Community preventive behaviour and perception on the severity of COVID-19 disease in Indonesia, 2021-2022: Structural equation modelling

Introduction

The world is currently besieged by the COVID-19 pandemic.¹ As of February 7, 2022, there were 394,381,395 confirmed cases of COVID-19, with the World Health Organization (WHO) reporting 5,735,179 deaths.² The COVID-19 pandemic in Indonesia from mid-2021 until the end of 2021 reached its peak (second wave) when it was dominated by the Delta variant with the total confirmed cases was 4.254.443 and 143.766 death (CFR: 3.4).³

COVID-19 prevention regulations largely depend on community compliance and behavior. While changing behavior is a significant challenge in health interventions. An AC Nielsen survey (2020) in six major cities with a total of 2,000 respondents, in collaboration with the United Nations International Children’s Emergency Fund (UNICEF), found that less than one-third (31.5%) of all respondents practiced all three-preventive activities (wearing masks, washing hands, and social distancing). Over one-third (36%) practiced only two of the three-preventive behaviors, less than one-fourth (23.2%) practiced only one of the three behaviors, and almost one-tenth (9.3%) did nothing at all.³ A number of studies described the associations between socio-demographic characteristics and people’s levels of perceptions about the severity of COVID-19 with their preventive behavior.^3–10

Based on the official government report (see https://covid19.go.id/monitoring-kepatuhan-protokol-kesehatan), some areas of Indonesia showed a compliance rate of only 60% till I January 2022.¹¹ The aim of this study was to examine changes in community behaviour to prevent local COVID-19 transmissions and changes in community perceptions about the severity level of COVID-19. The following research questions were formulated: 1) What constitutes as the community perceptions of COVID-19? 2) Is there an association between community perception and community preventive behaviour using the Health Belief Model (HBM) approach? And 3) what kind of model of preventive behaviour on COVID-19 can be predicted using the structural equation model (SEM)?

Methods

Community preventive behavior against COVID-19 transmission

Community preventive behavior was represented by six indicators in Figure 1 which include frequency of hosting (BHV1), frequency of visiting others (BHV2), frequency of work/study from office/school (BHV3), frequency of handshaking (BHV4), frequency travelling to a red zone (BHV5), and frequency of leaving the house when you are not feeling well (BHV6).

Figure 1. Proposed structural model of community preventive behavior against COVID-19.¹⁵

Results

The study results are presented in several parts sequentially, starting with the respondent’s characteristics, followed by the descriptive statistics of the independent variable (community perception of the level of severity of COVID-19 disease), the descriptive statistics of the dependent variable (composite variable of community behavior), and the SEM analysis results.

Most of the respondents lived in the city (46.1%) and a housing area (36.1%), and only a few lived-in villages (16.6%). A large majority of the respondents were women (74.5%), and approximately 80.0% were students and workers (Table 2).

In terms of perceived susceptibility (Table 3), more than half of respondents (>50%) chose to strongly agree to practice recommended behaviors such as social distancing at least 1-2 meters in public areas; almost half of them strongly agree that they had a low risk of getting COVID-19 when avoiding using public transportation, practicing work/study from home, handshaking, and travelling to a red zone. Most respondents also strongly agree that if they are exposed to COVID-19, it will affect their health if they do not practice recommended behaviors (perceived severity - Table 3). Likewise, concerning perceived benefit (Table 3), more than 50% of respondents strongly agree that they can avoid getting COVID-19 if they practice these behaviors. In terms of perceived barriers, less than 20% felt that it would be challenging/difficult to implement the recommended behaviors (Table 3). Regarding self-efficacy, this is defined as the conviction that one can successfully practice a certain behavior,¹³ the survey showed that less than 50% were confident that they could implement the recommended behavior (Table 3).

For each type of recommended behavior as mentioned in Table 1 to prevent the transmission of COVID-19, more than 50% of respondents never and rarely hosted, visited others, worked from the office/school, shook hands, travelled to a red zone, and left the house when not feeling well (Figure 2).

Figure 2. The description of behavior.

Notes: 1: Never; 2: rarely; 3: occasionally, 4: frequently, 5: always.

Based on the composite of each perceived item, it was found that only 16% of the respondents had a high percentage of good perceived susceptibility; perceived severity was 43%, perceived benefit was 54%, perceived barrier was 3%, self-efficacy was 19%, and only 60% of respondents practiced good behaviours.

SEM results

Structural model validity

Two analyses were conducted to evaluate the structural model validity:1) dependency test and 2) assessing the goodness-of-fit of the model.

Discussion

The results of categorized all perceived items revealed that only a small proportion of respondents (16%) held beliefs about their chances of experiencing COVID-19 (perceived susceptibility), while 43% of respondents believed in the severity of the COVID-19 effects were on their health. More than half of the respondents believed that the recommended behaviors to prevent COVID-19 infection could protect them from getting an infection (perceived benefit). Only 3% of respondents believed some things hindered the practice of recommended behaviors (perceive barrier), and only 19% believed that they could practice the recommended behaviors (perceive self-efficacy). This study’s results are similar to those conducted in India, Sri Lanka, Iran, and Ethiopia.^8,10,17,20 A study in Italiy conducted to nine hundred and seventy-eight Italian adolescents found a similar pattern where they had a low perception of COVID-19 risk, as well as perceived comparative susceptibility and perceived seriousness. They think that COVID-19 is not a potentially severe disease for them as many news stated that young people are less vulnerable to the COVID-19 effect.²¹ The HBM theory holds that people are likely to practice preventive behaviors or actions if: 1) They believe that it will reduce their risks,¹³ 2) They perceive themselves as susceptible to COVID-19 infection, 3) They believe that this condition would have a potentially serious impact, 4) They acknowledge the benefit of recommended actions/behaviours in reducing the susceptibility or severity of the virus, and 5) They believe that the anticipated benefits of doing preventive behaviours/actions outweigh the barriers.

Perceived susceptibility was not a significant predictor of behavior in this study. This finding is consistent with studies that measured adherence to COVID-19 precautionary measures in China²² and Korea.^23,24 A study that used a HBM framework to look at how Saudi Arabian students at Jazan University felt about the COVID-19 vaccine found that perceived susceptibility was not a good predictor of how they felt about getting the vaccine.²⁵ However, a similar study conducted in Malaysia found that high perceived susceptibility to COVID-19 infection was also associated with the behavior of vaccination intention.²⁶ A study that measured student behavior in the US related to non-pharmaceutical interventions (hand washing with soap and water, use of hand sanitizer, wearing a face mask in public, and practicing social distancing) found that perceived susceptibility was associated with multiple interventions more frequently.²⁷

A previous study conducted by Du Min et al. found that low perceived risk was associated with vaccine hesitancy.²⁸ As with perceived susceptibility, perceived severity was not a significant predictor of preventive behavior in this study. Several studies regarding behavior change using BHM had similar results.^24,26,27 Perceived severity, on the other hand, was a significant predictor of preventive behaviors.^8,22–24,29 Perceived benefit is one predictor that is not a significant predictor of preventive behavior. This finding is inconsistent with several studies that used BHM to predict behavior change, particularly in relation to COVID-19 prevention^{8,22,23,25–29} where it was found that there was no significant association between perceived benefit and preventive behaviour of.

This study demonstrated that the perceived barrier significantly predicted COVID-19 preventive behaviour. This result is contrasts with behaviour study conducted in Sri Lanka and Iran, which established a significant positive relationship between perceived benefit and self-efficacy in COVID-19 prevention behaviour.^10,30 Nevertheless, these findings were congruent with a study conducted in Ethiopia that employed the HBM theory to assess student eating behaviour the US³¹ and other behavioural studies in Iran, India, and Hong Kong.^8,17,32 This study also found that perceived self-efficacy was a significant predictor of COVID-19 preventive behaviour. The results show that, with lower self-efficacy, people were likely to practice COVID-19 prevention behaviour. This result is similar to several studies that examined behaviours using the BHM theory. Those studies found that perceived self-efficacy has a significant relationship to behaviour.^{10,17,20,23,30,33} In contrast, a study that assessed the student’s behaviour on the non-pharmaceutical intervention of COVID-19 found that perceived self-efficacy was not a significant predictor of behaviour change. In theory, an individual with good self-efficacy tends to practice recommended action/behaviour,¹³ which is the preventive behaviour of COVID-19. However, this study was unable to confirm these findings.

The findings in this study illustrated that most respondents (97%) had no barriers to practising the recommended behaviour. Still most respondents (81%) were not confident that they could fully implement the recommended prevention behaviours. As many as 60% of respondents practised COVID-19 prevention behaviour well. In knowledge attitude practice (KAP) studies conducted in Indonesia, this finding (percentage of good behaviour) tended to be lower than the other two findings in Indonesia, where the rate of those who performed the correct behaviour for the prevention of COVID-19 was more than 90%.^34,35 Studies conducted in other countries also found that respondents who practised COVID-19 prevention behaviours were relatively high (>70%), such as in China, Nepal, Malaysia, Vietnam, and India.^36–40

Fundamentally, perceived susceptibility and severity affect how a person decides to act.¹³ However, most respondents (84%) had low perceived susceptibility in this study. This means that most respondents did not believe that they were also at risk of being affected by COVID-19. This perception represented an obstacle for someone to implement recommended behaviour. It was also known that only 43% believed that if they were infected with COVID-19, they would experience harmful consequences for their bodies. Only half of the respondents believed that the recommended behaviour was able to protect them from COVID-19. This might relate to the information they obtained day-to-day. It was possible that most of them did not have a clear idea about the pathophysiology and epidemiology of COVID-19, made worse by unreliable news or hoaxes circulated on social media, which may have increased negative perceptions about COVID-19.^8,41

Regarding obstacles to implementing the recommended behaviour, a few respondents said it was extremely difficult to implement the behaviours (3%). This overall perception then leads to a low belief that respondents are able to implement the recommended behaviour. As a consequence, it will affect COVID-19 prevention behaviour practice. Perception is theoretically influenced by many factors, including demographics and level of knowledge.^{8,10,13,20,34,35} A study conducted in China found that knowledge was influenced by educational level and domicile.³⁶ Good knowledge can form a good attitude, which then creates a good perceived.^42,43 COVID-19 was a pandemic that touched almost every facet of human existence. People had to adjust their daily routines to accommodate local government rules in order to reduce the virus transmission, and these behavioural shifts may last long after the disease has passed.⁴⁴ Increasing the respondents’ knowledge is essential to narrowing the gap between knowledge and practice, including myths, hoaxes circulating about COVID-19, and facts. Several important factors that may affect perceived self-efficacy are related to social norms and trust⁴⁵ in the community. Unfortunately, neither of these factors (including knowledge) was investigated in this study.

The structural model of COVID-19 prevention behaviour in this study was not intended to describe the empirical phenomenon of preventive behaviour. Byrne stated that, if possible, researchers are advised to modify the model by using modification indices (MI) in SEM testing. He said that models with an MI score of more than 10 deserve attention for modification.⁴⁶ Already it has proved that to confront the increase in demand for care, the need for long-term care workforce, and the costs associated with care.⁴⁷ However, Hair et al. suggested that modifying the model should not change the model’s structure significantly.¹⁹

These several studies found a gap in knowledge to inform changes in policies on infection-prevention measures in the community, community infection procedures, the frequency of testing, etc.^8,41,48 The unreliable information may increase the negative perception of COVID-19, leading to the community’s obedience to suggested preventive behaviors.⁴¹ Moreover, people are more worried about their families and economic conditions due to the spreading pandemic than about complying with lockdown or restriction policies. Several factors that might be related were not examined in this study, such as level of knowledge, social norms, or trust.

Conclusions and recommendation

This study found that more than half of the respondents still had low perceived susceptibility (84%) and severity (67%), but more than half had high perceived benefits (54%). Only a few respondents had significant barriers to implementing COVID-19 transmission prevention behaviours (3%). Still, most respondents had low perceived self-efficacy (81%), and only 60% had good behaviors related to COVID-19 prevention.

It was found that sex had a statistically significant relationship with perceive susceptibility (p-value = 0.000); perceived susceptibility had a statistically significant relationship with perceived severity (p-value = 0.000); age (p-value = 0.004), sex (p-value = 0.030), and education ((p-value = 0.050) had significant relationships with perceived benefit; perceived benefit (p-value = 0.000), resident (p-value = 0.009), and sex (p-value = 0.000) had a significant relationship with perceive barrier; and perceive barrier (p-value = 0.000) and occupancy (p-value = 0.040) had a significant relationship with perceive self-efficacy. Regarding behaviour, only perceived barriers (p-value = 0.000) and self-efficacy (p-value = 0.000) had statistically significant relationships with COVID-19 prevention behaviour. However, the structural model in this study was not fit to the data and was not fit to describe the empirical phenomenon under study.

This study can be an input for public health policy development especially those related to behaviour change interventions/programs. By using HBM theory, policy makers or other stakeholders can consider which stages of behaviour change still require more intervention in addition to demographic factors which can also influence it. Based on HBM theory, for behaviour change to occur, individuals must perceive a threat in their current behaviour (perceived susceptibility and severity), believe that the change will bring meaningful or useful outcomes (perceived benefit), and possess the self-confidence to enact the change (self-efficacy). In this case, this research can be a reminder in terms of making evidence-based policies.

In the context of COVID-19 preventive behaviour change, we recommend to improve perceived susceptibility and severity (since this study found low of perceived and severity) by providing the correct information about COVID-19 in the local cultural context. It is expected that by improving perceived susceptibility and severity, there would be an increase in respondents’ knowledge, increasing perceived susceptibility and severity. The results and concept of this study can also be used/implemented for developing prevention policies against many types of diseases that require community behaviour changes which consider stages of behaviour change.

For further study, it is highly recommended to make inclusion and exclusion criteria prior to the data collection, to create a more rigorous data collection template to reduce selection bias effect, to collect the data onsite and if it is possible, to conduct the longitudinal study. The concept of HBM is one of the recommended theories to study about the health behaviour, hence it is replicable even in the different context of disease and area.

Ethics approval and consent to participate

The Commission approved this study for Research Ethics and Public Health Service, Faculty of Public Health, University of Indonesia Number: Ket-436/UN2.F10. D11/PPM.00.02/2021.

Consent for publication

Informed consent was obtained from all subjects involved in the study.

Author contributions

All authors made substantial contributions to this research and approved the final manuscript. TE and TS contributed to every step of the study (research concept, design, data interpretation, writing, and review). SP contributed to the research review.