Knowledge, attitude, and practice regarding COVID-19 outbreak in Bangladeshi people: An online-based cross-sectional study

Objectives: In Bangladesh, remarkable measures have been adopted to control the rapid spread of the ongoing COVID-19 epidemic. Adherence to control measures in general populations significantly influences on their perception, knowledge, attitudes, and practices (KAP) towards COVID-19. This study aimed to investigate KAP towards COVID-19 immediately after the locked down and during the rapid rise period of the outbreak. Study design: Cross-sectional study. Methods: This online-based cross-sectional study was conducted from March 29 to April 19, 2020, involving 2017 Bangladeshi residents aged (12–64) years. The sample was successfully recruited via the social media. After consenting, participants completed an online survey assessing socio-demographic variables, perception, and KAP towards COVD-19. Results: Of the 2017 survey participants, 59.8% and 40.2% were male and female, respectively. A majority of the participants were students (71.2%), aged 21-30 years (57.9%), having a bachelor's degree (61.0%), having family income >30000 in BDT (50.0%), and living urban area (69.8). The findings revealed that 48.3% of participants had more accurate knowledge, 62.3% had more positive attitudes, and 55.1% had more frequent practices for COVID-19 prevention. Majority (96.7%) of the participants agreed ‘COVID-19 is a dangerous disease’ and almost all (98.7%) of the participants wore a face mask in the crowded place, (98.8%) agreed to report a suspected case to health authorities, and (93.8%) good practicing of washing hands with soap and water. In multiple logistic regression analyses, the COVID-19 more accurate knowledge was associated with age, and residence. Whereas, the results reported that sociodemographic factors such as being older, having higher education, being unemployment, having a joint family, having monthly family income >30,000 BDT, and having more frequent prevention practices were the more positive attitudes’ factors; and the more frequent prevention practices factors were associated to being female, being older, having higher education, having monthly family income more than 30,000 BDT, urban area and having more positive attitudes. Conclusion: To improve the knowledge, attitudes, and practices of general populations are crucial during the rapid rise period of the COVID-19 outbreak. Therefore, an effective health education program must be initiated.


Introduction
Coronavirus disease 2019 (COVID -19) is a global public health threat creating a pandemic situation around the world. This ongoing threat is caused by a virus, severe acute respiratory syndrome, coronavirus 2 (SARS-CoV-2). 1 In Wuhan, China, the origin of COVID-19, where the first case was reported in December 2019. The ongoing COVID-19 epidemic has spread Bangladesh has not experienced similar viruses such as SARS or MERS diseases until now, and public healthcare systems are not completely prepared for this epidemic. In Bangladesh, the magnitude and rapid proliferation of COVID-19 through slightly symptomatic or asymptomatic infected people show the need to identify the behavioral responses of the population. However, in Bangladesh, some unprecedented measures have been adopted by the government to control COVID-19 transmission. Different global actions plan has been put along with government activities. The response from the government has become more forceful by placing restrictions on gathering and closing educational institutions. 8 The government (Bangladesh Public Administration Ministry) has also announced the shutdown of offices with 30 days of general holiday from March 26 to April 25, 2020, placing restrictions on leaving home after 6.00 pm in the evening, and imposing legal actions on individuals leaving their dwellings after 7.00 pm, 9 development of isolation units, personal hygiene maintenance, isolation and care for infected people and suspected cases. Until now, out of 8, 5 divisions were partially and 3 were completely locked down to avoid contacting, and of the 64 districts 61 were either partially or completely locked down. 10 Besides, on 24 April 2020, the Ministry of Religious Affairs announced that to maintain physical distancing during Ramadan amid the COVID-19 outbreak, the number of people in mosques for evening prayers is restricted to twelve (12) (including senior clerics and other mosque officials), and individuals or institutions are not allowed to arrange or participate in any gathering over Iftar meals at dusk during the  In Bangladesh, the fight against COVID-19 is continuing. People's adherence to these control measures is essential, which is affected by their knowledge, attitudes, and practices (KAP) towards COVID-19. 11, 12 There are a limited number of studies on knowledge and attitudes during epidemics that have been conducted in Bangladesh. However, the lessons learned from the studies conducted in other countries in an epidemic situation such as the SARS outbreak in 2003 suggest that knowledge and attitudes towards infectious diseases are associated with panic emotion among the population, which can further complicate attempts to prevent the spread of the disease. 13,14 Suggestions from another study In Latin America, during the outbreaks of chikungunya, zika, and dengue reported low levels of participation and commitment to the imposed control measures in populations. 15 Knowledge of the disease is an important cognitive key in public health regarding health prevention and promotion. It involves a range of beliefs about the causes of the disease and exacerbating factors, identification of symptoms, and available methods of treatments and consequences. 16 Beliefs about COVID-19 come from different sources, such as stereotypes concerning similar viral diseases, governmental information, social media and internet, previous personal experiences, and medical sources. The accuracy of these beliefs may determine different behaviors about prevention and could vary in the population. In many cases, the absence of knowledge, or if most of these medical-related beliefs are actually false may carry a potential risk factor. 17 In Hubei, China, one of the first studies analyzing attitudes and knowledge about  concluded that attitudes towards government measures to contain the epidemic were highly associated with the level of knowledge about COVID-19. 18 The authors reported higher the level of information and education was associated with more positive attitudes towards Covid-19 preventive practices. 6,18 Perception of risk is also a key factor in commitment to prevention during outbreaks of global epidemics. 6,[19][20][21][22] Knowledge of infection process precautions may be linked to citizens following government guidelines regarding prevention and quarantine measures. Numerous studies also reported level of knowledge directly affects the perception of disease susceptibility. 7,21,22 On the other hand, knowledge may also influence attitudes and practices. Attitudes are a motivational/emotional posture that prepares us to respond properly or not, in a specific context. In COVID-19 breakout, positive attitudes and good practices may mediate safe behaviors, prevention measures, and disease propagation. 23 It is notable mentioning that, despite the mandatory nature of public protective measures, the adherence to each of them is still poor among some populations.
Considering the lack of studies related to epidemics and how to facilitate outbreak management of COVID-19 in Bangladesh, and hence there is an urgent need to understand the public's perception and awareness of COVID-19. This study aimed to investigate the KAP towards COVID-19 among general people during the rapid rise period and immediately after the lockdown of Bangladesh of the COVID-19 outbreak.

Methods
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

Participation and procedure
A cross-sectional and anonymous online population-based survey was carried out involving 2,017 individuals aged (12-64) years. The survey was conducted from March 29 to April 20, 2020, immediately after the lockdown of Bangladesh. A semi-structured questionnaire was designed for the Google survey tool (Google Forms), and the generated link was shared on social media (i.e., Facebook, WhatsApp). The link was also shared personally to the contact list of investigators and research assistants. The investigators decided to collect the data via online due to maintaining social distance during the strict lockdown in Bangladesh. Initially, 2,068 potential respondents provided informed consent. Of them, 2,017 respondents completed the entire survey, generating a response rate of 97.5%. The inclusion criteria to participate in the study were being a Bangladeshi resident, having internet access, and volunteering. Those who did not willingly agree to participate in the study, serious and emergency patients were excluded from the study.

Measures
A semi-structured and self-reported questionnaire containing informed consent, questions regarding socio-demographics, knowledge, attitude, and practice.

Socio-demographic measures
Socio-demographic information was collected, including gender, age, education, occupation, marital status, nature of the family (nuclear/joint, with the joint being an extended family, often of multiple generations), number of family members, monthly family income, and permanent residence.

Knowledge, attitude, and practice
With regard to assessing the level of knowledge, attitude, and practice of the respondents, a total of 17 questions (including 6 for knowledge, 6 for attitude, and 7 for practice) was asked to the respondents during the survey. The survey questions were adapted and modified from previously published literature regarding viral epidemics related to MERS-CoV disease, 24,25 infection prevention and control measures for COVID-19 by World Health Organization, 26  is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 27, 2020. After an initial draft of the survey questionnaire, it was validated and adopted by the following two ways. Firstly, the questionnaire was sent to four experts who were well command over this field. After the coordination of all expert's opinions, the final questionnaire was designed.
Secondly, A pilot test was conducted on 30 individuals to confirm the reliability of the questionnaire. Finally, these 30 data were inputted in SPSS version 25 and performed reliability coefficient analysis. Cronbach's alpha coefficient of KAP questions was 0.73, which indicates acceptable internal consistency. 28 The knowledge section consisted of 6 items of a and each question was responded as "Yes", "No" and "Don't know". The correct answer was coded as 1, while the wrong answer was coded as 0. The total score ranged from 0-6, with an overall greater score indicates more accurate knowledge. A cut off level of ≥4 was set for more accurate knowledge.
The attitude section consisted of 6 items, and the response of each item was responded on a 3point Likert scale as follows 0 ("Disagree"), 1 ("Undecided"), and 2("Agree"). The total score ranges from 0 to 12, with an overall greater score indicates more positive attitudes towards the COVID-19. A cut off level of ≥11was set for more positive attitudes towards the prevention of COVID-19.
The practice section included 7 items practice measures responding to the COVID-19, and each item was responded as "Yes", "No", and "Sometimes". Practice items' total score ranges from 0-7, with an overall greater score indicates more frequent practices towards the COVID-19. A cut off level of ≥6was set for more frequent practices.

Statistical analysis
The data analysis was performed using Microsoft Excel 2019 and SPSS version 25.0. Microsoft Excel was used for editing, sorting, and coding. The excel file was then imported into SPSS software. Descriptive statistics (frequencies, percentages, means, standard deviations) and first-order analyses (i.e., chi-square tests) were performed. Binary logistic regression was performed with a 95% confidence interval to determine significant associations between categorical dependent and independent variables.

Ethical considerations
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 27, 2020. ; https://doi.org/10.1101/2020.05.26.20105700 doi: medRxiv preprint The study was conducted in accordance with the Institutional Research Ethics and the declaration of Helsinki. Formal ethical approval was granted by the Ethical Review Committee, Uttara Adhunik Medical College, Uttara, Dhaka-1260, Bangladesh. The consent form documented the aims, nature, and procedure of the study. Anonymity and confidentially were strictly maintained.

Results
A total of 2017 respondents were included in the final analysis, of which 59.8% male and 40.2% female, the average age was 24.36 years (SD=5.38) ranging from 12 to 64 years, and almost all were unmarried (80.8%). The majority were students (71.2%), had a bachelor's level of education (61.0%), came from urban areas (69.8%), lived in nuclear families (77.9%) and their monthly family income more 30000 BDT (50.0%) ( Table 1).

Perception towards the COVID-19 about mode of transmission, incubation period, symptoms, risk factors, treatments, prevention, initiatives, and challenges
With regard to assessing the perception towards the COVID-19, several questions were asked to the respondents during the survey ( Table 2) . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The respondents took the initiative to protect their family members: temporary closure of outside people coming inside the home (87.8%), arrange for handwashing with soap inside or outside the home (85.5%), and wash hands with soap after touching pets (39.4%). The respondents also reported that they faced many problems to create awareness among their family members: not being able to stop going out of the house (57.1%), negligence about the severity of the disease (40.3%), reluctance to use masks (25.5%), and only 19.7% had no problems.

Knowledge
For each question of knowledge, the distribution of responses from participants was presented in Table 3 with gender differences. There were no significant gender differences for each item of knowledge questions.
The findings revealed that 48.3% of respondents had more accurate knowledge, and 51.7% of respondents had comparatively inaccurate knowledge towards the COVID-19. The rates of adequate knowledge were significantly more likely to (ⅰ) be younger (12-20 years) (49.3% vs. 38.8% in aged more than 30 years, p=.029), and (ⅱ) be a respondent from a rural area (52.8% vs. 46.3% in those from an urban area, p=.008).
The sociodemographic factors of more accurate knowledge were 12-20 aged group vs. above

Attitude
For each question of attitude, the distribution of responses from participants was presented in Table 4 with gender difference. The response rates of "Agree" were significantly higher in . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 27, 2020. ; https://doi.org/10.1101/2020.05.26.20105700 doi: medRxiv preprint females (99.5% vs. 98.3% in males, p=.043) to the item of attitude section regarding "It is crucial to report a suspected case to health authorities". Furthermore, the response rates of "Agree" were significantly higher in females (99.6% vs. 98.1% in males, p=.011) to the item of attitude section regarding "It is important to use a face mask in a crowded place." The findings indicated that 62.3% of respondents had more positive attitudes, and 37.7% of respondents had comparatively less positive attitudes towards the COVID-19. The rates of more positive attitudes were significantly more likely to (ⅰ) be an older (age more than 30

Practice
For each question of practice, the distribution of responses from participants was presented in Table 5 with gender difference. The response rates of "Yes" were significantly higher in females (81.8% vs. 73.5% in males, p<.001) to the item of practice section regarding "Do you use tissues or hanker chips during coughing/sneezing?". The response rates of "Yes" were significantly higher in females (95.6% vs. 92.5% in males, p=.023) to the item of practice section regarding "Do you wash hands frequently using water and soaps?". The response rates of "Yes" were significantly higher in females (96.2% vs. 87.1% in males, p<.001) to the item of practice section regarding "Do you maintain social distance (or home quarantine)?". The . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 27, 2020.

Discussion
Currently, COVID-19 pandemic suffering both the number of cases and deaths are increased around the world. The number of newly infected Bangladeshi people is increasing as the days go by. Considering this outbreak, this study was conducted aiming at measuring the level of knowledge, attitude, and practice of COVID-19 and perceptions regarding it among Bangladeshi people. No data have been published on the knowledge, attitudes, and practice among Bangladeshi people during the outbreak of COVID-19. Therefore, this is the first study in Bangladesh examining the KAP towards COVID-19 among general people in Bangladesh.
The findings of this study may be useful when planning health education programs about this emerging infectious disease in the near future.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 27, 2020. ; https://doi.org/10.1101/2020.05. 26.20105700 doi: medRxiv preprint In perception towards COVID-19, most of the study participants (99.35%) reported some symptoms are recognized as COVID-19 related: fever, dry cough, breathing difficulty, however, this has not been confirmed as part of the diagnosis, 29 participants also reported other important symptoms such as sore throat and blocked nose (51.16%), followed by headache, and diarrhea. Only 8 (.3%) respondents did not know the symptoms which were 56.7% in Srichan et al. study. 30  The most identified gap in knowledge among participants was related to disease treatment. The present study reported 18.3% of participants believed that there is no treatment for COVID-19.
In contrast to 47.3% participants addressed COVID-19 is a treatable disease revealed in Srichan et al. study. 30 However, present also found 80.7% agreed with using supportive treatment for COVID-19, and only 1% of the participants reported in the presence of a vaccine for preventing COVID-19, which was 31.2% reported by Srichan et al. 30 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 27, 2020. The results of our study showed that the responses of knowledge about COVID-19 were differed regarding gender but insignificant. The majority of participants (96.7) reported COVID-19 is dangerous and 60% reported it affects only humans. However, recently, there is no evidence that companion animals/pets such as dogs, cats can be infected by the new coronavirus. 26 Participant's more accurate knowledge was significantly more likely to be a young adult (21-30 years, p=.029), and be a respondent from a rural area (p=.008). This may be due to most of the participants of our study were students and they all went back home mostly to rural areas during this locked down period.
The main sociodemographic factors of more accurate knowledge were being young (12-20 aged group: p=.012; 21-30 aged group: p=.010), and residence of "rural area" (p=.008). In The findings showed highly more positive attitudes among the participants towards reporting health authorities to the suspected case of COVID-19 (98.8%), wearing a face mask before going to a crowded place (98.7%), and in the following other recommendations. These findings . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 27, 2020. ; https://doi.org/10.1101/2020.05.26.20105700 doi: medRxiv preprint were similar to a very recent study conducted in China, during the rapid rise of COVID-19 outbreak, reported optimistic attitudes towards COVID-19: most residents took precautions to prevent infection by COVID-19: not going to crowded places and wearing masks when going outside. 18 Saqlain et al. also reported highly more positive attitudes among all the healthcare professionals towards wearing gowns, gloves. 30 The present study findings also revealed an optimistic attitude presenting less likelihood of 'undecided' and 'disagree' answers in regard to 'wash hands and face after coming outsides ('Agree'=99.5%)', and 'health education can play an important role for COVID-19 prevention ('Agree'=95.6%)'. The multiple logistic regression findings of our study revealed that sociodemographic variables associated to more positive attitudes regarding COVID-19 were being an older (p<.001), having higher education (above bachelor) (p<.001), being unemployment (p=.048), having joint family (p=.029), having monthly family income > 30,000 BDT (p=.005), and having more frequent practices (p<.001). Using similar analysis in a study conducted by Zhong et al. found marital status of "others" (p<.001), female gender (p <.001), age-groups of 16-29 years (p =.013), education levels of associate's degree or higher (p ≤.008), occupations of . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 27, 2020. ; https://doi.org/10.1101/2020.05.26.20105700 doi: medRxiv preprint unemployment (p =.001) and students (p =.043) and COVID-19 knowledge score (p <.001) were significantly associated factors. 18 The results of our study revealed that the majority of the participants had more frequent practices in the following precautionary measures. Highest (93.8%) more frequent practices were observed in washing hands with water and soap which was almost similar to the findings Considering Bangladesh, a multi-ethnic country, with different economic income statuses, education levels, traditions, and the quick spread out of the COVID-19 and government response; it is expected that the levels of knowledge, attitude, and prevention display differences in the population. In our study, although a good proportion of the sample has good KAP.
Nonetheless, governmental policy and more health promotion and prevention are needed. We consider that exist susceptible clusters of people that may be vulnerable to prevention practices, related to less attitude and low knowledge about COVID-19. 23 Accessibility to information, preventive behavior, and sanitary educational measures are essential, especially in rural areas . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 27, 2020. ; and among old people, which may have difficulties getting access to novel information or financial/technical adversity to performing preventive measures.
In COVID-19 breakout, the knowledge accuracy of the disease is constantly confronted against the current state of objective medical knowledge. 16 It is usually believed that a more educated population/patients about the disease and those have more accurate knowledge will comply better with the preventive and treatment measures. 39 Regarding this quick implementation of preventive measures, psychological adaptation is a key factor promoted by health prevention/promotion programs.

Limitations
This study has some limitations. First, this study followed a cross-sectional study design.
Therefore, causal inferences may not be established. Second, compared with face-to-face interviews, self-reporting has limitations including multiple biases. Third, related to the limited sample representativeness, this study used an online-based survey method to avoid possible transmission, causing the sampling in our study to be voluntary and conducted by an online system. In this regard, the possibility of selection bias must be considered, so this study was not representative, for example, of those who have limited access to the internet like vulnerable populations including those have lower income or have a lower level of education (less than 13% have secondary and intermediate level of education), older people, and rural people. Therefore, KAP study towards the COVID-19 pandemic of vulnerable populations warrants special research attention in today's Bangladesh. Fourth, we used a limited number of questions to measure the level of knowledge, attitude, and practice. Thus, additional assessments would be important, using all aspects of KAP towards COVID-19, to determine the actual extent of KAP among the general population. Additionally, the unstandardized and inadequate assessment of attitudes and practices towards COVID should be developed via focus group discussion and in-depth interviews and constructed as multi-dimensional measures.

Conclusion
Our findings suggest that Bangladeshi general populations of a relatively high-income status and higher education status, being young, in particular female, have had more knowledge, more positive attitudes, and more frequent good practices. Whereas being older, and living in urban area have had less accurate knowledge towards COVID-19 after the immediate locked down . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 27, 2020. ; https://doi.org/10.1101/2020.05.26.20105700 doi: medRxiv preprint and during the rapid rise period of the COVID-19 outbreak. In addition, more favorable knowledge is associated with more positive attitudes and more frequent practices towards COVID-19 but insignificant. However, the figure of the proportion of more accurate knowledge, more positive attitudes, and more frequent practices was not remarkable, suggesting that effective health education programs aimed at improving COVID-19 knowledge, which boosts more positive attitudes and maintaining safe sustained practices.  . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 27, 2020. ; https://doi.org/10.1101/2020.05.26.20105700 doi: medRxiv preprint  . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.