vaccinated. Our findings may provide information for the management authorities and stakeholders to promote and improve attitudes, knowledge and perceptions towards COVID-19 vaccination uptake among HCWs.

Adane et al. BMC Public Health (2022) 22:128 Background COVID-19 infection rates are accelerating in Africa, including in Ethiopia. By October 24, 2021, Ethiopia had reported 362,088 COVID-19 cases and 6,347 deaths [1]. Many countries, including Ethiopia have implemented various strategies to control COVID-19, including declaring a state of emergency, issuing restrictions on mass gathering, enforcing stay-at-home orders, and promoting the use of personal protective equipment [2]. As part of the worldwide strategy, efforts have been made to develop and distribute vaccines. The COVAX program, backed by the World Health Organization (WHO) and other multilateral bodies, aims to supply 600 million doses to Africa, enough to vaccinate at least 20% of the population. However, by April 2021, only 18 million doses, representing 2% of all vaccine doses administered globally, had been administered by 41 African countries [3]. An additional 400 million doses of the Johnson & Johnson vaccine are scheduled to be shipped to the African Union starting in the third quarter of 2021 [4]. The first large shipment (2.2 million doses) of the AstraZeneca vaccine was received by Ethiopia in early March 2021 through the COVAX program. The country expected to receive an additional 5.4 million doses by May 2021 [5]. In addition, the African Vaccine Acquisition Trust (AVAT) announced the first monthly shipment of 108,000 doses of Johnson & Johnson vaccine to Ethiopia [6]. By April 2021, Ethiopia had administered only 430,000 doses of COVID vaccines, covering 0.2% of the country's population with 2-shot vaccinations [7], but expected to cover 20% of the population by the end of 2021 [5,8]. As of October 26, 2021, while more than 3.84 billion people worldwide had received one dose of a COVID-19 vaccine (equal to about 50% of the world population), 24 African countries had vaccinated less than 3% of their populations and only 0.9% of Ethiopia's population had been fully vaccinated [9].
Ethiopia began on March 13, 2021, to vaccinate HCWs for COVID-19 with the AstraZeneca vaccine. Healthcare workers include medical doctors, laboratory technicians, nurses, midwives, pharmacists, radiographers, anesthesiologists, public health and environmental health officers, any any other professionals including non-medical auxiliary staff who work in a healthcare facility. Non-medical auxiliary staff are other than health professionals such as financial workers, human resource workers, janitors, card room and documentation workers, porters, guardians and others.
Elderly people above the age of 60 and those above 50 years who have chronic diseases are also prioritized by the Ethiopian Ministry of Health. The launch of the vaccine program was attended by the WHO country representative Boueri Hama Sambo, who urged "communities and community leaders to build trust" as well as "demand for the vaccine" and to "promptly address "misinformation" [8]. During that event, one health worker was not convinced, saying "I don't believe in this vaccine. The virus is a sign of God's wrath upon us, so I prefer to pray" [8]. Negative attitudes and mistrust towards COVID-19 vaccines are major barriers to increasing vaccine coverage worldwide [10].
Despite the scarcity of COVID-19 vaccines in Africa, several countries including Sierra Leone and Malawi have discarded large numbers of vaccines doses that expired before they were used due to low demand in the population, a situation also anticipated in Uganda [11]. In addition to the low accessibility of COVID-19 vaccines and hesitancy to be vaccinated for COVID-19, a preference for traditional medicines may constitute another barrier to vaccine acceptance in African countries, including Ethiopia. Indigenous medicines have been used to treat COVID-19 in Tanzania and Madagascar without meeting safety and efficacy standards [12]. The use of these plant products to treat COVID-19 carries safety risks and may hinder use of scientifically proven COVID-19 vaccinations. A study in Nepal revealed that the use of medicinal plants has increased since the outbreak of the COVID-19 pandemic [13]. The use of traditional medicine for COVID-19 and its impact on vaccination rates has not been investigated in Ethiopia. However, it is known that some plant extracts have traditionally been used to treat influenza [14], and people may prefer using them rather than accepting COVID-19 vaccines in the absence of diagnostic tests.
Willingness to receive a COVID-19 vaccine is a challenge in many countries [15]. COVID-19 vaccine acceptance rates in the general population were highest in Vietnam (98%), India (91%), China (91%), Denmark (87%), and South Korea (87%) and lowest in Serbia (38%), Croatia (41%), France (44%), Lebanon (44%), and Paraguay (51%) [16]. High hesitancy rates were also reported worldwide among HCWs, who play a central role in reducing the burden of the pandemic through their role be vaccinated. Our findings may provide information for the management authorities and stakeholders to promote and improve attitudes, knowledge and perceptions towards COVID-19 vaccination uptake among HCWs.
Keywords: attitudes, COVID-19 vaccine, healthcare workers, knowledge, perceptions, vaccine refusal, Ethiopia in modeling preventive behavior and administering vaccinations. A review of 35 studies revealed that vaccination hesitancy rates varied from 4.3% to 72% worldwide [17]. Major concerns of health workers were vaccine safety, efficacy, and potential side effects. Higher socioeconomic status, directly working with patients, perceived risk and fear of COVID-19, and a history of influenza vaccination were associated with higher vaccine uptake [17].
Reluctance of HCWs to accept COVID-19 vaccination may not only increase the risk of virus transmission to their patients but also reduce the likelihood that HCWs will encourage patients to be vaccinated [17]. There is no documented evidence regarding attitudes and perceptions of Ethiopian health workers towards COVID-19 vaccination. This study examines knowledge, attitudes, perceptions towards the COVID-19 vaccine, level of refusal and associated factors of COVID-19 vaccination among HCWs in Dessie City, Ethiopia.

Study design, period and setting
This institution-based cross-sectional study was conducted in Dessie City in May, 2021. Dessie is the capital of South Wollo Zone, located about 400 km from Addis Ababa. The city is located at an altitude of 2,470 to 2,550 meters. According to Dessie City administration health office, the city has 8 health centers, 2 government hospitals, 3 private hospitals and 15 clinics.

Source population, inclusion and exclusion criteria
The source population for the study included all HCWs in Dessie City. The study population was all HCWs in the selected health facilities. All HCWs present at the selected health institutions during the survey were included in the study. Those who were absent (on sick leave, annual leave, or maternity leave) were excluded and not considered in sampling procedure (interval calculation) based on payroll lists.

Sample size determination and sampling procedure
A total of 404 study participants were identified by using the single population proportion formula and considering the following assumptions: where: n: is the minimum sample size required, Z α/2 is the standard normal variable at (1-α) % confidence level (α is 0.05 with 95% CI, Z α/2 = 1.96), P is an estimate of the attitude towards COVID-19 vaccine (50%) and d is the margin of error (5%). Based on these assumptions, 384 individuals were estimated. Then 5% of the sample size n = z a/2 contingency was added to minimize errors arising from the likelihood of non-compliance, giving a final sample size of 404.
To select study subjects, 11 health institutions (2 government hospitals, 2 private hospitals, 4 health centers, and 3 private clinics) were randomly selected. For the selected healthcare facilities, study subjects' proportional allocation was made based on their number regardless of the profession. Then each institution's payroll for all professions and systematic random sampling was used to identify study participants. The first participant was selected by the simple lottery method. Professions represented in the study included medical doctors, laboratory technicians, nurses, midwives, ophthalmologists, pharmacists, public health officers, radiographers, environmental health specialists, anesthesiologists, porters, medical record workers, administrative staff, cleaners, laundry workers, and guards.

Outcome measurement and explanatory variables
The independent variables were socio-demographic factors, knowledge, attitudes, and perceptions and the dependent variable was refusalof to receive COVID-19 vaccination. To compute the three outcome variables of knowledge (good or poor), attitude (positive or negative), and perception (good or poor), we used the mean score of responses to 9 questions about knowledge and perceptions and 8 questions about attitudes. A score of 1 was assigned for poor knowledge, negative attitude and poor perception and a score of 2 for good knowledge, positive attitude and good perception. Scores above the mean value for each category were considered as good knowledge, positive attitude, and good perception. In addition, refusal to receive the COVID-19 vaccine was correlated as an outcome variable with socio-demographic, knowledge, attitude, and perception as independent variables.

Data collection and quality assurance
The data were collected using a structured questionnaire adapted from relevant literature. Data were collected both through a self-administered questionnaire and interviews with selected professionals. Ten nurses with BSc degrees and 5 environmental health professionals with BSc degrees were employed as data collectors and supervisors. To ensure the acquisition of reliable data, both data collectors and supervisors were trained by the principal investigator for one day on the objectives of the study, the content of the questionnaire, ethical issues, and methods of data collection. We also pre-tested the questionnaire on 5% of the sample size among HCWs in nearby Haik Town before actual data collection and made adjustments in the questionnaire where necessary. The collected data were checked daily by the supervisors for completeness. To ensure accurate and reliable data, the reliability of the questionnaire was assessed using the Cronbach alpha test, giving values of 0.82, 0.79, and 0.87 for knowledge, attitude and perception questions, respectively. We also assessed the validity of this questionnaire by recruiting experts to evaluate the content of the questionnaire using face validity, resulting in the experts agreeing that the test was a valid measure of the concept being measured. This means that they evaluated whether each of the measured items matched any given domain of the concept.

Data management and statistical analysis
The collected data were checked, coded, and entered into EpiData Version 4.6 and exported to Statistical Package for Social Sciences (SPSS) Version 25.0 for data cleaning and analysis. Descriptive statistics, such as frequency distribution (n) and proportions (%) were computed. We ran a logistic regression analysis using socio-demographic, knowledge, attitudes and perceptions as independent variables and refusal of vaccination as the dependent variable. First, bivariate logistic regression analysis (crude odds ratio [COR] at 95% confidence interval [CI]) was performed and then variables with p< 0.25 were retained for multivariablelogistic regression analysis.
From the multivariable analysis, variables with a significance level of p< 0.05 with adjusted odds ratio (AOR) at 95% CI were taken as statistically significant factors that were independently associated with the refusal of vaccination.
Model goodness of fit was checked using the Hosmer and Lemeshow test; P-values greater than 0.05 showed a value of 0.27, indicating the model was fit. Multi-collinearity between independent variables was also checked using the standard error of the coefficient of the model with a cut off value of 2. A maximum standard error value of 0.683 was reported from our model indicating the absence of multi-collinearity between independent variables.

HCWs knowledge about COVID-19 vaccines
The mean score of the knowledge was 14 HCWs considered the differences in the effectiveness among the Pfizer, Moderna, and Astra Zeneca COVID-19 vaccines to be large. More than three-quarters (77.6%) of the HCWs considered the major mode of COVID-19 transmission to be touching contaminated surfaces and touching one's eyes, nose, and mouth, followed by shaking hands, hugging, and kissing (70.4%) and inhaling the virus (69.9%) ( Table 2).
COVID-19 disease was associated with untreated infections. Two hundred thirteen (54.3%) of the respondents stated that it could result in pneumonia and 279 (71.2%) in respiratory failure (Fig. 1).

HCWs attitudes towards COVID-19 vaccines
The mean score of attitudes was 13.31, with a standard deviation of 1.529. The overall positive attitude rate about the COVID-19 vaccine was 52.3% with 95% CI (47.7 -57.4%). Of those who had a positive attitude towards the COVID-19 vaccine, 189 (79.75%) were planning to get vaccinated when a vaccine becomes available.   Table 3).

Proportion of refusal to be vaccinated for COVID-19
One pharmacists refused to be vaccinated but this association was not statistically significant (Table 6).

Discussion
On-going efforts are been made to end the COVID-19 pandemic. Various COVID-19 vaccines have been distributed in many countries, including Ethiopia. The overall good knowledge, good perception, and positive attitude rates of the HCWs about COVID-19 vaccines were 62.5%, 60.5%, and 52.3%, respectively. In our study, 64.0% of the HCWs planned to be vaccinated and 36.0% refused to do so. A similar vaccine acceptance rate was reported by a study in Iraq (61.7%) [18], which was higher than in two studies in the USA, where more than half of all HCWs were undecided and delayed the decision to be vaccinated [19,20]. Low acceptance rates were also reported among healthcare workers in Ghana (39.3%) [21]), the Democratic Republic of Congo (27.7%) [22]), Egypt (21%) [23]), and Nepal (38.3%) [24]). Studies in Nigeria and Saudi Arabia reported intended vaccine uptake rates of 50.2% [25] and 50.52%, respectively [26]. Low rates may be due to earlier study dates (when prospects of the vaccine rollout were uncertain), HCWs limited knowledge about vaccines, lack of trust in government management capacity, and concerns about vaccine safety. Some of the highest vaccine acceptance rates (above 95%) were reported from South and Southeast Asia, where HCWs were willing to be vaccinated because as they perceived the pandemic to be severe, considered the vaccines to be safe, experienced few financial constraints and little stigmatization of being vaccinated, and trusted the health authorities [27]. Studies in China and Vietnam reported intended vaccine uptakes of 76.63% [28] and 76.10%, respectively [29], slightly higher than our findings. These high rates were associated with good knowledge regarding the severity of COVID-19, HCWs' trust in the vaccines, and earlier study dates than in our study.
In this study, high rates of not intending to be vaccinated when vaccines become available were associated with negative attitude and low perception about COVID-19 vaccine. Our study also showed that being a nurse and midwife was significantly associated with acceptance to be vaccinated for COVID-19. Consistent with our findings, a study in India found vaccine acceptance to be highest among nurses [30]. However, in a study in the Democratic Republic of the Congo, the highest acceptance rates among HCWs were reported for medical doctors [22] and in Debre Tabor Hospital in northern Ethiopia, nurses had the lowest acceptance rates [31].
Our findings corroborate a study from the Democratic Republic of the Congo that reported that having a positive attitude towards COVID-19 vaccines was significantly associated with the willingness to receive a vaccine [22]. A study among HCWs of an inner-city hospital in New York [32] reported similar results. A study in Libya indicated that having a family member or friend infected with COVID-19 was positively associated with the likelihood of vaccine acceptance [33]. A community-based study in Wolaita Zone, Ethiopia, showed that family members and friends who had been tested for COVID-19 were significantly associated with the acceptance of COVID-19 vaccine [34]. In our study, having COVID-19-infected family members, colleagues, or friends was associated with increased vaccine uptake intention only in the bivariate analysis but not in the final adjusted logistic model (Table 5). In addition, as indicated in Table 4, 39.5% of the HCWs associated vaccines with possible COVID-19 infections and 46.9% of them thought that vaccines could worsen any pre-existing medical conditions. This is very concerning, as it indicates that HCWs are not convinced about the safety of COVID-19 vaccines and are also misinformed. The reason for this widespread misinformation is not known but the public health agencies in Ethiopia may need to increase their efforts to educate HCWs about the safety of these vaccines.

Limitation of the study
Bias may have affected the result of our study. Social desirability bias in which HCWs answered questions in a manner that would be viewed favorably by others may have resulted in over-reporting of good attitudes and  perceptions as well as intended uptake of vaccines. Moreover, causal inferences cannot be drawn from this study due to the nature of the study design.

Conclusion
The willingness to get vaccinated against COVID-19 was moderately high among healthcare workers in Dessie City. Negative attitudes and poor perceptions of HCWs towards COVID-19 vaccines were the most significant factors in the refusal to accept vaccines. Because HCWs are scheduled to be among the first to receive vaccines and play a central role in their administration in the population, key factors in their decision-making process, such as knowledge about the safety of vaccines, must be addressed as early as possible. Our findings may inform health planners and administrators in developing relevant interventions that promote COVID-19 vaccination uptake among healthcare workers in Ethiopia. In particular, the large proportions of HCWs who considered vaccines to worsen any pre-existing medical conditions or cause COVID-19 is very concerning, as it seems that HCWs are not convinced about the safety of the vaccines and are also misinformed. These issues need to be urgently addressed by the public health agencies in Ethiopia. Working on behavioral change communication and social mobilization towards COVID-19 vaccines using culturally appropriate manner might increase the attitudes and perceptions, which in turn increasing the rate of COVID-19 vaccination. We also recommend studies in urban and rural communities to investigate the religious conspiracy about COVID-19 vaccines that may ensure increasing the vaccination rate of COVID-19.

Implication of the study
The information on attitudes, knowledge and perceptions pertaining to the refusal of COVID-19 vaccine among HCWs can inform policy makers and administrators about opportunities and constraints in distributing vaccines among HCWs in Dessie and other Ethiopian cites. The findings may thus contribute to developing a strategy for controlling the pandemic by addressing factors significantly affecting vaccination uptake.
Abbreviations AOR: adjusted odds ratio; CI: confidence interval; COR: crude odds ratio; HCWs: healthcare workers. we received when needed. We thank the data collectors and supervisors for their assistance and all study participants for their participation and valuable information.
Authors' contributions AA and MA conceptualized, designed, analyzed, and interpreted the data, prepared a draft manuscript and edited the manuscript. HK conceptualized,