Predictors of adherence to COVID-19 prevention measure among communities in North Shoa Zone, Ethiopia based on health belief model: A cross-sectional study

Introduction Coronavirus disease 2019 (COVID-19) is an emerging respiratory infections and is known to cause illness ranging from the common cold to severe acute respiratory syndrome. At present, the disease has been posing a serious threat to the communities, and it is critical to know the communities’ level of adherence on COVID-19 prevention measures. Thus, this study aimed to identify the predictors of adherence to COVID-19 prevention measure among communities in North Shoa zone, Ethiopia by using a health belief model. Methods Community-based cross-sectional study design was employed. A total of 683 respondents were interviewed using a structured and pre-tested questionnaire. The data were collected by using a mobile-based application called “Google form.” Logistic regression was performed to analyze the data. Estimates were reported in adjusted odds ratios with 95% confidence intervals (CI) and a significant association was declared at p-value of less than 0.05. Result The overall adherence level of the community towards the recommended safety measures of COVID-19 was 44.1%. Self-efficacy (AOR = 0.23; 95% 0.14, 0.36), perceived benefits (AOR = 0.35; 95% 0.23, 0.56), perceived barriers (AOR = 3.36; 95% 2.23, 5.10), and perceived susceptibility of COVID-19 (AOR = 1.60; 95% 1.06, 2.39) were important predictors that influenced the adherence of the community to COVID-19 preventive behaviors. Conclusions In this study, the overall adherence level of the community towards the recommended safety measures of COVID-19 was relatively low. It is vital to consider the communities’ self-efficacy, perceived benefits, perceived barriers and perceived susceptibility of COVID-19 in order to improve the adherence of the community towards the recommended safety measures of COVID-19.

numbers of COVID-19 cases were South Africa, Egypt, Morocco, Nigeria and Ethiopia [11]. 54.3% of those infected with SARS-CoV-2 were male with a median age of 56 years. Patients who required intensive care support were older and had multiple comorbidity such as cardiovascular, cerebrovascular, endocrine, digestive, and respiratory disease. Those in intensive care were also more likely to report dyspnea, dizziness, abdominal pain, and anorexia [12].
The impact of the COVID-19 outbreak on the East and Horn of Africa is expected to be far-reaching and more catastrophic than for high income countries, These pre-existing conditions such as the population size, status of health systems and health workforce are expected to worsen any health impacts from COVID-19 [13,14].
Ethiopia is the second most populous nation in Africa, and home to close to 9% of the African population. Ethiopia respond to COVID-19 started early. On January, the government introduced passenger-screening protocols at Addis Ababa's international airport, and further preparation continued in January and February. The first report of a COVID-19 case in Ethiopia was on March 13, two days after the global pandemic was declared. National responses were scaled up soon after, and a state of emergency was declared on April 8 [15]. In response to COVID-19, the federal and regional governments in Ethiopia have been taking a series of policy actions. These include closing schools, restricting use of public transportation, banning large meetings, and suspending sporting and religious gatherings [16].
The following basic measures to reduce transmission of COVID-19 have been recommended by the WHO and have been adopted by the Ethiopian Government, such as wash hands frequently using soap, maintain social distancing, stay informed and follow advice given by healthcare provider, stay at home. If you begin to feel sick and if you develop fever or cough or experience difficulty breathing, seek medical advice and call in advance the center. Even though COVID-19 case diagnosis and knowledge on prevention remain vital components of the control strategies, little is known about the knowledge, risk perception and perceived selfefficacy of the community in the context of Ethiopia. The Ethiopian government is expanding its diagnosis system in order to increase accessibility of COVID-19 diagnosis and treatment to regional and zonal levels. There is no study to date that has been conducted to assess the knowledge of the community on prevention, risk perception and evaluate the implementation of policies and regulations set for the prevention and control of COVID-19 by the community in Ethiopia. Evidence on this subject would contribute to the prevention, control of COVID-19 as well as on the management of its impact. Therefore, this cross-sectional survey is aimed to assess the predictors of adherence COVID-19 prevention measure among communities in North Shoa Zone, Ethiopia based on health belief model. The Health Belief Model (HBM) provided the theoretical base for this study. The HBM was developed by psychologists in the United State Public Health Service in the 1950s as a 6 way to explain participation in medical prevention and disease detection programs [17]. The model suggests that changes in preventive health behaviour are originally based on six factors, susceptibility: perceived personal vulnerability to or subjective risk of a health condition, seriousness: perceived personal harm of the condition, benefits: perceived positive attributes of an action, barriers: perceived negative aspects related to an action, health motivation refers to beliefs and behaviours related to the state of general concern about health and confidence is defined as the belief that one can successfully perform a behaviour that will then lead to a desirable outcome [18,19]. The aim of this model is to increase the perception of individuals about a health threat and direct their behaviors towards health. Likewise, it focuses on a person's health related behavior and belief in predicting future actions. The perceived risk of people developing COVID 19 is considered to be the primary motive to change within the Health Belief Model, which assumes that the higher the perceived threat, the more likely an individual will modify his or her behavior to avoid that threat.

Study setting and period
The study was conducted from May 1-30 /2020 at North Shoa zone, Amhara regional state, Ethiopia. Debre Berhan town (the zone city) is located 130 kilometers far from Addis Ababa (the capital city of Ethiopia). The zone has 21 rural districts and 3 city administrations; 370 rural kebeles, 523,338 households; 370 health posts; and 740 health extension workers. The zone has 1 comprehensive, specialized referral hospital, 7 primary hospitals, 95 health centers, 389 health posts, 2 private hospitals, 34 private medium level clinics, 97 private primary clinics, 8 pharmacies and 1 private rural drug vendor (unpublished zonal health department report, 2020).

Study design
A community-based cross-sectional study design was employed to identify predictors of adherence to COVID 19 prevention measures among communities using the health belief model.

Sample size determination and sampling procedures
The sample size was determined by using single population proportion formula. The following assumption was made: margin of error was 5%, confidence interval was 95%, the proportion of adherence was 50%, design effect was 1.5 and non-response rate was 10%. In this study, a multi-stage sampling technique was applied. Initially, the zone was stratified into two strata: urban and rural districts. Thirty percent of the districts were selected from each stratum (seven from rural district and two from urban districts). In the second stage, a total of 27 kebeles were selected randomly. At kebele level, family folders (list of households taken from health extension workers) were used as the sampling frame and the required numbers of households were selected using systematic random sampling technique. The first household was selected by using a lottery method, and then every 6 th household was included in the study. If data collectors could not find both of the household heads with three visits, they shifted to the next immediate household.

Instrument and measurement
Data was collected by "Google forms". A structured questionnaire was adopted from a WHO survey tool for COVID-19 [20]. The questionnaire was prepared first in English and translated to Amharic and back translated to English before data collection process. The questionnaire had four parts: socio-demographics variables, health belief model constructs, adherence of the community towards COVID-19 prevention measures and knowledge of the community towards COVID-19. Health belief model constructs were a 5-likert scale item (1 = strongly dissatisfied, 5 = strongly dissatisfied), then during analysis strongly dissatisfied and dissatisfied merged to disagree and strongly agree and agree merged to agree. Similarly, the adherence level of the community towards COVID-19 prevention measures were a 5-likert scale item (1 = never, 2 = rarely, 3 = sometimes, 4 = often, 5 = always). Higher scores indicate greater safety measure practice, while lower scores indicate less safety measure practice. Survey questions were tested for content validity and internal validity (reliability). The content validity of questionnaire was evaluated by public health and nursing expertise. Based on their endorsements, modifications to the survey tool were made. Moreover, the Cronbach alpha coefficients of HBM constructs were calculated. The reliability scores are presented in Table 1.
In addition, the reliability score (cronbach alpha) of adherence tool was 0.91. Based on these questions, mean scores of perceived susceptibility, perceived severity, perceived benefits, perceived barriers and cues to action were calculated to classify the respondents into two groups; above the mean and below the mean.

Data collection
Data was collected by 27 health care workers. Google platform was used to collect data. The Universal Resource Locator (URL) address was sent to each data collector through their mobile telegrams. Training was provided for data collectors and supervisors on the objective of the study, data collection procedures, COVID-19 precaution, and ethical considerations.

Data analysis
Data were checked for completeness and inconsistencies. The online datasheet was downloaded from the Google form in excel format and then exported to SPSS version 21 software for further analysis. Descriptive statistics (mean, median, standard deviation, Inter Quartile Range, percentage, and frequency distribution) were computed. Logistic regression was performed to analyze the data. Those independent variables which were statistically significant in the bivariate model (p-value < 0.05) were entered into the multivariable analysis. In the final model, a significant association was declared at a p-value of less than 0.05. The results were presented in texts and tables with adjusted odds ratio (AOR) and the corresponding 95% CI.

Ethical approval and considerations
Ethical clearance was obtained from the Institutional Review Board (IRB) of the Debre Berhan University. Written informed consent was obtained from each participant. All the information obtained from the study participants were kept confidential throughout the process of study, and the name of the participant was replaced by code. Withdrawal from the study at any point if they wished was assured.

Socio-demographic characteristics of participants
Six hundred eighty three respondents were participated in this study. The majority, 390 (57.1%) of participants were male. The median age of the respondents was 38 (IQR 30-48 years). A high number of the respondents 492 (72%) were married. Three hundred forty eight (51%) of the respondents did not attend formal education. More than half of the respondents, 377 (55.2%) were from rural and 251 (36.7%) were farmers. Most of them, 566 (82.9%) had no chronic illness (Table 2).

Knowledge of the community towards COVID-19
The mean knowledge of the participants towards COVID-19 was 10 with (SD ±2.66). Nearly, half of participants (47.1%) had inadequate knowledge of COVID-19.

Health belief model constructs for COVID-19
Perceived sensibility. The overall perceived susceptibility of COVID-19 infection was calculated using the mean score. The mean score of perceived susceptibility of COVID-19 was 13.88 (±3.38 SD). Forty-five percent (44.8%) of the respondents were above the mean score, and they perceived themselves susceptible to COVID-19 (Table 3). Perceived barriers of COVID-19 preventive measure. In our study area, the study participants pointed out their perceived barriers of COVID-19 infection prevention. As demonstrated in Table 4, more than half of participants (55.6) do not have barriers of COVID-19 infection prevention.
Perceived benefit of preventive safety measures of COVID-19. Our study participants also reported the benefit of preventive safety measures of COVID-19 infection. Consequently, nearly three-fourth of our respondents (72.0%) agreed with the usefulness of practicing the recommended safety measures to prevent COVID-19 infection (Table 5). Cues to action to COVID-19. There were signals in the community to act upon the prevention of COVID-19 infections. Accordingly, more than eighty percent of our respondents (84.3%) had cues to act on COVID-19 infection prevention.

The adherence level of the community towards COVID-19 safety measures
The overall adherence level of the community towards the recommended safety measures of COVID-19 was 44.1% (95% CI = 41.1, 48.2). Only nine percent of participants did not practice handwashing with soap and 42.2% of the respondents did not utilize sanitizers to clean hand ( Table 6).

Factors associated with the adherence of the community towards COVID-19 safety measures
The odds of adherence to safety measures of COVID-19 were 1.60 times higher among communities who perceived that they were not susceptible to COVID-19 than their counterparts (AOR: 1.60, 95%CI: 1.06, 2.39). Members of the community who did not have a perception of barriers of COVID-19 measures was 3.36 times higher on poor adherence of COVID-19 recommended preventive measures as compared to those who had a perception of barriers of COVID-19 measures (Table 7).

Discussions
COVID-19 pandemic has created fear circumstances of the human race globally. The impact progress to social activities, education, health services and other services were affected [21]. Ethiopia has implemented preventive guidelines recommended by WHO to protect human to human transmission of COVID-19, such as social distancing, ban on public gathering, religious gatherings, regular personal hygiene, use of face masks, cover the mouth and nose while sneezing or coughing, one per seat in public vehicles, temporary closing of schools, Colleges and Universities [22]. The achievement of the world's fight against COVID-19 depends upon people's adherence to the control measures [23,24]. The battle against COVID-19 in Ethiopia is still in its infancy. The first confirmed case was announced on March 13/2020. The community's adherence to these control measures is largely affected by their health belief towards COVID19 and its preventive measure. Since Ethiopian society is vulnerable to COVID-19, this health belief on risk perception gap is potentially dangerous and should be addressed to mitigate the disease spread.
The HBM implies that an individual's perception of his/her susceptibility to a disease, attached to his/her belief that the disease has potentially serious consequences (perceived seriousness) equals the perceived threat which leads to a behavior such as using safety measure and testing. If the person believes that behaviors such as using safety measure are beneficial and outweigh his/her perceived barriers, then he/she is more likely to adopt the new behavior. Cues to action are events, people or things that encourage people to change their behavior. Examples of agents of cues to action include family, friends, media, and health care providers. Other modifying variables include age, gender, ethnicity, personality, socioeconomic, knowledge and motivation [17].
Findings from this study showed that 102(14.9%) participants were had a chronic disease, whereas study done at Turkish showed that 54 (15.7%) participants had a chronic disease [25].
According to the current study, the mean knowledge score was 10 with (SD ±2.66). More than half 361(52.9%) of the respondents had adequate knowledge on COVID-19. A study done in Pakistan medical students showed that, 80% of participants had sufficient knowledge about coronavirus [26]. A study done in Bangladesh showed that, the mean knowledge score for participants was 10.77 (SD ±0.588). Around 82.85% of the participants were having adequate knowledge on COVID 19 and other finding, 61.2% participants had adequate knowledge to COVID 19 [27,28]. A National Survey of Syrian indicated that, 75.6% Participants were shown a good level of awareness regarding COVID-19 [29]. Finding from Chinese resident, the mean COVID-19 knowledge score was 10.8 (SD±1.6), around 90% of participants were knowledgeable [30]. Study done on quarantine population in Tigrie region, Ethiopia showed that, the mean knowledge score was 8.73 (±2.64). Less than half, 42.9% of the study participants were knowledgeable [31]. Regarding perceived susceptibility, forty-five percent (44.8%) of the respondents were above the mean score, and they perceived themselves susceptible to COVID-19 infection. While a cross-sectional study in Myanmar showed that, 77.2% were had moderate risk perception and 22.8% were had high risk perception [32]. Study done South Korea showed, a high proportion of the respondents were perceived the seriousness of COVID-19 [33]. Cross-sectional studies in America, 50% of study participants were perceived the risk of COVID 19 [34].
A worldwide cross-sectional study across the ten sampled countries, risk perception were varied between 4.78 and 5.45 [35].
Regarding to the perceived severity of COVID-19 infection, 572 (83.7) participants were perceived severity of COVID-19 infection. Study done in India 90% of the participants considered COVID-19 as a serious disease [36]. Study in Egypt and China, most participants were believed that COVID-19 is a life-threatening, so any member of the families have risk of infection [37], whereas 80% of china's were believed that COVID 19 severely harms health [38].
In this study, more than half of participants (55.6) did not have barriers to implement to COVID-19 prevention measures, whereas study in Egypt, more than two fifths (22.7%) of study participants were believed that infection is associated with stigma [37]. In our Study, nearly three-fourth of the study respondents (72.0%) was agreed on the usefulness of practicing the recommended safety measures. The global based study conducted in ten selected countries showed that, the act of practicing recommended safety measure for the greater benefit of society is relevant [35].
The respondents' belief of COVID-19 infection was formulated from HBM constructs. As a result, eighty-four percent of the respondents have cues to perform prevention measure activities towards COVID-19 and sixty-four percent of our study participants have self-efficacy to apply preventive measures of COVID-19 infections. This result indicated that there are still gaps on the belief of COVID-19 infection. Since Ethiopian communities have a strong societal ties and poor socioeconomic basis which are now risk factors for COVID-19 transmission [39,40], the communities need to ensure self-efficacy to apply preventive measures of COVID-19 infections.
This study suggests that, 93 (13.6%) participants were always wearing a mask, 125 (18.3%) were always washing hands with soap for 20 seconds, 90 (13.2%) participants were always practiced avoiding touching eyes, nose, and mouth with unwashed hands, 85 (12.4%) were always staying at home, when have cold or fever and 127 (18.6%) were always applying recommended social distancing. The overall level of adherence to recommended safety measures was 44.8%. The study done in India, showed that 93.2% of the participants have adopted preventive measures to control COVID-19 and 97.6% were wear a mask, 97.3% were frequently wash hands, 97.8% were avoid handshaking and hugging, and 95% were avoid going out [36]. Another study done in Bangladesh, showed that 87.97% of participants were followed social distancing in their daily life [41]. Community based study Done at Pakistan showed that, 92.8% of respondents were practiced maintaining safe physical distance [42]. In Punjab 71.8% participants were always wearing a mask when going outside, 91.8% participants were followed the stay at home policy, 53.8% participants were followed the steps of hand hygiene [43]. Based on an impact survey done America, about 94.7% of participants reported washing or sanitizing hands, 39.6% reported working from home, and 9.6% reported visiting a doctor or hospital [44]. Survey of Bangladish showed that, more than half 51.6% of respondents had good practices of COVID 19 preventive measure [28]. Study done on quarantine population in Tigrie region, Ethiopia showed that, nearly half, 165 (49.8%) of the participants have gone to crowded places, Forty-six percent of the participants did not use a face mask when leaving home, more than half, 54.4% of participants did not obey the preventive measures given by local health care authorities [31]. A study done in US, more than half of patients (58.6%) were reported that the corona virus caused them to change their daily routine activities [45] other study done in China, majority of the 96.4% participants were had not visited any crowded place and 98.0% participants were wore masks when going out [30,45]. Moreover, the reported prevalence of adherence level of community was likely to be conservative as this was based on community recall and self-reports which usually overestimate or underestimate community adherence levels. The cultural and socioeconomic characteristics of the community also determine the community adherence level. Here in the study area the communities are highly socially bonded by "Edir", "Ekub" and funereal ceremonies any other social structures which may affect the adherence level of the community on recommended safety measures.
Regarding with factors, community perception on the susceptibility of COVID-19 measures, barriers on prevention measures of COVID-19 and the presence of self-efficacy showed significant association with level of recommended safety measures adherence on COVID-19.
Poor adherence of recommended safety measures of COVID-19 were 38% less likely among household members who share a common bed compared from the counterpart. This was available evidence implies that, at least 1 meter physical distancing was associated with a large reduction in infection, and distances of 2 meters were more effective. These data also suggest that wearing face masks protect people (both healthcare workers and the general public) against infection by corona viruses, and that eye protection could confer additional benefit (3). These may be possibly due to that if family members share a common bed they may face difficulty in keeping social distancing.
The odds of poor adherence on recommended safety measures of COVID-19 were 1.60 times higher among communities who perceives they are not susceptible for COVID-19 than the counterpart (AOR: 1.60, 95%CI: 1.06, 2.39). A Community that has no barriers on preventive measures were 3.36 times higher on poor adherence of COVID-19 recommended preventive measures. Community members who had self-efficacy 77% more likely adherence to COVID 19 safety measure.

Conclusions
In this study, the overall adherence level of the community towards the recommended safety measures of COVID-19 was relatively low. It is vital to consider the communities' self-efficacy, perceived benefits, perceived barriers and perceived susceptibility of COVID-19 in order to improve the adherence of the community towards the recommended safety measures of COVID-19.

Limitation of the study
This study has different limitations. Firstly, this study followed a cross-sectional study design that cannot establish causal inferences. Secondly, were used face-to-face interviews to collect data, which is prone to social desirability biases. Thirdly, there may be sample selection biases. There was no previously standardized and validated tool to assess the Knowledge and community health belief towards COVID-19. The area is not well studied, so we didn't find adequate studies to compare and contrast our findings with others, and it makes our discussion narrow. Despite these limitations, the study finding provides valuable information about the adherence level of the communities towards the recommended safety measure of COVID 19.
Supporting information S1 Annex. English version of the survey questionnaire. (DOCX)