Understanding citizen health protocol behavioral intention: The mediation role of perceived risk

Amid the outbreak of the Covid-19 virus, there are still many people who do not comply with health protocols, so this condition has prompted the Indonesian government through the Ministry of Health to create a #changeusirwabah campaign with a risk communication approach. This study aims to determine what and to what extent the factors in the campaign can change people's behavior to live a healthy life based on the Health Belief Model and TRA (Theory of Reasoned Action) theory. This study uses a quantitative research approach with regression analysis techniques carried out by taking 170 samples of people in DKI Jakarta who are aware of the Change Banish Outbreak campaign. The role of mediation in this study is exciting research finding. The perceived risk variable used as a communication approach by the Ministry of Health has a vital role in encouraging the influence of campaign messages on behavioral intention. The results in this study can be used as guidelines for an institution to change public health behavior through health campaigns using a risk approach. © 2021 by the authors. Licensee Bussecon International, Istanbul, Turkey. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution 4.0 International license (CC BY) (http://creativecommons.org/licenses/by/4.0/).


Introduction
At the end of 2019, the worldwide community was hit with the Covid-19 pandemic, including Indonesia. Based on data from the Pusdatin of the Ministry of Health published on the official website covid19.go.id (2020), 244,676 confirmed community cases of COVID-19 (as of September 20, 2020). Cases of the spread of COVID-19 fluctuate every month; it is recorded that the accumulated data on the increase in confirmed cases of the COVID-19 virus from March 2020 to September 20, 2020, continues to rise 244,676 people. In response to this, both the government and corporate institutions are trying to prevent the spread of the Coronavirus Disease 2019 . Through the Ministry of Health of the Republic of Indonesia, the government, in collaboration with USAID and PT XL Axiata Tbk, initiated a campaign called Berubah User Wabash (Room, 2020b). The Campaign to Berubah User Wabash was conducted virtually in Jakarta, considering the number of people infected with the COVID-19 virus in DKI Jakarta province was still high compared to other provinces.
Based on data from the Ministry of Health of the Republic of Indonesia (2020), it was noted that DKI Jakarta was the province with the highest case rate -60,828 people (25.3%) of the total national cases of 244,676 people. Of the total cases, 1,527 people died, and 47,176 people were declared cured. Meanwhile, 12,125 people are still in intensive care and self-isolation. Given the high number of positive cases of COVID-19, the Ministry of Health has created a campaign called Berubah Usir Wabah. The campaign is expected to change people's behavior and awareness towards the potential risks of the COVID-19 outbreak, which is in line with the increase of COVID-19 spread cases every month. This campaign is a Behavior Change Communication Strategy in Covid-19 Prevention, while the 321 Interactive Voice Response Service is one of the educational media for COVID-19 prevention. Besides being educational media for COVID-19 prevention, it can also change people's behavior to be healthier (Gunawan, 2020).
The Interactive Voice Response 321 service is one of the communication channels that the general public can access to obtain information about COVID-19. This service provides an opportunity for more Indonesians who have limited access to the internet to accelerate the spread of information about COVID-19 more evenly to all regions in Indonesia, considering that all cellular phone users in Indonesia can access this service. In addition, all information contained in this service is also provided in Indonesian. The Interactive Voice Response 321 service is available in 34 provinces in Indonesia (Ministry of Health, 2020b). The Change Out of Plague Campaign was launched virtually in Jakarta on Tuesday, May 19, 2020. The launch of the Change Banish Plague campaign reflects our shared responsibility to reach out to people who live in remote areas to get information quickly about COVID-19 and various other information, which helps improve the health and welfare of the community (Room, 2020c). One of the activities carried out by the Ministry of Health is by making a Public Service Advertisement Video: Change to Drive Out the Plague-United Against COVID19, which is uploaded on the official website of the Ministry of Health, Directorate of Health Promotion and Community Empowerment and through the official Youtube account of the Ministry of Health. The Ministry of Health utilizes social media Youtube to reach a broader target group, as shown in figure five above. The one-minute video was uploaded on May 14, 2020, with a total number of views of 2,082 views on October 31, 2020. The Change to Banish the Plague Campaign is carried out using a risk communication approach that refers to increasing public knowledge of the potential risks and threats of health problems so that they able to decide on steps and actions that can protect themselves from these problems (Ministry of Health RI Directorate General of Health Promotion and Community Empowerment, 2020b). Public Relations, in general, exist to foster good relations between all components of a corporate institution to provide understanding, foster motivation and participation. All of this aims to grow and develop the public's goodwill and obtain favourable public opinion for the company (Kemal & Ernungtyas, 2020).
Public Relations plays a good role in building a brand, company, or institution to create an identity for the public to recognize. In this study, the #BerubahUsirWabah campaign is carried out by the Ministry of Health to provide knowledge to the public so that they can change their healthy lifestyle by following health protocols. They quoted from Mufarida (2020), based on a survey conducted by the Balitbangkes of the Ministry of Health in July 2020. Five provinces are in a high position of public confidence that they are not exposed to COVID-19. The first is DKI Jakarta with 30 percent, then East Java 29 percent and Central Java 18 percent, then West Java 16 percent, and South Kalimantan 14 percent. With this belief, people are reluctant to follow health protocols due to a lack of public knowledge about COVID-19.
For this reason, it is necessary to look at and consider the campaign message factors and the perceived risk perception of the community to determine the extent to which these factors can change people's behavior to live a healthy life based on the Health Belief Model and TRA (Theory of Reasoned Action) theory. This is supported by previous research conducted by Kemal and Ernungtyas (2020), which found that the public's interest in behavior was influenced by cognitive campaign messages. Meanwhile, Yoo, Kim, and Lee (2018) found that health belief factors influenced people's behavioral interest by looking at risk perception and self-efficacy factors. The results of this study explain that individual assessment of risk is seen as an essential element of most theoretical models of health and risk behavior, including the Self-Efficacy Theory by Bandura (1994), the Health Belief model by Rosenstock (1974), Theory of Reasoned Action by Fishbein and Ajzen (1975) and the Theory of Planned Behavior by Ajzen (1985). This study aims to examine, examine and re-confirm the relationship between the Covid-19 campaign message #berubahusirwabah and behavioral intention through perceived risk in different situations and conditions from the case studies in previous studies. However, this research has industry similarities with previous research, namely the health industry. Furthermore, this research will integrate several theories of the Health Belief Model and TRA (Theory of Reasoned Action) as the theoretical basis.

Theory of Reasoned Action
An individual's behavior about certain things must have a reason; this is in line with the Theory of Reasoned Action (TRA) developed by Icek Ajzen andMartin Fishbein (1975 and1980). This theory assumes that, in general, humans behave consciously, consider the available information, and implicitly or explicitly consider the implications of the actions taken. Taking a deeper look, it can be found that an individual's intention to do (or not to do) behavior is a direct determinant of the action or behavior (Jogiyanto, 2007, p.31).

Health Belief Model
According to Yoo et al. (2018), health belief or belief in healthy life consists of two parts, namely perceived risk and self-efficacy that exist in a person. Perceived risk, consisting of perceived severity (i.e., perceived magnitude of risk) and perceived vulnerability (i.e., perceived likelihood of experiencing risk), and self-efficacy (i.e., a person's ability to carry out recommendations) have been argued to motivate healthy behaviors. Moreover, it has been widely used in health contexts, as mentioned in Yoo et al. (2018) research. The health risk behavior model consists of several fundamental theories: the Health Belief Model, motivational protection theory, Reasoned Action Theory, and Trans-Critical Model. Some of these theories produce various primary hypotheses, where human behavior is rational and intentional.
Individual assessment of risk is considered an essential component of most theoretical models of health and risk behavior, including the Self-Efficacy Theory, Health Belief Model, Reasoned Action Theory, Planned Behavior Theory, Self-Regulatory Theory, Subjective Culture, and Interpersonal Relations Theory. These theories assume that people's beliefs about the consequences of their actions and perceptions of their susceptibility to those consequences hold critical positions in behavior. Although the question arises whether the existing studies adequately address this hypothesis, the strength of the logical association between risk perception and behavior is reassuring. As a result, according to Halpern-Felsher et al. (2001), the perception of risk plays an essential role in behavioral intervention programs, which try to get adolescents to recognize and acknowledge their vulnerability to adverse outcomes.
The Health Belief Model has been used to predict many prevention and adherence health behaviors such as risk-taking behavior and adherence programs. Like other health behavior social cognition models, the Health Belief Model has potential use in identifying several key factors that are important for predicting whether or not a person will engage in health-protective behavior, because the construction of the Health Belief Model is predictive of health behavior, changing these beliefs can lead to changes behavior (Albery & Munafò, 2008). After more research was conducted on Health Belief Models and their use, and after being used to develop and implement programs on various health issues such as patient adherence to medication, preventive behavior, and participation in health screening, the theory was added with two more components. The first is the idea that instructions for action are needed as a direct motivation to take action, and the next, the sixth component, namely one's belief in one's ability to take action, which is commonly known as "confidence in one's abilities" (self-efficacy) (Edberg et al., 2010).  (Becker & Maiman, 1975) In essence, the health belief model is known as a value expectation model, which refers to the assumption that people will engage in healthy behavior if; 1) they judge the outcome (being healthy) about the behavior and 2) they think that the behavior is likely to produce that outcome. The Health Belief Model focuses specifically on threat perception and evaluation of health-related behaviors as the main aspects for understanding how a person represents health measures. So it can be concluded that the factors from the Health Belief Model that can be the cause of the emergence of behavior are: 1) Perception of vulnerability, namely the degree of risk that a person feels towards health problems, 2) Perception of severity, namely the level of belief in a person that the consequences of health problems will be The more severe, 3) Perception of benefits, namely positive outcomes that a person believes as a result of actions, 4) Perceptions of obstacles, namely adverse outcomes that are believed to be the result of actions, 5) Instructions for action, namely external events that motivate a person to act, 6) Self-efficacy that is, a person's belief in his ability to take action.

Campaign Message
The content of the message delivered in a company or organization as a communicator in a campaign. When the message concept is designed for a campaign, a communicator must see whether the audience will find the helpful message informative, trustworthy, convincing, and motivating. The message delivered must be persuasive by providing additional knowledge and adding confidence to the audience. In addition, the message in the campaign must also be in line with the existing facts. The message must be delivered in a good and creative way. The message is published or promoted to be known and understood, which would be accepted by the target audience. Andriani and Putra (2019) shows that business firms may adopt social media platforms to deliver message. Research conducted by Kemal and Ernungtyas (2020) found that campaign messages are one of the factors that influence people's behavioral interests. As a result of those study, the following hypothesis was formulated: H1: Campaign messages have an influence on people's behavioral interest in the #BerubahUsirWabah campaign.

Perceived Risk
According to Slovic, Monahan, and MacGregor (2000), each individual's different decision-making is determined by their perception of the risks they face and the importance of their effects. Risk perception is a form of interpretation or assessment of a risk situation based on experience or belief. According to Slovic (1992), on the psychometric paradigm approach, the risk is assessed subjectively inside the mind influenced by psychological, social, institutional, and cultural factors. According to Hillson and Murray-Webster (2005), an assessment of the likelihood, environmental conditions, and scale of change determine the effectiveness of risk depending on how well people perceive change and its impact when things are different from what was expected.
When faced with the same decision-making situation, several people will make different decisions depending on each person's perception and understanding of the risks and impacts. Some people may feel very uncomfortable with uncertainty and tend to avoid, minimize threats and take advantage of opportunities to avoid uncertainty. Some individuals may feel comfortable with uncertainty and perceive it as acceptable and thus have no desire to avoid threats. Some people are quite able to tolerate uncertainty yet have less desire to respond to it. At the same time, others may feel uncomfortable with uncertainty in the long term, so they take short-term actions to acquire long-term results. Previous research conducted by Yoo et al. (2018) found that risk perception is one factor that influences people's behavioral interests. This is supported by research conducted by Khasawneh and Alfandi (2019), which conclude the same findings. As a result of this study, the following hypotheses were formulated: H2: Perceived risk has an influence on people's behavioural interest in the #BerubahUsirWabah campaign.

Behavioral Intention
Understanding public behavior will facilitate management and Public Relations practitioners to develop their products or services according to the needs and desires of consumers. Consumer behavior is often based on the possibility of actions to be taken by consumers. According to Jogiyanto (2007), behavioral intention is a desire experienced by an individual to perform a specific behavior. In addition, according to Jogiyanto (2007, p.11), behavior is the actions or reactions of an object or organism. Behavior can be conscious or unconscious, overt or covert, voluntary or involuntary. Humans often consider why certain things are done during day-to-day activities, even consider how those activities impact others. This is known as general behavior or expected behavior. Humans evaluate their actions against the standards of social norms and regulate them using social control. Behavior is the operationalization and actualization of an individual's or a group's attitude towards a situation and environmental condition in terms of natural, community, technology, or organizational.
In contrast to intentions which are still in the form of desires that have not been realized in action, behavior can be seen in real life. Notoatmodjo (2007) explains Snehandu B. Kar's theory that behavior intention is an individual's intention to behave about health care. The behavior to be carried out is the behavioral intention (intention) which is the intention (intention) to take regular health actions, where there is an increasing probability of taking the health action, Conner and Norman (2005). Intention (intention) is a collection of beliefs that can be called intention. Previous research conducted by Sar and Anghelcev (2012) found that risk perception is a variable capable of mediating existing relationships. The more negative a person's mood, the higher the perceived risk of contracting the disease mentioned in the message. For this reason, from the results of this study, the following hypotheses were formulated: H3: The message of the #BerubahUsirWabah campaign has an influence on perceived risk.
H4: Perceived risk mediates the effect of campaign messages on people's behavioral interest in the Change Banish Outbreak campaign.

Research & Methodology
The research approach used in this study is a quantitative method with a causality research approach known as explanatory research to identify the extent and nature of cause-and-effect relationships. The population in this study is the people of Jakarta who are aware of the Change Out of Plague campaign carried out by the Ministry of Health, USAID, and PT XL Axiata Tbk, where this condition makes the total population unknown. Hair et al. (2014) suggests that the general rule of sampling technique for an infinite population is that the sample quantity must be aligned with the number of question indicators applied to the questionnaire using the assumption of observed variable x 10 (indicator) so that in this study it was found that the minimum sample in this study is 170 (17 x 10). This study uses a purposive technique because not all samples have criteria that match the phenomenon under study. The considerations in this study are (1) People who have seen, read, heard of the #BerubahUsirWabah campaign at least twice (2) Are residents of the city of DKI Jakarta. With this descriptive research using the IBM SPSS version 26 program, various information will be collected in answering questions regarding each related variable.
To measure each variable in this study, the Campaign Message variable (X) was measured by 8 statement indicators adapted from the dimensions put forward by Cutlip Center and Broom (in Mukarom & Laksana 2019), namely courtesy, concreteness, completeness, correctness, conciseness, clarity, and consideration. For the second variable, Perceived Risk (Z) is measured by 5 statement indicators with dimensions adopted from Beneke et al. (2012), namely operational risk, financial risk, physical risk, psychological risk, and social risk, while the variable (Y) namely Behavioral Intention is measured by 3 statement indicators adopted from Machrus and Purwono (2010).

Figure 3: Framework
In this study, there is one independent variable, namely campaign message (X), and one dependent variable, namely Behavioral Intention (Y), with perceived risk (Z) as the intervening variable. Furthermore, these variables will be tested further to prove the hypothesis built in this study, namely to determine the effect of the campaign message on the Change to Banish Outbreak campaign on behavioral intention (Behavioral Intention) with perceived risk as an intervening variable.

Characteristics of Research Respondents
Respondents in this study used a purposive sampling technique where the research was only conducted on respondents who had seen, read, and heard the #BerubahUsirWabah campaign with a minimum of 2 conditions and were residents of the city of DKI Jakarta. The total number of respondents willing to answer a series of questions through the google form service is 201 respondents. The number of observed variables (indicators) in this study amounted to 17 indicators, based on the theory of Hair et al. (2014), then a minimum sample size of 170 samples is required. It can be concluded that the adequacy of the sample in this study has been met by obtaining 201 respondents. The characteristics that can be collected from the respondents are gender and domicile with several screening questions.
Based on the results of data processing, from a total of 200 respondents, researchers got 201 respondents, or as many as 100% who met the criteria for the first sample, namely residents of DKI Jakarta and who knew the #BerubahUsirWabah campaign as many as 170 respondents with a percentage of 85% were out of a total of 170 respondents whom there are dominated by those who have seen, read, and heard only two times in the last six months with a total of 155 respondents. The remaining 15 respondents have seen, read, and heard three times. The results showed that most respondents were aware of the #BerubahUsirWabah campaign from Youtube with a total of 101 respondents, then from Instagram with a total of 21 respondents, then Facebook with a total of 18 respondents, and as many as 30 respondents who answered others. Furthermore, as many as 58 respondents were male, and 112 respondents were female. So it can be concluded that women dominated the respondents in this study. This is by data conducted by the Central Statistics   The majority of 53 respondents live in the South Jakarta area, 51 respondents live in the Central Jakarta area, 20 respondents live in the North Jakarta area, 21 respondents live in the West Jakarta area, and 25 respondents live in the East Jakarta area. Isnawa Adji's statement as the acting Mayor of South Jakarta, who stated that South Jakarta was in the fifth position with the highest active COVID-19 cases in Indonesia at 4,238 active COVID-19 cases as of February 26, 2021 (Prodjo, 2021). Further findings showed that of the 170 respondents, the majority of 94 respondents were aged 21 to 25 years, then 39 respondents were aged 26 to 30 years, then 21 respondents were under 20 years old, and 16 respondents were over 30 years old. This is because the use of the internet is more dominated by millennials, where according to data from APJII in 2018 (in Haryanto, 2019), millennials dominate the top position of internet use in Indonesia, namely 20-24 years with a penetration of 88.5%.
Findings from the characteristics of the work found that as many as 100 other respondents answered which criteria were not provided in the statement, then 29 respondents worked as private employees, then 25 respondents worked as entrepreneurs, and 16 respondents as civil servants. This is supported by data from the Central Statistics Agency (2020), which reveals that DKI Jakarta residents have various professions in the fields of agriculture, plantations and forestry; Mining and excavation; Processing industry; Electricity and gas; Water, garbage, waste, recycling; Construction; Trading; Transportation and warehousing; Accommodation and eating and drinking; Information and communication; financial and insurance services; Real estate; Company services; Government administration; Education services; Health services; Other services (Central Bureau of Statistics, 2020).

Research Data Analysis
In analyzing the effect of mediation, this study uses three forms of simple linear regression equations. The first simple linear regression equation analyzed the relationship between campaign messages (X) and Behavioral Intention (Y). Meanwhile, the second simple linear regression equation was used to analyze the relationship between perceived risk (Z) and Behavioral Intention (Y), and the third simple linear regression equation was used to analyze the relationship between campaign messages (X) and perceived risk (Z). Data analysis in this study begins by looking at the quality of the research data by testing the classical assumptions. According to Noor (2017), the normality test is a test carried out to see and find out whether the data received comes from a normally distributed population or not. The normality test of this study uses the help of the IBM SPSS Statistics 26 program; if the collection of points approaches a straight line on the normal-probability plot test, then the assumption of normality can be obtained (Kadir, 2015). Based on the picture above, it can be seen that the points approach the diagonal straight line, which shows the data is normally distributed, so the regression model fulfills the assumption of normality. In addition to using the average probability plot graph, the Kolmogorov Smirnov test can also detect normality detection. Based on the Kolmogorov Smirnov test results, it was found that the three existing regressions had a significance value above 0.05, then Ho was accepted. Therefore it could be concluded that the residuals in this study met the assumption that the data were normally distributed because the significance value was > 0.05. The next test is done by looking at the value of the coefficient of determination (R 2 ) which measures how far the model's ability to explain the variation of the dependent variable is. Based on the findings, it is known that the coefficient of determination (R Square R 2 ) in the first regression is 45.4%, which means that the contribution of Behavioral Intention can be explained by the independent variable, namely campaign messages. So the remaining 54.6% is explained by other variables or factors not examined in this study. Meanwhile, the coefficient of determination (R Square R 2 ) in the second regression is 74.3%, which means that the contribution of Behavioral Intention can be explained by the independent variable, namely perceived risk. So the remaining 25.7% is explained by other variables or factors not examined in this study. In the third regression, the coefficient of determination (R Square R 2 ) is 62.3%, which means the contribution of perceived risk can be explained by the independent variable, namely campaign messages. So the remaining 37.7% is explained by other variables or factors not examined in this study. Based on the results of regression testing, the campaign message on Behavioral Intention was found to have a positive and significant effect, so it can be concluded that the hypothesis (H1) is accepted because it has a value of 0.273 which means it has a positive relationship, at the value of 11.823 (>1.974) which means it has a positive relationship. Significant and the probability value of sig. of 0.000, which means Ha is accepted. In the second regression, perceived risk on Behavioral Intention was found to have a positive and significant effect, so it can be concluded that the hypothesis (H2) is accepted because it has a value of 0.625 which means it has a positive relationship, t value of 22,028 (>1,974) which means it has a significant relationship. Significant and the probability value of sig. of 0.000, which means Ha is accepted. Meanwhile, in the third regression, it was found that the influence of campaign messages as the X variable had a positive and significant influence on the perceived risk variable as the Z variable, so it can be concluded that Normal P-Plot of Reg I Normal P-Plot of Reg II Normal P-Plot of Reg III the hypothesis (H3) is accepted because it has a value of 0.441 which means it has a positive relationship, while The result of the ttest obtained is 16,678, which is greater than the t-table (1.974). Then the significant value generated in the X variable, namely (campaign message), to the Z variable (perceived risk) is 0.000, which means that the significant value is smaller than the probability value that has been set, which is 0.05. From the analysis results above, the researcher concludes that the higher the campaign message that is packaged with a risk communication approach, the higher the perception of risk felt by the community. Thus, the Ministry of Health's strategy to change people's behavior by using a COVID-19 risk communication strategy is said to be able to make the public aware so that they can discipline themselves based on the health risks that are given when individuals do not comply with health protocols and policies that the government has set. The results of the estimated changes in the coefficients of the indirect influence path in this model will be further analyzed using the Variance Accounted For (VAF) method according to Hair et al. (2014) as follows: Based on the calculation of VAF, found to mediate partially with a magnitude of 50% (Partial). However, when referring to the theory development carried out by Hair et al. (2017), perceived risk was found to have a complimentary mediation mediating effect, which means that whether there is a direct or indirect relationship, the campaign message can influence behavioral intention and perceived risk is an essential variable between the two variables. Research conducted by Yoo, Kim supports these findings, and Lee (2018), which found that risk perception is one of the factors that influence people's behavioral interests; this is supported by research conducted by Khasawneh and Alfandi (2019) and Chao (2019), who found the same results.

Conclusion
Based on the results of data calculations and analysis conducted regarding the effect of campaign messages on Behavioral Intention through perceived risk as a mediating variable, the research can be concluded that all hypotheses are accepted where the campaign message on Behavioral Intention was found to have a positive and significant influence. These findings are supported by research conducted by Kemal and Ernungtyas (2020), which found that campaign messages are one of the factors that influence people's behavioral interests. In The Theory of Reasoned Action (TRA), Behavioral Intention can be influenced by several factors, one of which is campaign messages, as evidenced by research conducted by Kemal and Ernungtyas (2020). However, the campaign message is not the main factor influencing Behavioral Intention, where the main factor forming the intention is subjective attitudes and norms. The study results show that the higher the intensity of the campaign messages conveyed, the higher the public's interest in behavior to be aware of the potential risks of the COVID-19 outbreak so that it can change people's behavior to be healthier. The Ministry of Health's strategy to change people's behavior is considered successful by giving messages through the #BerubahUsirWabah campaign, and people gain knowledge that they previously did not know about health behavior patterns related to the spread of the Covid-19 virus.
The results in this study found that perceived risk to Behavioral Intention has a positive and significant influence. These findings are supported by research conducted by Yoo, Kim, and Lee (2018), which found that perceived risk influences people's behavioral interests. In addition, research conducted by Khasawneh and Alfandi (2019) and Sohn, Lee, and Yoon (2016) found the same findings as this study. In the theory of health belief model, perceived risk is one of the main factors to form intentions in health behavior, where when someone knows and perceives the risks that exist, it will cause a solid intention to perform health behaviors in order to avoid the risks they accept and perceive. . From the results of the analysis above, the researcher concludes that the higher the risk perception felt by the community, the higher the public's interest in behavior for a healthy life. The Ministry of Health's strategy to change people's behavior by using a strategy to communicate risks is considered successful. People gain knowledge that they previously did not know about non-preventive behavior and their consequences when they suffer from Covid-19. Campaign messages were also found to have a positive and significant influence on perceived risk, where the researcher concluded that the higher the campaign message that was packaged with a risk communication approach, the higher the perceived risk perception by the public. Thus, the Ministry of Health's strategy to change people's behavior by using a COVID-19 risk communication strategy is said to be able to make the public aware so that they can discipline themselves based on the health risks that are given when individuals do not comply with health protocols and policies that the government has set.
In theory, TRA and Health Beliefs themselves have not specifically discussed the role of mediation in detail, but previous research conducted by Yoo, Kim, and Lee (2018) succeeded in combining two theories, TRA and Health Beliefs, to find a mediating role. In this study, perceived risk has a role or effect of partial mediation or partial mediation considering that there is no change in the direct and indirect relationship where if previously it was known that campaign messages and Behavioral Intentions had a positive and significant influence, the effect was found to be no change if, through perceived risk, campaign message and Behavioral Intention have a positive and significant influence. From these results, it can be concluded that the presence or absence of perceived risk in the relationship between campaign messages and Behavioral Intentions has an effect and is a connecting variable between existing variables. For this reason, the researcher can conclude that the provision of campaign messages containing knowledge of the risks of spreading Covid-19 is considered very effective because, with the knowledge of the risks received, the public will better understand and change patterns of healthy living behavior by complying with health protocols. This study has limited research that can be used as suggestions to increase the number of respondents and widen the coverage area outside DKI Jakarta, considering the #BerubahUsirWabah campaign is given to people throughout Indonesia. By adding the scope of respondents, it is expected to find other findings that were not found in this study.
Author Contributions: Conceptualization, MAR., FKR., MHZ., YRP.; methodology, MAR., FKR., MHZ., YRP.; Data Collection, MAR., FKR., MHZ., YRP.; formal analysis, MAR., FKR., MHZ., YRP.; writing-original draft preparation, MAR., FKR., MHZ., YRP.; writing-review and editing, MAR., FKR., MHZ., YRP.; Author has read and agreed to the published the final version of the manuscript. Institutional Review Board Statement: Ethical review and approval were waived for this study, due to that the research does not deal with vulnerable groups or sensitive issues. Data Availability Statement: The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy.

Conflicts of Interest:
The author declares no conflict of interest.