Background

COVID-19, a respiratory infection caused by SARS-CoV-2, was first described in late 2019 and identified as an international public health emergency in January 2020 [1,2,3]. Response to COVID-19 included rapid development of protocols for diagnosis, treatment, containment and prevention efforts [2]. A variety of different prevention measures were implemented, including closures, virtual events or telework, social distancing, mask wearing, and surveillance programs [2, 4,5,6,7].

However, effective implementation of the majority of these measures relies on individual behavior change, which can be challenging. Prior to the pandemic, reported rates of compliance with standard infection control prevention policies among hospital staff were reported between 50–70% [8,9,10], and even less than 50% [11]. A recent review of 56 papers found staff compliance with preventive measures was associated with working in emergency or ICU settings as well as high level of risk perception and concern [12]. This raises concern that compliance among individuals in other settings, where adherence would not affect employment, could be even lower.

In order to facilitate better adherence, it is important to understand characteristics correlated with adherence. In studies prior to the COVID-19 pandemic, females were more likely to wear masks in the early stage of the H1N1 influenza epidemic [13] and had a higher perception of risk compared to males [14]. Those with more education were more likely to wear masks in public areas [13]. Non-white participants, compared to white study participants, were more likely to follow protective guidelines [15] or indicate a willingness to get an influenza vaccine [16]. A systematic review of 26 studies during the swine flu pandemic showed that female, older, more educated and non-white participants adopted protective behaviors more than other demographic groups [17].

These findings are largely consistent with studies conducted during the COVID-19 pandemic. Females tended to follow the suggested guidelines (e.g., social distance, hand hygiene practice) more than males, and their perceived risks from COVID-19 were also higher than that of their male counterparts [18,19,20,21]. Additionally, respondents from countries that were severely affected by COVID perceived protective measures as more important than that from countries that were less affected [21, 22].

At the time this research was initiated, there were very few reports of attitudes towards prevention among students, which is a notable gap as student perceptions and adherence to prevention guidelines may vary substantial from other groups. However, since the start of the COVID-19 pandemic, research has explored the rate of students’ compliance with preventive measures [23,24,25,26] as well as attitudes or perceptions towards preventive measures among Iranian high school students [27] or among university students from the Middle East [28, 29], Asia [30, 31], Europe [32, 33], and black college students from North Carolina, USA [34]. Therefore, the purpose of this study is to add to this growing body of literature to explore the perceptions and awareness of COVID-19 prevention measures among students at a large United States university.

Methods

Study design

This was a cross-sectional, online survey of 643 undergraduate and graduate students at Purdue University, Indiana, USA from October 2020 through December 2020. During this time, Purdue University was among the minority of US universities to offer at least some classes in a face-to-face format and a majority of students were physically on campus.

Purdue University developed COVID-19 prevention strategies (“Protect Purdue”); under these guidelines the university operated with both online and hybrid in-person/online instruction during 2020–2021. Briefly, social distancing was required whenever feasible. Cloth mask use was required indoors and when social distancing was not possible outdoors. When social distancing was not possible indoors, face shields were required with masks. These guidelines were applied to both on- and off-campus events. All students were required to have a COVID-19 test prior to arrival on campus and to participate in random surveillance testing throughout the school year. All students, faculty, and staff signed a pledge indicating their willingness to follow these guidelines. Violators could potentially face strict penalties, including suspension [35]. The protocols were consistent throughout the data collection period.

Study population

Eligibility criteria included 1) undergraduate or graduate student status at Purdue University; 2) ≥ 18 years old; 3) completion of informed consent. Individuals were excluded if they did not complete the full survey. Participants provided informed consent prior to completing the survey. This study was determined to be exempt from Institutional Review Board review by the Purdue University Biomedical Institutional Review Board.

Participants were recruited through flyers and direct emails. Flyers posted across campus included a URL and QR code that students could use to access the survey: N = 342 individuals used the URL and N = 121 individuals used the QR code. Direct emails were sent to a random sample of 4000 undergraduate and graduate students (8.7% of enrolled students). Of these, 473/4000 (11.8%) accessed the survey. Altogether, 936 students started the survey (Supplemental Figure S1). Participants were excluded due to not meeting inclusion criteria (N = 119) or not completing the survey (N = 171). An additional 3 participants who indicated nonbinary gender were excluded due to a lack of power to interpret this category. A total of 643 participants (69.0% of those who accessed the survey) were included in analyses.

Study variables

The 21-item online survey covered demographics, student status, experience and perceptions of COVID-19 as well as COVID-19 prevention measures. The survey was developed by the investigators. Demographic/student status variables included gender, race/ethnicity, residency (in-state, out-of-state, or international), class, and college. Seven options were provided for race/ethnicity; however, due to low numbers in some categories this was recoded to white, Asian/Pacific Islander, or other/multiracial for analysis. Awareness/experience variables included “do you know anyone who has tested positive for COVID-19?” and “how concerned are you about COVID-19?”. Participants were also asked whether they agree or disagree with the statement “I am aware of students who are hosting social events that violate the Protect Purdue guidelines”.

Participants were asked to indicate the extent to which they agreed or disagreed with 12 statements related to COVID-19 prevention measures using a 5-point Likert scale (Supplemental Table S1). The Cronbach’s alpha for these 12 items was 0.74. Responses were scored from 1 = “strongly disagree” to 5 = “strongly agree”. Most statements were originally written so that a response of “strongly agree” was consistent with increased prevention measures or belief in the effectiveness of prevention measures. Three statements which did not meet this pattern were recoded for analyses. These 12 items were grouped into five broader categories: participant’s own perception of the effectiveness of prevention measures (“protocol effectiveness (self)”); how others perceive the effectiveness of prevention measures (“protocol effectiveness (others)”; participant’s own adherence to preventive measures (“protocol adherence (self)”); others adherence to preventive measures (“protocol adherence (others)”; and perception of protocol violation consequences (“consequences of protocol violation”).

Data analysis

Analyses were completed using Stata Version 16 (College Station, TX); a p-value < 0.05 was considered to be statistically significant. Descriptive analyses were completed to assess the frequency and distribution of all variables. Linear (unadjusted) regression and multiple (adjusted) linear regression were used to compare predictor variables with perceptions of COVID-19 measures, knowledge of anyone holding social events which violated these measures, and self-reported level of concern about COVID-19. Predictor variables were all coded as categorical variables and included gender, race, residency, college, class and knowledge of someone with COVID-19. Multiple linear regression models included all predictor variables as covariates. Results are presented as change in perception score on a scale of 1 to 5 with a higher score indicating a higher use or belief in preventive measures.

Results

Overall, 643 students were included in analyses (Table 1). Approximately 60% of participants were female and 62% had in-state residency. The majority of participants were white (N = 465, 72.3%), the remainder were Asian or Pacific Islander (N = 101, 16.7%) or other/multiracial (N = 77, 12.0%). The College of Engineering (N = 146, 22.7%) and College of Health and Human Sciences (N = 111, 17.3%) had the most participants. Most of respondents were undergraduate students (N = 455, 70.8%), the remainder were graduate students (N = 188, 29.2%). The majority of participants reported knowing someone who was COVID-19 positive (N = 511, 80.3%). Most students reported being “moderately concerned” about COVID-19 (N = 240, 37.3%), with the remainder roughly equally divided between being less concerned and more concerned. Just under half of the participants reported that they “somewhat agree” or “strongly agree” with the statement that they are aware of social events in violation of prevention policies (N = 305, 48.0%).

Table 1 Population characteristics
Full size table

Mean scores for the questions related to adherence or perception of COVID-19 prevention measures are presented in Table S1 and Fig. 1. “Protocol adherence (self)” (mean score: 4.60; SD: 0.71) and protocol adherence (other) (mean score: 3.91; SD: 0.78) received the first and second highest scores, respectively. Protocol effectiveness for both self (mean score: 3.45; SD: 1.01) and others (mean score: 3.45; SD: 1.09) received the next highest mean scores, and consequences of protocol violation received the lowest scores (mean score: 3.12; SD: 1.11).

Fig. 1
figure 1

Mean scores (dot) and 95% confidence interval (line) for student’s agreement with statements about COVID-19 prevention protocols. Outlined/dashed markers indicate individual questions; black markers indicate scores for the overall category

Full size image

Results from linear (unadjusted) regression models are presented in Supplemental Tables S2 and S3 and were similar to results from multiple linear regression (adjusted) models. All multiple linear regression models included gender, race, residency, college, class and knowledge of someone with COVID-19 as covariates; results are presented in Tables 2–3. Interactions between predictor variables were not evaluated in these models. Scores for protocol effectiveness (self) were significantly higher among women (versus men), Asian/Pacific Islander or other race/multiracial (versus white), out of state (versus in-state), as well as juniors, senior or graduate students (versus freshman) (Table 2). Those who reported having a positive COVID-19 test themselves (versus not knowing anyone who tested positive) had significantly lower scores on protocol effectiveness (self). There were significantly lower scores for protocol effectiveness (others) among the participants who knew friends, “someone else”, or multiple people who tested positive for COVID-19 (versus those who reported none).

Table 2 Adjusted difference in opinions on effectiveness or adherence to prevention protocols associated with selected characteristics
Full size table
Table 3 Adjusted difference in knowledge of violating guidelines or self-reported level of concern about COVID-19 associated with selected characteristics
Full size table

Scores for protocol adherence (self) were significantly higher among graduate students versus freshman in multiple linear regression analysis (Table 2). However, participants who reported having a positive COVID-19 test and multiple categories of known COVID-19 individuals (versus reporting none) had significantly lower scores in protocol adherence (self). For protocol adherence (others), we found significantly lower score in participants who knew family members, friends or multiple categories of known people who tested positive for COVID-19 versus reporting none. Scores for consequences of protocol violation were significantly higher among females (versus males) and Asian/Pacific Islanders (versus white participants); however, they were significantly lower among participants who tested positive for COVID-19 themselves (versus reporting none).

In multiple linear regression analyses, scores were significantly higher for knowledge of violating guidelines for participants in other or multiple colleges (versus engineering), seniors (versus freshman), as well as those reporting that a friend or others/ multiple categories tested positive for COVID-19 (versus those who reported none) (Table 3). Scores for concern about COVID-19 were significantly higher in females (versus males), Asian/ Pacific Islander (versus White), out-of-state (versus in state) as well as juniors, seniors, or graduate students (versus freshman).

Discussion

In this study, we found that characteristics most likely to be correlated with higher belief in or adherence to preventive measures were female, Asian/Pacific Islanders or graduate students. Meanwhile, respondents who reported knowing someone who had COVID-19 tended to report lower belief in or adherence to these preventive measures.

Female participants showed higher belief in or adherence to protective measures compared to their male counterparts. This is consistent with several previous surveys also conducted during the COVID-19 pandemic which report that females followed protective guidelines [18,19,20,21] and had a higher perception of risk compared to males [20]; these trends were also seen among studies conducted in university settings[29, 33]. Consistent with these observations, males have also been reported to practice riskier behaviors than females [36].

Higher level undergraduates and graduate participants showed higher belief and adherence with the preventive measures than the younger students; similar to results from results from a multi-university study from Albaqawi and colleagues [29]. This could reflect differences in age or in educational levels. One study conducted in Germany found that men with more education were more worried about COVID-19 than the men with less education [37]. Level of education has also been found to be a significant factor impacting the knowledge and perceptions of the Nipah Virus in Bangladesh [38]. Several other studies have found correlations of age with taking preventive measures. For example, Luo et al. report that older generation are more likely to take preventive measures than younger generation [39].

Several outcomes were significantly higher among self-reported Asian/Pacific Islanders. This is consistent with prior literature noting that non-white participants were more likely to wear masks [15, 16]. However, there were insufficient data to explore these association with other self-reported racial/ethnic groups in our study.

There were several limitations to this study. First, data were collected solely from Purdue students, so results may not be generalizable to other settings. It is also possible that response bias could have influenced our results. However, many of our key findings were similar to those reported by studies conducted in the general population, which suggests that there is at least some degree of generalizability of our results. Additionally, this analysis is based on self-reported data, but we are not aware of any potential bias that could influence these data. It is possible that there could be some difference in response related to calendar time, as the survey was available over a few months and our knowledge of COVID-19 was rapidly changing throughout this period. However, campus COVID-19 prevention policies on campus did not change over the course of the data collection period.

There are also several strengths to this study. There have been few studies of COVID-19 prevention within institutions of higher education to date. Increased knowledge for this specific setting is important as a university community may be highly vulnerable to disease outbreaks as they typically have a large amount of social interaction and travel. Several prior studies have documented disease outbreaks focused within university communities [40,41,42,43,44,45,46,47,48]. Additionally, our results provide more information regarding which characteristics are strongly correlated with student belief and adherence to prevention protocols, which may help administrators in the design and application of any future policies related to infectious disease prevention.

At the time of writing, incidence of COVID-19 has decreased to a point where many universities and institutions, including Purdue, have currently relaxed their preventive measures. However, the emergence of COVID-19 variants or other emerging infectious diseases is a possibility [49, 50], and these could trigger reinstatement of similar protocols. Results from this study and similar studies are promising in that they suggest that the policies used in the COVID-19 pandemic were largely effective. However, improvements could be made in education and/or risk communication to men, younger students, and students with a close contact that was ill.

Conclusions

Results from our online survey of Purdue University students are largely consistent with prior literature related to perceptions and knowledge of COVID-19 or other infectious disease prevention measures. Females, Asians/Pacific Islanders, and graduate students reported higher belief and adherence in the COVID-19 prevention measures..