Liu-Ting Lin , Fan Zhang?, Chan-Juan Zhao, Guo-Tian Lin, Yu-Fei Zhai, Hua-Nuo Feng

1. School of Public Health, Hainan Medical University, Hainan Haikou 571199, China 2. Hainan Health School, Hainan Haikou 570311, China

Keywords:Hainan COVID-19 Knowledge-Attitude-Behavior Infulencing factor

ABSTRACT

1. Introduction

At the end of December 2019, Chinese and international researchers sequenced the samples of pneumonia patients with unknown cause in Wuhan and found that it was an unknown beta virus and then it was named as severe acute respiratory syndrome coronavirus 2 (sars-cov-2).The novel coronavirus disease (corona virus disease 2019 (COVID-19) [1]) has been identified by WHO as the cause of infection of sars-cov-2 and it has been confirmed by human to human transmission [2]. On January 20, 2020, COVID-19 was included in class B infectious diseases and managed as class A[3]. On January 22, 2020, the first confirmed COVID-19 case was reported in Hainan. On January 25, Hainan provincial government launched level I emergency response for public health emergencies. On March 11, the general director- of the World Health Organization (WHO),Tan Desai, announced in Geneva that COVID-19 was "pandemic", which alarmed a global public health event that seriously threatens human beings. On April 17, 2020, there were 168 confirmed COVID-19 cases in Hainan Province, with a total number of 2,101,130 cases worldwide. The author investigated the cognition, attitude and behavior of COVID-19 for mobile phone users in Hainan Province during the rising period of the epidemic. The results are as follows.

2. Objects and methods

2.1 Subjects investigated

From January 31 to February 8, 2020, COVID-19 epidemic in Hainan was on the rise, using the network "questionnaire star" platform through WeChat, QQ and other social software to distribute questionnaires to 901 Hainan Province mobile phone users. The subjects answered independently and anonymously. 901 questionnaires were collected, of which 804 were effectively responded , with a response rate of 89.2%.

2.2 Research methods

2.2.1 Sampling

The research group invited mobile phone users of Hainan Province to fill in the online questionnaire through the online recruitment method, and sent the QR code or link of the online questionnaire through snowball way for investigation. In order to distinguish more heterogeneous research objects, the members of this project selected 10 research objects as "first-level seeds" to improve credibility by considering various factors such as age and sex. The questionnaire was then forwarded to 10 of their peers who were considered suitable for the survey as a "secondary seed" and posted in their WeChat circle of friends [4].

2.2.2 Survey tools

The questionnaire was compiled in the framework of health belief model, referring to the relevant literature at home and abroad, and consulting epidemiologists. After pre-survey, the questionnaire was revised many times to form the questionnaire on cognition, attitude and behavior of residents of Hainan Province on COVID-19.The questionnaire consist of four parts:(1)basic demographic characteristics: sex, age, nationality, education, conscious health condition, etc.;(2)COVID-19 and its disease cognition;(3)attitude towards COVID-19and its disease; (4) personal protective behavior.

2.2.3 Marking

COVID-19 related knowledge cognition 5 multiple topics, one 20 points, less or wrong selection were counted as 0 points. There were 6 questions in the attitude questionnaire, and the main evaluation basis was the frequency of the question meaning, which was divided into 5 levels: Full approval, Approval, Uncertainty, Incomplete approval and Disapproval, with the scores of 2, 1, 0, - 1 and - 2 in turn. The higher the score was, the more correct the attitude will be. There were 6 questions in the behavior questionnaire, and the main evaluation basis was the frequency of the question meaning, which was divided into 3 levels: Yes, Unsure, No. The higher the score, the better the behavior. The total score of knowledge was 100 (＜ 60 was poor awareness, ≥60 was good awareness), the total score of attitude was 12 (＜ 11 was non-supportive attitude, ≥ 11 was supportive attitude), the full score of behavior was 12 (＜ 11 was negative behavior, ≥ 11 was positive protective behavior).

2.2.4 Evaluation indicators

COVID-19 related knowledge awareness rate (total score ≥60 points / single answer total number)×100%, ＜ 80% was poor awareness, and ≥80% was good awareness. Supportive attitude rate = (Number of supportive attitude / total number of respondents to the question) ×100%, and positive protective behavior rate = (number of people who choose to positive protective behavior / total number of respondents to the question) ×100%.

2.2.5 Quality control

The members of the research team checked the questionnaire according to the survey location (not in Hainan Province), filling time (＜ 2min), mobile IP address (duplicate item) and the logic of the questionnaire (such as the questionnaire with age of 1 year, education background of doctor, etc.), and rejected the unqualified questionnaire. 901 questionnaires were collected, and 804 valid questionnaires were obtained after the time of filling in the questionnaire, the residents of the survey site, the IP address of mobile phone and the obvious logical errors in the verification. The effective response rate of the questionnaire was 89.2%.

2.3 Statistical treatment

Data extraction was exported by questionnaire stars, and analyzed by Excel 2016, and statistical analysis was carried out by SPSS 25.0 software. Demographic characteristics were described by continuous variables and categorical variables. The change of attitude could be divided into two categories: supportive attitude and non-supportive attitude; the change of behavior could be divided into two categories: positive behavior and negative behavior. Behavior measures as dependent variables, demographic characteristics, COVID-19 related cognition and attitudes were independent variables. The chi square test was used for single factor analysis. Then, Spearman correlation analysis was conducted on the scores of knowledge and belief. Finally, the two factors were selected for P＜0.05 binary logistic analysis (α = 0. 05) to analyze COVID-19 knowledge level and its influence on personal protective behavior.

3. Results

3.1 Basic characteristics of the subjects

The youngest was 13 years old, the oldest was 65 years old, and the average age was 27 years old. The specific distribution of sex, ethnicity , age, education, region, income, health condition, etc. was detailed in Table 1.

Tab. 1 Basic characteristics of subjects [number (%)]

3.2 COVID-19 knowledge and single factor analysis

The score of COVID-19 related cognition was (61.22±22.453), in which good awareness accounted for 71.5% (575/804). Details of the responses were given in tab.2.Through single factor analysis, the cognitive differences of different gender, age, education level and economic status were statistically significant (P＜ 0.05),as detailed in Tab.3.There was no significant difference in cognition of different ethnicities (χ2= 2.161, P = 0.142), regions (χ2= 3.611, P = 0.164), and health conditions (χ2= 1.473, P = 0.479).

Tab. 2 Awareness of epidemiological characteristics [number, (%)]

Tab. 3 Comparison of cognition status of COVID-19 infection in different populations (number (%)]

3.3 Attitude towards COVID-19

The average score of attitude among 804 subjects was (10.69 ± 2.26), 584 of which (72.6%) were ≥11. 98.0% agreed to " wash hands frequently with soap and running water ", 98.5% agreed to " keep enough sleep, nutrition, avoid excessive fatigue, enhance body immunity ", 98.8% agreed to " keep indoor ventilation and ventilation ", 93.4% agreed to " reduce handshake and hug", 97.9% agreed to " have fever, fatigue, dry cough and other symptoms, and timely medical treatment", 97.5% agreed to "no contact, hunting Selling, purchasing, processing and eating wild animals ". The statistically significant differences of attitude were found in enthnicities, age, education, region, income and health condition (P＜0.05), as shown in Tab. 4. There was no significant difference between different genders (χ2= 0.444, P = 0.505).

Tab. 4 attitudes of COVID-19with different characteristics (number (%))

3.4 COVID-19 related positive protective behaviors

Among 804 subjects, 592 (73.6%)≥11 points. COVID-19 infection would be actively understood in 96.1%. 95.6% would actively publicize knowledge about COVID-19 infection to family members and friends, 96.8% would discourage family members and friends from going to parties, 96% would go out with medical masks, 68.7% would monitor the body temperature voluntarily at home, and 81.3% would be vaccinated with a COVID-19 vaccine. COVID-19 behavior differences were statistically significant (P＜0.05) in different age, education, health condition, cognition and attitude, as shown in Tab. 5. There was no significant difference in behavior among different genders (χ2= 0.001, P = 0.976 ＞ 0.05), regions (χ2= 5.790, P = 0.055 ＞ 0.05), family income (χ2= 0.72, P = 0.788 ＞ 0.05).

Tab. 5 Protective behaviors COVID-19 among different characteristics (number of people (%)]

3.5 Correlation analysis of KAP scores

COVID-19 related knowledge scores were weakly positive correlated with the scores of attitude (r=0.183, P＜0.001). The scores of knowledge were strongly positive correlated with the scores of behavioral (r=0.770, P=0.03), and the scores of behavior were weakly positive correlated with the scores of attitude (r=0.230, P＜0.001).

3.6 Binary logistic analysis of COVID-19 protective behaviors

Indexes with statistical significance of single factor analysis results were included in the multifactor model. COVID-19 protective behavior model was established by stepwise backward logistics regression analysis. The -2 times logarithmic likelihood ratio was 902.751. Hosmer's pseudo fit test was used to test P＞0.05. H0 hypothesis was accepted. The data fitted well with the regression model. The prediction rate of the model reached 73.8%＞60%.

The Tab. 6 for the variable assignment table of logistic regression analysis, set the protective behavior as the dependent variable (y), and set attitude (x1), cognition (x2), education (x3) and age (x4) as the independent variables in the binary logistic regression model. It could be seen from tab. 7 that there was no statistically significant difference in education background and age (P＞ 0.05), and these two variables were excluded from the model. The possibility for subjects with positive attitude was 1.635 times higher of having protective behavior than those with negative attitude (95% CI: 1.154-2.316), and the possibility for subjects with good awareness was 1.515 times higher of having protective behavior than those with poor knowledge (95% CI: 1.072-2.140).

Tab. 6 COVID-19 protective behavior in different characteristics by Logistic regression analysis variable assignment

3.7 Access to epidemic information of COVID-19

During the outbreak, COVID-19 related information was mainly obtained from social software (80%), social platform (69.9%), TV (67%), website (65.4%), friends and relatives (38.9%), broadcast (36.3%), magazine (18.5%).

4. Discussion

4.1 Mobile phone users COVID-19 in Hainan requirement to be improved, and health education can be strengthened through new media.

In total, the mobile phone users' COVID-19 accounted for only 71.5% of the total number of patients, and the positive attitude was only 72.6%. The 73.6% of the patients were well protected. The Hainan mobile phone users' KAP status of COVID-19 was not good enough, and it still needs to be improved by strengthening health education. Knowledge score of COVID-19 was relatively low and the dispersion degree was relatively large. Only 28.4% of the 804 subjects knew that atypical symptoms of COVID-19 were present. The atypical symptoms of patients are the most easily neglected source of infection.

During the rise of the epidemic in Hainan, 31.7% of the respondents believed that books delivered by Wuhan express would be infectious, while 9.7% believed that medical masks made in Wuhan would also be infectious, and only 22.5% of the respondents believed that they should "open the doors and windows without going out". This shows that COVID-19 is still lacking in vigilance and correct understanding among the majority of the population. The survey found that COVID-19 related knowledge was acquired through various ways, including social software (80%) and social platform (69.9%). It suggested that relevant departments should make full use of the advantages of new media on the basis of traditional media, and strengthen health education and information disclosure.

4.2 COVID-19 is correlated with KAP, but strong and weak.

The higher the public's awareness of the disease, the higher the correctness and timeliness of healthy behavior [5-6]. Domestic scholars such as Chen Shihai and Shi Jian [7] analyzed the knowledge of AIDS prevention and treatment of migrant workers in Nanning and its correlation with attitude and behavior. The results showed that there was a positive correlation between knowledge and attitude, attitude and behavior, and the more correct attitude, the higher the degree of patients' behavior coordination. The survey found that COVID-19 related awareness and attitude were weakly correlated (r=0.183), and awareness and behavior were strongly correlated (r=0.770), and behavior and attitude were weakly correlated (r=0.230), which was consistent with the related research results. The level of public knowledge and public health attitude are the main factors affecting behavior. The necessary condition of behavior transformation is awareness, but it is not a sufficient condition. The change of attitude and belief is based on awareness, but not necessarily. Obviously, there is a causal relationship among awareness, attitude and behavior, which is not a directional development [8].

4.3 Mobile phone users' good awareness of COVID-19 and positive attitude are the main influencing factors of COVID-19 protective behavior in Hainan.

Multivariate regression analysis showed COVID-19 was goodawareness (OR=1.515, 95%CI:1.072-2.140) and supportive attitude (OR=1.635, 95%CI: 1.154-2.316) was the most important factor to adopt personal protective behavior. According to the theory of health education, in the process of changing from recognition to behavior, it is influenced by many factors such as education level and age [9]. The results of COVID-19 were 96% of the respondents. However, only 68.7% of the patients would take the initiative to monitor their body temperature at home. This shows that there are still some loopholes in personal protective measures, which is not conducive to early detection of low fever patients. Fever is one of the typical symptoms of COVID-19.It is suggested that in the implementation of health education, the low educated should be regarded as the key population of health promotion[10],and the public health literacy should be improved, and the protective behavior should be improved by improving cognition and setting up a positive attitude. At the same time, we should give full play to the supervision role of grid managers in grass-roots communities (health centers), further strengthen the publicity and guidance of residents' self-monitoring body temperature, and early detection and early reporting of fever cases.

Tab. 7 COVID-19 protective behaviors in different characteristics by multivariate Logistic regression analysis

4.4 Research deficiencies and limitations

During the rising period of the epidemic, all areas in China are at the first level of emergency response state [11], a strict randomized surveywas not able to be applied, so we used the non randomized sampling online survey, and because of the tight time, the sample size was small and there must be a certain bias. In the later stage, further research can be carried out by expanding the sample size and improving the sampling method.