Coronavirus disease (COVID-19) poses a major health crisis globally (1). With the implementation of COVID-19 vaccination, it is expected that the risk of infection, morbidity and mortality will be reduced (2, 3). However, there is a significant disparity in the distribution of vaccines across countries (4, 5). This is further hampered by the slow uptake of the vaccine by the community (general public) in majority of countries due to uncertainty about side effects and complications following vaccination (6, 7). The continued surge of the infection that has been seen in numerous countries and regions is ascribed partly to the slow vaccination rate (8, 9) and partly to the emergence of COVID-19 variants at the community level (10). This is the apparent situation locally which is evidenced by the third wave with an unprecedented increase in the number of cases reported daily, and the increasing number of deaths (11). Similar dire situations are evident elsewhere in South Asia, South East Asia, South America and the African continent (11).

The challenges faced by the Malaysian authorities is how to fine-tune the balancing act of applying the movement control order (MCO) to restrict movement of people on the one hand, and allowing businesses to remain open to protect the economy on the other (12). Further, even with clear guidelines on standard operating procedures (SOPs) and implementation of different levels of MCO, the preventive measures such as wearing a face mask when leaving the home seems to be lacking locally (51.2%) (13), causing an unstable pattern of daily new cases (14).

In the university setting, staff and students alike face numerous challenges including the change in teaching and learning modes, implementation and observance of strict measures when required to be on campus, not to mention the lifestyle adjustments needed to mitigate the risk of COVID-19 infection and managing the resultant mental stress. How well the university community adjusts to this existential challenge is likely to be dependent on numerous individual and environmental factors.

Numerous studies conducted among the Asian population had pointed to some of these factors including knowledge, perception and practice in relation to COVID-19. Overall, with respect to the university community, the level of knowledge was found to be related to educational level (15), nature of courses pursued (15–17) and gender (17, 18). Further, it appeared that COVID-19 knowledge was positively correlated with risk perception and preventive behavior (19).

With respect to behavior, specifically pertaining to the use of preventive measures to minimize the risk of infection, the results are quite varied and concerning. For example, the use of face masks, an essential practice particularly in crowded places, has been reported to be <20% among university students in Pakistan (20), while that in Indonesia was quite high at 86.9% (17). Likewise, the findings for other key preventive measures such as hand hygiene (16, 20–22) and social distancing (18, 21, 23) are also quite variable.

A study from Indonesia reported that predictors of good preventive behavior were male gender and being a medical student (17) while another study from Egypt found that female gender and high perceived risk of infection were the predictors (18). However, risk perception was found to be inversely correlated with preventive measures among medical students in Iran (24) highlighting the differences in results from studies across the region. There are relatively few studies on COVID-19 risk perception among the Asian population, one of which recorded that 27.3% of Indian university students consider that they have a high probability of getting COVID-19 (22). In another study among Chinese university students, it was found that good knowledge, female gender and being non-medical students were predictor of high risk- perception for COVID-19 non-medical students and those with good knowledge in COVID-19 (25).

Based on the literature review, it is apparent that the predictors of knowledge on COVID-19 and the practice of preventive behavior among the university community is quite variable across the region. This could partly be due to difference in local culture across different nationalities as well as difference in the success or lack of in managing the infection in different countries. In addition, there is no single study that provides a comprehensive analysis on predictors of COVID-19 knowledge, preventive behavior, and risk perception. Nevertheless, with the available information, we hypothesize that the factors mentioned in the preceding paragraph, could have a significant association with COVID-19 knowledge, perception and behavior. Hence, this study provides an overall assessment of these parameters and their associated factors, in order to provide a clearer overview of the status among our university community.

A total of 434 responses were received comprising 93 (21.4%) from staff members and 341 (78.6%) from students. The characteristics of the respondents are presented in Table 1. The age among the staff is 36.6 ± 12.1 and that among students is 21.6 ± 2.1 (mean ± SD). Overall, females make up 58.4% of the respondents. Among the staffs, slightly under half (47.3%) are married, 78.5% have attained higher level education and a minority (6.5%) live alone. Among students, the majority (88.5%) are pursuing an undergraduate degree; all except 3 are single; the large majority (97.7%) live in households of 2 or more members.

Overall, the large majority of the respondents do not have any chronic illness; With regards to living arrangement, 32.3% live with elderly members and 35% with dependents under age 18; 320 (73.7%) reside in high-risk or red zones. Lastly, the large majority are satisfied with the support services provided and are compliant with the public health measures recommended, albeit to varying degrees. The distribution of individual socio-demographic variables in both groups are statistical different in terms of age, marital status, and household size including the respondent (p < 0.05).

Table 2 is a summary of the outcome measurements. Overall, there is no statistical difference in the total score of knowledge on COVID-19, of self-risk perception, preparedness & perceived self-efficacy, preventive (own) measures, as well as unwanted and desirable behaviors between staffs and students. The descriptive analysis for each outcome measure is detailed in the subsequent sections. Subgroup analysis by gender for the following assessment was shown in (Supplementary Table S8).

The crisis caused by the COVID-19 virus has far-reaching effects in the field of higher education. In Malaysia, schools and institutes of higher learning were closed for varying duration during the first and second waves of the pandemic, with short breaks in between, spanning a period of almost 12 months at the time of this survey. This necessitated the use of on-line modes for communication, meetings and teaching-learning activities. All categories of staffs and all students were required to adapt fast to this change. During the short periods of relaxation of the movement restriction between the two waves, staffs and students were allowed on campus, in smaller numbers, with strict observation of public health measures to minimize the risk of infection. Hence, it is very important that everyone is knowledgeable about the infection and familiar with the preventive measures. This survey was conducted with the objective of determining the level of knowledge, the preventive measures practiced, COVID-19 risk perception, preparedness and perceived self-efficacy, and behavior among staffs and students, and their inter-relationships.

We found that the level of knowledge on COVID-19 was quite high among the respondents, with an average score of 18.7/22 (85.1%). While the knowledge score was good overall, there were some gaps with respect to the less frequent symptoms of COVID-19. More concerning is the misconception that antibiotics could be an effective way to prevent COVID-19 among 47.9% of the respondents, a finding similar to that found in a study which reported that as high as 66.4% of Australians believe that taking antibiotics regularly could possibly prevent COVID-19 (34). The proliferation of information over social media, including misinformation, could be a factor in this observation as well as in promoting risky behavior (35, 36).

The predictors of higher knowledge score included older age and residence within red (high-risk) zones. The association between older age and knowledge score is not surprising within the context of our study, as over half of the staff, who are obviously older than students, have postgraduate and/or professional degrees. The positive association between age and knowledge is corroborated by several studies (37–40). Further, Olaimet et al. (15) and Habatu et al. (16) have demonstrated that students engaged in postgraduate studies scored significantly higher compared to undergraduate students, providing support to the plausible association between educational level and knowledge. The association with location of residence can be understood from the perspective of the increased risk of infection if the residence is within a zone of high transmissibility of the SARS-CoV-2; hence the existential risk and the need to be more vigilant. This is likely to lead to more attention being paid to information related to the infection and its spread.

Other factors that have been reported to be associated with the level of knowledge include gender (16–18), and specifically in the case of students, the nature of the course taken (15–17). The association with gender is unclear as reported results are quite varied. We did not find a significant association between gender and knowledge similar to study by Olaimet et al. (15). We did, however, find significant association between knowledge and being in the medical course among the students, similar to the results from the above quoted studies.

The use of preventive measures against COVID-109 spread was relatively high among our respondents (overall score of 8.71/10). Specifically with regards to the use of face masks, the score was indeed 100%. In comparison, in an earlier study on the local population, the use of face masks was only 51.2% (13). This difference could be due to various factors, the most important being the fact that the quoted study was done in the early phase of the local outbreak, at which point in time, the reported daily number of cases were much smaller and presumably, the perceived risk of infection lower among the people, not forgetting the fact that face masks were not mandatory then. Data on the use of face masks as a preventive measure from other countries is also rather varied ranging from under 20% (20) to over 85% (17). Factors, other than that already mentioned, such as socio-economic and cultural difference could also account for such diversity of the reported data.

We found a significant association between good preventive behaviors and good level of knowledge. It is logical to assume that good knowledge of the disease transmission would promote the use of preventive measures; hence, the finding of similar results in several other studies (41–44). The other predictor for better preventive measure score was female gender, which was also seen in previous studies conducted in China, Turkey and Saudi Arabia (41, 45, 46). Such a finding could indicate that females tend to be more concerned about the infection and thus more likely to comply with safety measures (47). A meta-analysis published in 2016 in the context of respiratory epidemics such as influenza, found that females are more likely to practice non-pharmacological protective behavior than men (48).

It may be relevant to note that certain precautionary measures, including exercising caution when opening parcels and disinfecting surfaces, are practiced less frequently among the respondents. However, there is no consensus regarding the transmission of coronavirus through inanimate surfaces. Some studies have shown that the SARS-CoV-2 was detectable on plastic up to 5 days, and on paper between 3 h to 5 days; the variation of survivability of the coronavirus may be due to different environmental conditions such as temperature or humidity (49, 50). However, it was pointed out that the chance of transmission through such means is high only if an infected person coughs or sneeze on the surfaces and someone else touches that surface soon after that (within 1–2 h), thereby making this mode of transmission less likely (51).

Self-risk perception was analyzed from three aspects, the probability of being infected, the susceptibility to infection and the severity of illness if infected. Overall, the perception of high risk for disease contagion, susceptibility and severity was relatively low at 6.7, 7.6, and 9.9% respectively. In early 2020, an investigation of the risk perception among the public across several countries in Europe, United States of America and Asia was carried out by Dryhurst et al. (52). It was reported that the perceived risk, which varied from country to country, was nevertheless relatively high. The authors highlighted that overall, personal experience with the virus, prosocial values, and amplification about the infection through family and social contacts significantly influenced risk perception in more than half the countries studied. Although the risk perception among the participants in this study is comparatively low, the notable finding of a positive association between risk perception and presence of COVID-19 cases within social contacts, was similar to the observation of Dryhurst S et.al. In this paper, it was also pointed out that risk perception correlated significantly with reported adoption of preventative health behaviors in all the countries studied. We did not, however, find such an association. Another finding in our study is a positive association of higher self-risk perception with having elderly family members, which is not unexpected as older people who are at much higher risk of being infected could pose an increased risk to the entire family.

A negative association between perception of high self-risk and chronic medical illness was observed in our respondents. However, it is noted that in our study the number of participants with chronic illness were very small, which may negate the relevance of this finding. We note that the effect of medical comorbidities on risk perception has been mentioned in studies by Yan et al. (53), He et al. (54) and Laires et al. (55). In these studies, contrary to our observation, a positive association was found between chronic illness and risk perception. This is not unexpected in view of the widespread dissemination, through various media, about the increased risk for COVID-19 among people with chronic illnesses. Lastly, we also did not find any association between gender and self-risk perception in contrast to results from other studies (56, 57).

Preparedness and perceived self-efficacy were positively associated with the living in households with children and negatively associated with the presence of COVID-19 cases among social contacts. It is generally believed that children are at lower risk of infection compared to adults, with reported prevalence ranging from 1.2% in Italy (58) to 2% in China (59). Further, they are at much lower odds of being infected (pooled odds ratio = 0.56, 95% CI = 0.37–0.85) compared to adults (60). Nevertheless, owing to the relative lack of information on the long-term effects of COVID-19, it has been recommended that children should not be exposed to people outside the household and any individual who is unwell (61), highlighting the risk, albeit not as high as among adults. The negative association with the presence of COVID-9 among social contacts is rather more difficult to explain, as the opposite would seem to be more likely.

Overall, unwanted behavior was found to be significantly associated with the male gender and COVID-19 positive status. Knowledge on COVID-19 and being married were the predictors of desirable behaviors, a finding that is consistent with another study (62). This is not unexpected as married people, especially those with children and/or are staying with elderly parents would more likely adopt better preventive measures and behavior, within the context of local cultural norms – being more prosocial than individualistic.

Out of the six items on unwanted behavior, three were about lifestyle which included physical activity, alcohol consumption and unhealthy dietary intake. The fourth was on stereotyping people who are considered to be high risk individuals and the remaining two were on health literacy. In terms of lifestyle, males apparently tended to adopt more unwanted behavior, in particular with respect to diet and alcohol consumption (63). In the present study, we observe that males exercised less (p = ≤ 0.001) and consumed alcohol more (p < 0.016); however, the difference in terms of unhealthy diet was not significant (p = 0.117). With respect to stereotyping, we found that females tended to do this more (p = 0.005). In a study on Malaysians, Jaafar et al. reported that females have better sufficiency in health literacy than their male counterparts (64), a result that was consistent with the finding from other countries (65–68). We did not, however, observe this in the present study.

Regarding the confirmed diagnosis of COVID-19 as one of the predictors for unwanted behaviors, we note that there was only one person with a history of COVID-19 infection, a 44 years old male, with no history of chronic medical illness, lived in the green zone and with no known COVID-19 among his immediate social contacts. This person had affirmed that he had practiced all the unwanted behaviors.

The findings of this research could inform relevant management staff and policy makers of the university with regards to the gaps in knowledge, the perceptions and behavior of the staff and students related to the COVID-19 pandemic, thereby allowing adjustments to existing measures and policies with the aim to improve upon them if deemed necessary. As the pandemic is still evolving locally, and transmission has accelerated with the appearance of highly infections virus variants, measures must be continually monitored and adjusted with the changing landscape. While the results from this study may not be generalizable, it nevertheless provides a glimpse into the effects of the pandemic on a typical university community and provides a reference point for future studies.

The limitation of this study lies in the fact that it only involved the university community, and only one university. In addition, due to low response rate and sampling method used, sampling bias or non-response bias is expected. Therefore, the findings of this study cannot be generalized to the whole Universiti Tunku Abdul Rahman community or to university communities at large. Further, being a cross-sectional and largely descriptive study, no causal associations can be drawn from the results. Due to the dynamic nature of the pandemic, these results should be taken in context.

COVID-19 pandemic is a major challenge to all aspects of living. This study showed that the majority of the university community had a good level of knowledge achieving an overall knowledge score of 18.7 per out of 22. Likewise, preventive action with an overall score of 8.7 out of 10. The perceived self-risk was relatively low implying that the respondents considered themselves to be quite safe. This could be related to their high level of preventive action. However, a low level of risk perception could lead to lowering the guard against the infection. Hence the needs for constant and timely update to inform and remind via social media, television and radio regarding the deadly nature of the infection is essential to aid in efforts to fight COVID-19 despite the introduction of vaccination.

Raw data supporting the conclusion can be obtained from corresponding author on special request.

The studies involving human participants were reviewed and approved by Universiti Tunku Abdul Rahamn. The patients/participants provided their written informed consent to participate in this study.

KWL, SFY, HTO, PPL, NMH, and MSL: conceptualization, investigation, and methodology. MSL: data curation. KWL, SFY, and MSL: formal analysis. SFY: funding acquisition. HTO, PPL, and NMH: project administration and resources. KWL: software. SFY and MSL: supervision. HTO, PPL, and MSL: validation. KWL: visualization. KWL, SFY, and MSL: writing—original draft and writing—review and editing. All authors contributed to the article and approved the submitted version.

This research was funded by Universiti Tunku Abdul Rahman, Grant Number: IPSR/RMC/UTARRF/2020-C2/Y01.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.