Advertisement

SARS-CoV-2 Seroprevalence: Demographic and Behavioral Factors Associated With Seropositivity Among College Students in a University Setting

      Abstract

      Purpose

      Examine SARS-CoV-2 seroprevalence and the association of seropositivity with demographic, geographic, and behavioral variables among University of North Carolina Chapel Hill (UNC-CH) undergraduate students enrolled in the fall 2020 semester.

      Methods

      All UNC-CH undergraduate students were invited to participate in the Heelcheck study; participants were weighted to the UNC-CH undergraduate population using raking methods. We estimate SARS-CoV-2 seroprevalence at study entrance (11/12/2020–12/10/2020) and bivariable associations using log-binomial regression.

      Results

      SARS-CoV-2 seroprevalence was 7.3% (95% confidence interval (CI): 5.4%–9.2%) at baseline. Compared to students who were living off-campus in the Chapel Hill/Carrboro area (CH) for the Fall 2020 semester (8.6% seroprevalence), students who never returned to CH had lower seroprevalence (1.9%, prevalence ratio (PR), 95% CI: 0.22, 0.06–0.81), whereas, students who started the semester on-campus and moved to off-campus CH housing had 18.9% seroprevalence (PR, 95% CI: 2.21, 1.04–4.72) and students who spent the semester living in a Sorority/Fraternity house had 46.8% seroprevalence (PR, 95% CI: 5.47, 2.62–11.46). Those who predicted they would join an indoor party unmasked had 3.8 times the seroprevalence of those who indicated they would not attend (PR, 95% CI: 3.80, 1.58–9.16). Compared to students who disagreed with the statement “…I am not going to let COVID-19 stop me from having fun…”, those who agreed had higher seroprevalence (14.0% vs. 5.7%; (PR, 95% CI: 2.45, 1.13–5.32)).

      Discussion

      Increased seroprevalence was associated with congregate living and participation (actual or endorsed) in social activities. During pandemics, universities must create safe socializing opportunities while minimizing transmission.

      Keywords

      Implications and Contribution
      Increased SARS-CoV-2 was observed for undergraduate students in certain congregate living settings (e.g., Sorority/Fraternity housing) and those who participated (actual or endorsed) in social activities. This work should inform the creation of safe socializing opportunities in university settings during waves of the COVID-19 pandemic and future pandemics.
      The COVID-19 pandemic has caused significant morbidity and mortality globally. In the United States alone, the Centers for Disease Control and Prevention estimate that nearly 53 million SARS-CoV-2 infections and 2.4 million hospitalizations may have occurred during February-September 2020 [
      • Reese H.
      • Iuliano A.D.
      • Patel N.N.
      • et al.
      Estimated incidence of coronavirus disease 2019 (COVID-19) illness and hospitalization-United States, February-September 2020.
      ]. While many subgroups are at risk of becoming infected with SARS-CoV-2, young adults (<30), especially those living in congregate settings such as universities, are particularly at risk of acquiring and transmitting SARS-CoV-2 [
      • Boehmer T.K.
      • DeVies J.
      • Caruso E.
      • et al.
      Changing age distribution of the COVID-19 pandemic - United States, May-August 2020.
      ,
      • Salvatore P.P.
      • Sula E.
      • Coyle J.P.
      • et al.
      Recent increase in COVID-19 cases reported among adults aged 18-22 years - United States, May 31-September 5, 2020.
      ,
      • Walke H.T.
      • Honein M.A.
      • Redfield R.R.
      Preventing and responding to COVID-19 on college campuses.
      ]. In addition, young adults play an important role in overall community transmission dynamics [
      • Boehmer T.K.
      • DeVies J.
      • Caruso E.
      • et al.
      Changing age distribution of the COVID-19 pandemic - United States, May-August 2020.
      ]. While COVID-19 vaccines have transformed our capacity to inhibit SARS-CoV-2 transmission and mitigate the pandemic's effects, the possibility for genetic variants to emerge and potentially escape natural and vaccine-induced immunity exists [
      • Baden L.R.
      • El Sahly H.M.
      • Essink B.
      • et al.
      Efficacy and safety of the mRNA-1273 SARS-CoV-2 vaccine.
      ,
      • Shah A.S.V.
      • Gribben C.
      • Bishop J.
      • et al.
      Effect of vaccination on transmission of SARS-CoV-2.
      ,
      • Harvey W.T.
      • Carabelli A.M.
      • Jackson B.
      • et al.
      SARS-CoV-2 variants, spike mutations and immune escape.
      ]. Furthermore, vaccination coverage rates are currently suboptimal, particularly among young people, even in the U.S. where vaccine accessibility is high [
      Does the public want to get a COVID-19 vaccine? When? Kaiser Family Foundation.
      ,
      • Baack B.N.
      • Abad N.
      • Yankey D.
      • et al.
      COVID-19 vaccination coverage and intent among adults aged 18-39 years - United States, March-May 2021.
      ]. As such, there is a need for continued use of additional prevention measures, including masking and physical distancing, especially in transmission-prone settings, such as college campuses.
      College campuses and many of the typical activities associated with campus life are ideal settings for SARS-CoV-2 transmission [
      • Losina E.
      • Leifer V.
      • Millham L.
      • et al.
      College campuses and COVID-19 mitigation: Clinical and economic value.
      ,
      • Cohen A.K.
      • Hoyt L.T.
      • Dull B.
      A descriptive study of COVID-19-related experiences and perspectives of a National Sample of College students in Spring 2020.
      ,
      • Currie D.
      • Moreno G.
      • Delahoy M.
      • et al.
      Interventions to disrupt coronavirus disease transmission at a university, Wisconsin, USA, August–October 2020.
      ]. Previous studies have identified Greek life event participation and off-campus social settings as predictors of SARS-CoV-2 acquisition [
      • Fox M.D.
      • Bailey D.C.
      • Seamon M.D.
      • Miranda M.L.
      Response to a COVID-19 outbreak on a university campus - Indiana, August 2020.
      ,
      • Vang K.E.
      • Krow-Lucal E.R.
      • James A.E.
      • et al.
      Participation in fraternity and sorority activities and the spread of COVID-19 among residential university communities - Arkansas, August 21-September 5, 2020.
      ,
      • Segaloff H.E.
      • Cole D.
      • Rosenblum H.G.
      • et al.
      Risk factors for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and presence of anti–SARS-CoV-2 antibodies among university student dormitory residents, September–November 2020.
      ]. Many U.S. colleges were eager to bring students back to campus for in-person instruction in Fall 2020, and quickly returned to distanced learning following rapid transmission of SARS-CoV-2 among the students on campus [
      • Walke H.T.
      • Honein M.A.
      • Redfield R.R.
      Preventing and responding to COVID-19 on college campuses.
      ,
      • Wilson E.
      • Donovan C.V.
      • Campbell M.
      • et al.
      Multiple COVID-19 clusters on a university campus - North Carolina, August 2020.
      ]. At one large public university, SARS-CoV-2 prevalence among students tested on campus increased from 3% during the week before Fall 2020 classes began to 32% 2 weeks after classes resumed [
      • Wilson E.
      • Donovan C.V.
      • Campbell M.
      • et al.
      Multiple COVID-19 clusters on a university campus - North Carolina, August 2020.
      ]. Many college students have adopted COVID-19 prevention measures; in a survey of 725 full-time college students in the U.S., 95% were sheltering in place per public health guidelines as of late April 2020 [
      • Cohen A.K.
      • Hoyt L.T.
      • Dull B.
      A descriptive study of COVID-19-related experiences and perspectives of a National Sample of College students in Spring 2020.
      ]. Nevertheless, the experience of college life is a unique and important time of development for young adults, especially undergraduate students who are likely leaving their home environment for the first time, and social interactions are a typical component of that experience [
      • Gruber J.
      • Prinstein M.J.
      • Clark L.A.
      • et al.
      Mental health and clinical psychological science in the time of COVID-19: Challenges, opportunities, and a call to action.
      ,
      • Morris M.E.
      • Kuehn K.S.
      • Brown J.
      • et al.
      College from home during COVID-19: A mixed-methods study of heterogeneous experiences.
      ,
      • Lederer A.M.
      • Hoban M.T.
      • Lipson S.K.
      • et al.
      More than inconvenienced: The unique needs of U.S. college students during the COVID-19 pandemic.
      ]. It is essential to understand the unique risk factors for SARS-CoV-2 transmission in the college campus setting, especially within social contexts characteristic of undergraduate campus life, to inform policies and procedures that support in-person instruction and student life. This understanding will be vital as we navigate potential surges in the current COVID-19 pandemic and during future pandemics [
      • Morens D.M.
      • Fauci A.S.
      Emerging pandemic diseases: How we got to COVID-19.
      ].
      The University of North Carolina at Chapel Hill (UNC-CH), like many campuses across the U.S., transitioned to remote learning in March 2020, shifted to a combination of in-person, online, and hybrid classes for the Fall 2020 semester (which started on August 10, 2020), and, shortly after, on August 19, returned to online learning exclusively after a significant rise in COVID-19 cases. The Heelcheck cohort study was designed to determine the seroprevalence of SARS-CoV-2 and identify demographic and behavioral predictors of seroprevalence among undergraduate students enrolled in the fall 2020 semester at UNC-CH. Here, we had three descriptive aims: to (1) estimate SARS-CoV-2 seroprevalence among undergraduate students; (2) describe the distribution of characteristics uniquely tied to the undergraduate college experience during COVID-19, including location of residence and decision-making around COVID-19 prevention behaviors; and (3) identify demographic, behavioral, and social variables associated with SARS-CoV-2 seroprevalence among undergraduate students at a large, public university in the Southern U.S.

      Methods

      Heelcheck cohort study

      The Heelcheck cohort study consisted of a series of three online questionnaires (administered at three separate time points over 11 months) and three corresponding serology tests. The overarching study objective was to estimate SARS-CoV-2 seroprevalence over time, and to identify demographic and behavioral predictors of seroprevalence among UNC-CH undergraduate students. All UNC-CH undergraduate students enrolled in the Fall 2020 semester (n = 19,395) were sent an email invitation to participate in the Heelcheck study on 11/12/2020. The Heelcheck study received approval from the UNC-CH Institutional Review Board.

      Baseline heelcheck questionnaire

      The first (baseline) Heelcheck questionnaire was programmed using Qualtrics and included 64 questions. We used multiple-choice, multiple answer, and open-ended question formats. The baseline questionnaire took approximately 20 minutes to complete and covered a wide range of topics including socio-demographics, involvement in campus activities (e.g., Greek life, UNC athletics), COVID-19 symptoms and prior diagnoses, adherence and compliance with COVID-19 guidelines, COVID-19-related exposures (e.g., attending parties, sporting events) and perceived risk, and attitudes toward vaccination and other prevention measures.
      We used two, unique, scenario-based questions to examine decision-making in social contexts specific to undergraduate campus life [
      • Icenogle G.
      • Cauffman E.
      Adolescent decision making: A decade in review.
      ]. We asked respondents to place themselves in two different situations and indicate (via multiple choice) the behavior they think they would most likely exhibit in the situation (Appendix).
      Finally, we asked two questions about location of residence during the Fall 2020 semester (Appendix) and used the two responses to construct a single variable capturing location of residence type throughout the Fall 2020 semester.

      Seropositivity measure–the Tasso-SST device

      To collect blood and assess SARS-COV-2 seropositivity among the participants, we used the Tasso-SST device—a capillary blood collection device that can be self-administered at home without supervision (Appendix) [
      • Hendelman T.
      • Chaudhary A.
      • LeClair A.C.
      • et al.
      Self-collection of capillary blood using Tasso-SST devices for anti-SARS-CoV-2 IgG antibody testing.
      ]. Blood samples were received at UNC-CH between 11/18/2020 and 01/19/2021 where they underwent anti-SARS-CoV-2 Spike antibody testing. The serology assay is based on spike receptor-binding domain antigen and measures total Ig (IgG, IgA and IgM) with an estimated sensitivity of 95% and specificity of 96% [
      • Narowski T.M.
      • Raphel K.
      • Adams L.E.
      • et al.
      SARS-CoV-2 mRNA vaccine induces robust specific and cross-reactive IgG and unequal strain-specific neutralizing antibodies in naïve and previously infected recipients.
      ,
      • Premkumar L.
      • Segovia-Chumbez B.
      • Jadi R.
      • et al.
      The receptor binding domain of the viral spike protein is an immunodominant and highly specific target of antibodies in SARS-CoV-2 patients.
      ]. Serology test results were categorized as positive, negative, indeterminate, or non-viable (due to poor sample quality). Baseline seropositivity was defined as at least one positive serology result (as some individuals submitted two Tasso samples at baseline).

      Analytic sample and statistical analysis

      For this descriptive analysis, our analytic sample included individuals who consented to the baseline questionnaire and for whom we also had a serology result (n = 680). To minimize the potential for nonresponse bias, weighting techniques were used such that the weighted respondents represented the target population based on a set of characteristics thought to be associated with both study participation and the variables of interest. Using demographic data from the UNC-CH Registrar's Office (as of August 2020), we used iterative proportional fitting (i.e., raking) methods to weight the analytic sample to the marginal distributions of the entire UNC-CH undergraduate population with respect to three characteristics: sex (female, male), class year (first year, sophomore, junior, senior), and Race/Ethnicity (Asian, Black/African-American, Hispanic (all races), White, other)—the latter of which we considered a socio-political rather than biological construct [
      • Izrael D.
      • Hoaglin D.
      • Battaglia M.
      A SAS macro for balancing a weighted sample. Abt Associates Inc.
      ,
      • Williams D.R.
      • Wyatt R.
      Racial bias in health care and health: Challenges and opportunities.
      ,
      • Williams D.R.
      • Collins C.
      Racial residential segregation: A fundamental cause of racial disparities in health.
      ]. The iterative raking procedure continued until the weighted marginal proportions differed from the target population's proportions by <0.5% for each raking variable. Race/Ethnicity was collapsed into a five-level variable to ensure adequate sample sizes for weighting.
      We calculated descriptive statistics, including estimation of SARS-CoV-2 seropositivity, for both the analytic sample consisting of baseline Heelcheck respondents who had a viable baseline serology sample (n = 680) and the weighted population. We described the distribution of COVID-19 related attitudes, exposures, and behaviors, and the scenario-based responses for the weighted population. Next, we used log-binomial regression to estimate the unadjusted prevalence ratios and 95% confidence intervals for the associations between COVID-19 seropositivity and each of the demographic/behavioral variables in the weighted population [
      • Fox M.P.
      • Murray E.J.
      • Lesko C.R.
      • Sealy-Jefferson S.
      On the need to revitalize descriptive epidemiology.
      ,
      • Conroy S.
      • Murray E.J.
      Let the question determine the methods: Descriptive epidemiology done right.
      ]. The 95% confidence intervals in the weighted analysis were calculated using the Taylor Series Linearization method for proportions and the empirical sandwich (robust) variance estimator for prevalence ratios. The results among the analytic sample (n = 680) can be found in the Tables A1–A3. All statistical analyses were conducted using SAS 9.4 (Cary, NC), and the weighting procedure used the SAS macro, RAKING [
      • Izrael D.
      • Hoaglin D.
      • Battaglia M.
      A SAS macro for balancing a weighted sample. Abt Associates Inc.
      ].

      Results

      Demographic characteristics

      According to data from the UNC-CH Registrar's office, there were 19,395 undergraduate students enrolled in the Fall 2020 semester (as of August 2020), of whom 1,147 (6%) consented to the baseline Heelcheck questionnaire. Of the 1,147 baseline respondents, 680 (59%) provided at least one viable blood sample using the Tasso device and were included in the analytic sample (Table 1). Members of the analytic sample completed the baseline questionnaire between 11/12/2020 and 12/10/2020.
      Table 1Demographic and geographic characteristics in the unweighted and weighted populations
      Missing data was imputed based on logical imputation when possible, otherwise the mode response was used.
      Characteristicheelcheck respondents with baseline serology result (N = 680)Weighted population (n = 19,395)
      Weighted frequencies might not sum to the weighted sample size total given use of a consistent rounding rule.
      Survey completion date, range
      The date of the earliest baseline questionnaire record; for individuals with two baseline records, only the first was analyzed.
      11/12/2020–12/10/202011/12/2020–12/10/2020
      Sex at birth
       Female490 (72%)11,636 (60%)
       Male190 (28%)7,759 (40%)
      Race/Ethnicity
       Asian81 (12%)2,931 (15%)
       Black or African American23 (3%)1,716 (9%)
       Hispanic or Latinx (all races)63 (9%)1,868 (10%)
       White459 (68%)11,092 (57%)
       Other
      Other includes American Indian or Alaska Native, Native Hawaiian or Pacific Islander, two or more races, and no response/missing.
      54 (8%)1,788 (9%)
      Age
       <21 years392 (61%)10,982 (60%)
       21+ years254 (39%)7,196 (40%)
       Missing341,216
      Current year at UNC
       First year150 (22%)3,211 (17%)
       Sophomore134 (20%)4,296 (22%)
       Junior167 (25%)5,572 (29%)
       Senior229 (34%)6,316 (33%)
      Location and housing Fall 2020
      Survey respondents were asked to report on their location/housing type on “August 10” (2020) and their location “up until the semester ended”, (November 24, 2020).
       CH off campus Aug 10 and end316 (47%)9,157 (48%)
       Campus housing/dorm Aug 10 and left CH by end136 (20%)3,464 (18%)
       Not in CH Aug 10 or end123 (18%)3,850 (20%)
       Campus housing/dorm Aug 10 and CH off campus at end33 (5%)687 (4%)
       Campus housing/dorm Aug 10 and end25 (4%)801 (4%)
       Chapel Hill off campus Aug 10 and left CH by end23 (3%)690 (4%)
       Sorority/Fraternity house Aug 10 and end11 (2%)356 (2%)
       Other
      Other includes: Not in CH Aug 10 and living in CH off campus by end; Sorority/Fraternity Aug 10 and living in CH off campus by end; Not in CH Aug 10 and campus housing/dorm by end; Sorority/Fraternity Aug 10 and left CH by end.
      6 (1%)153 (1%)
       Missing7236
      Member of UNC Athletic Team
       Yes14 (2%)377 (2%)
       No658 (98%)18,777 (98%)
       Missing/no response8241
      Member of Sorority or Fraternity
       Yes88 (13%)2,518 (13%)
       No582 (87%)16,563 (87%)
       Missing/no response10313
      Tasso date, range
      The date on which the Tasso kit and blood specimen was received at the UNC-CH laboratory, which was likely 2–3 weeks after the specimen was collected by the participant.
      11/18/2020–01/19/202111/18/2020–01/19/2021
       Missing832,178
      SARS-CoV-2 seropositivity
       Positive60 (9%)1,415 (7%)
       Negative or inconclusive620 (91%)17,980 (93%)
       Missing00
      a Missing data was imputed based on logical imputation when possible, otherwise the mode response was used.
      b Weighted frequencies might not sum to the weighted sample size total given use of a consistent rounding rule.
      c The date of the earliest baseline questionnaire record; for individuals with two baseline records, only the first was analyzed.
      d Other includes American Indian or Alaska Native, Native Hawaiian or Pacific Islander, two or more races, and no response/missing.
      e Survey respondents were asked to report on their location/housing type on “August 10” (2020) and their location “up until the semester ended”, (November 24, 2020).
      f Other includes: Not in CH Aug 10 and living in CH off campus by end; Sorority/Fraternity Aug 10 and living in CH off campus by end; Not in CH Aug 10 and campus housing/dorm by end; Sorority/Fraternity Aug 10 and left CH by end.
      g The date on which the Tasso kit and blood specimen was received at the UNC-CH laboratory, which was likely 2–3 weeks after the specimen was collected by the participant.
      Among baseline respondents in the analytic sample, the majority were female at birth (72%) and identified their Race/Ethnicity as White (68%) (Table 1). After the iterative weighting procedure, the distribution of sex, Race/Ethnicity, and class year in the weighted population was nearly identical to that of the entire UNC-CH undergraduate population indicating successful weighting.

      SARS-COV-2 seroprevalence

      The serology specimens were received at UNC-CH between 11/18/2020 and 01/19/2021 (median: 12/21/2020). In the weighted sample, 1,415 of 19,395 students (7.3%, 95% CI: 5.4%, 9.2%) had at least one positive SARS-COV-2 serology test at Heelcheck study entrance.

      COVID-19-related exposures, risk perceptions, and behaviors in the weighted population of undergraduate students

      Among undergraduate students, 377 (2%) were current UNC-CH athletes and 2,518 (13%) were members of a Sorority or Fraternity (Table 1). When comparing where students were living at the beginning of the semester (August 10, 2020) to where they were located at the conclusion of the semester (November 2020), nearly half (48%) were living off-campus in the Chapel Hill/Carrboro area (CH) at both time points. An additional 20% were not living on-campus or in CH at either time point, and 18% started the semester on-campus (dorms), but had left CH by the end of the semester. The remaining 14% started and concluded the semester at a combination of on-campus and off-campus settings in CH and elsewhere, including 2% who started and concluded the semester living in a Sorority/Fraternity house.
      Twenty-three percent of students reported a possible history of SARS-CoV-2 infection, though the majority of this group (77%) had not received a positive test or physician's diagnosis, and only suspected infection due to possible symptoms (Table 2). When asked about their perceived risk of getting COVID-19 in the next year, only 8% responded that there was a high chance; 51% responded moderate chance, 38% slight chance, and 2% indicated almost no chance.
      Table 2Weighted distribution of COVID-19-related exposures, perceptions, and risk-related behaviors
      Heelcheck survey questionWeighted NWeighted %, (95% CI)
      What do you think your risk or chance of getting COVID-19 is in the next year?
       Almost no chance4352%, (1.1%, 3.5%)
       Slight chance7,18638%, (33.5%, 41.8%)
       Moderate chance9,80451%, (47.2%, 55.6%)
       High chance1,5778%, (6.0%, 10.5%)
       No Response850%, (0.0%, 1.0%)
       Missing307
      Have you ever had or been suspected of having Coronavirus/COVID-19?
       Yes, positive blood test920%, (0.0%, 1.0%)
       Yes, a positive nasal swab or saliva test9405%, (3.3%, 6.5%)
       Yes, a medical diagnosis, but no test200%, (0.0%, 0.3%)
       Yes, possible symptoms, but no diagnosis by a test3,45518%, (14.6%, 21.4%)
       No, not to my knowledge14,55476%, (72.3%, 79.6%)
       No Response1041%, (0.0%, 1.2%)
       Missing229
      Was anyone in your household ever diagnosed with Coronavirus/COVID-19 while you were living together?
       Yes2,14911%, (8.8%, 13.6%)
       No17,00389%, (86.1%, 91.1%)
       No response370%, (0.0%, 0.6%)
       Missing207
      Have you lived with any essential workers since COVID-19 began?
       Yes8,65445%, (41.0%, 49.3%)
       No10,37954%, (50.0%, 58.3%)
       No response1311%, (0.0%, 1.4%)
       Missing232
      Once the first coronavirus vaccine becomes widely available, how likely are you to get vaccinated?
       Very or somewhat unlikely7934%, (2.5%, 5.8%)
       Unsure1,4648%, (5.3%, 10.1%)
       Very or somewhat likely16,78088%, (85.3%, 91.0%)
       Missing358
      Within the past month, how often have you done the following when leaving your home:
       Worn a mask inside when you're around other people not in your household
      Never or rarely8564%, (2.8%, 6.2%)
      Occasionally or often2,83015%, (11.7%, 17.9%)
      Most or all of the time15,45281%, (77.3%, 84.1%)
      Missing257
       Maintained physical distancing (at least 6 feet of distance) between yourself and others not in your household
      Never or rarely4232%, (1.0%, 3.4%)
      Occasionally or often4,72925%, (21.0%, 28.4%)
      Most or all of the time13,99873%, (69.3%, 76.9%)
      Missing246
       Only met with others in a group of 10 people or less
      Never or rarely7164%, (2.2%, 5.4%)
      Occasionally or often2,19912%, (8.8%, 14.4%)
      Most or all of the time16,05185%, (81.5%, 87.8%)
      Missing429
      Please indicate your level of agreement with the following statements:
       I am not worried about getting COVID-19 because I'm not living with or interacting with high risk people
      Disagree or strongly disagree13,59872%, (68.3%, 76.0%)
      Neither agree nor disagree2,36313%, (9.8%, 15.3%)
      Agree or strongly agree2,88315%, (12.1%, 18.5%)
      Missing551
       It is important to take part in contact tracing and quarantine as an essential part of controlling the spread of COVID-19
      Disagree or strongly disagree3042%, (0.6%, 2.7%)
      Neither agree nor disagree5543%, (1.6%, 4.3%)
      Agree or strongly agree18,07695%, (93.8%, 97.1%)
      Missing460
       Young people should get COVID-19 so that they get immunity to the virus and then can get on with their normal activities
      Disagree or strongly disagree15,97984%, (81.3%, 87.6%)
      Neither agree nor disagree2,29212%, (9.3%, 15.0%)
      Agree or strongly agree6473%, (2.0%, 4.9%)
      Missing476
       I am not worried about getting COVID-19 because I think COVID-19 is a hoax
      Disagree or strongly disagree18,69799%, (97.9%, 99.5%)
      Neither agree nor disagree2201%, (0.4%, 2.0%)
      Agree or strongly agree200%, (0.0%, 0.3%)
      Missing458
       College only happens one time and I am not going to let COVID-19 stop me from having fun or from doing the social activities that are part of the college experience
      Disagree or strongly disagree15,14580%, (76.5%, 83.1%)
      Neither agree nor disagree2,42513%, (10.1%, 15.4%)
      Agree or strongly agree1,4087%, (5.3%, 9.6%)
      Missing417
      Please rate the following concerns about COVID-19 contact tracing programs:
       My friends being angry at me for sharing their names and contact information
      Not at all concerned11,45362%, (57.7%, 65.8%)
      Slightly or somewhat concerned5,24728%, (24.6%, 32.0%)
      Moderately or extremely concerned1,84210%, (7.6%, 12.3%)
      Missing854
       My friends having to be in quarantine if I report them as my contact
      Not at all concerned9,01048%, (43.7%, 52.2%)
      Slightly or somewhat concerned6,14433%, (28.8%, 36.7%)
      Moderately or extremely concerned3,61919%, (16.1%, 22.5%)
      Missing622
       Getting in trouble with UNC if I report being at a gathering or some other event I am not supposed to be at
      Not at all concerned10,86358%, (53.9%, 62.2%)
      Slightly or somewhat concerned4,84826%, (22.3%, 29.5%)
      Moderately or extremely concerned3,00216%, (13.1%, 19.0%)
      Missing682
       My health and personal information being shared with the University
      Not at all concerned10,92258%, (53.6%, 62.0%)
      Slightly or somewhat concerned5,89431%, (27.3%, 35.1%)
      Moderately or extremely concerned2,08511%, (8.4%, 13.6%)
      Missing494
       Having to be in isolation/quarantine if I test positive or report being a close contact to someone who has tested positive
      Not at all concerned7,61140%, (36.1%, 44.4%)
      Slightly or somewhat concerned6,47434%, (30.3%, 38.2%)
      Moderately or extremely concerned4,82726%, (21.9%, 29.1%)
      Missing483
       My organization/club/team/fraternity/sorority getting in trouble or being closed as a result of something I report
      Not at all concerned13,95179%, (75.5%, 82.4%)
      Slightly or somewhat concerned2,08712%, (9.1%, 14.6%)
      Moderately or extremely concerned1,6319%, (6.8%, 11.6%)
      Missing1,726
      In the past month, have you:
       Been in a car with people you do not live with, no masks
      Yes11,90362%, (57.9%, 66.1%)
      No7,28538%, (33.9%, 42.1%)
      Missing207
       Been to an indoor party with 10 or more people, no masks
      Yes3,94321%, (17.1%, 24.0%)
      No15,24579%, (76.0%, 82.9%)
      Missing207
       Been to a restaurant or bar indoors
      Yes10,71656%, (51.7%, 60.0%)
      No8,47244%, (40.0%, 48.3%)
      Missing207
       Been to an athletic event such as a football or soccer game
      Yes2,66814%, (11.2%, 16.6%)
      No16,52086%, (83.4%, 88.8%)
      Missing207
      Scenarios
       You and two roommates decide to go out on a Friday night. You all leave the house and are wearing masks. You go to a friend's house who invited you over for a socially distanced, outdoor party. When you get there everyone is inside, no one is wearing a mask. Do you:
      Go inside with your mask on1,7889%, (7.1%, 11.6%)
      Go inside with your mask on, but eventually take if off1,7929%, (7.0%, 11.8%)
      Go inside and take your mask off8584%, (2.7%, 6.3%)
      Talk to the host and see if you could move the party outside2,39113%, (9.8%, 15.3%)
      Leave because there are too many people and no masks inside11,69761%, (57.1%, 65.3%)
      No response5773%, (1.5%, 4.5%)
      Missing291
       You're studying late in the common space of your house. One of your roommates comes home with three friends who do not live in your household. They all start hanging out in the room you are in. None of them are wearing masks. Do you:
      Hang out with them with; no mask on2,84515%, (12.0%, 17.7%)
      Hang out with them, but put a mask on1,2116%, (4.3%, 8.3%)
      Ask the friends to put on masks3,77420%, (16.4%, 23.0%)
      Ask the friends to leave1,7989%, (6.9%, 11.8%)
      Leave the room to go somewhere else9,17148%, (43.7%, 52.1%)
      No response3432%, (0.6%, 3.0%)
      Missing254
      95% CI: 95% confidence interval.
      A large majority of respondents expressed a strong commitment to COVID-19 prevention measures. When asked about behavior in the month preceding the questionnaire, 81% reported mask wearing indoors most or all of the time when around non-household contacts, 73% reported maintaining physical distance with non-household contacts most or all of the time, and 85% reported meeting in groups of ≤10 people most or all of the time. Furthermore, 88% confirmed that they were “very or somewhat likely” to get the COVID-19 vaccine once it became available.

      Decision-making and social contexts related to campus life

      In the first scenario question, which asked participants how they would respond if they showed up to a party that they thought was going to be held outdoors with social distancing and instead was being held indoors without face masking, the majority of students reported that they would leave (61%), while 13% reported that they would talk to the host and see if the party could be moved outside (Table 2). The remainder predicted that they would go inside masked (9%), go inside masked, but eventually remove the mask (9%), go inside unmasked (4%), or had no response (3%).
      In the second scenario, which asked students to imagine what they would do if one of their roommates brought three friends, all unmasked, to hang out in the common space of the house that the student was studying in, the three most common responses were: leave the room to go somewhere else (48%), ask the friends to put on masks (20%), and join them and hang out without a mask (15%). In addition, 9% said they would ask the three friends to leave, 6% would hang out with them, but put on a mask, and 2% indicated no response.
      Ninety-nine percent disagreed or strongly disagreed with the statement: “…COVID-19 is a hoax”. When asked their level of agreement with the following statement “College only happens one time and I am not going to let COVID-19 stop me from having fun or from doing the social activities that are part of the college experience”, 80% disagreed or strongly disagreed, however 20% either agreed or did not have an opinion. When asked about contact tracing-related concerns, 26% were extremely or moderately concerned about having to be in isolation/quarantine if they or a close contact tested positive, and 19% were extremely or moderately concerned about “my friends having to be in quarantine if I report them as my contact”.
      In the month before questionnaire completion, 62% had been in a car, unmasked, with people outside of their household, 21% had been to an indoor party, unmasked, with 10+ people, 56% had visited an indoor restaurant or bar, and 14% had attended an athletic event.

      Predictors of seroprevalence in the weighted sample of undergraduate students

      SARS-COV-2 seroprevalence varied by housing location status, and attitudes and behaviors related to COVID-19 risk and prevention (Figure 1, Tables 3 and A4). Compared to students who were living off-campus in the Chapel Hill/Carrboro area (CH) at the beginning and end of semester (8.6% seroprevalence), students who started in campus housing and left CH had a seroprevalence of 5.7% (prevalence ratio (PR), 95% confidence interval (CI): 0.79, 0.39–1.56), and students who never reported living in CH had a seroprevalence of 1.9% (PR, 95% CI: 0.22, 0.06–0.81) (Figure 1). Whereas students who started the semester on-campus and moved to off-campus CH housing had a seroprevalence of 18.9% (PR, 95% CI:2.21, 1.04–4.72) and students who spent the entire semester living in a Sorority/Fraternity house had seroprevalence 46.8% (PR, 95% CI:5.47, 2.62–11.46). Members of a Sorority/Fraternity had four times the prevalence of SARS-COV-2 antibodies compared to nonmembers (20.3% vs. 5.1% (PR, 95% CI:4.00, 2.32–6.88)), and having a household contact with a COVID-19 diagnosis was associated with higher seroprevalence compared to those without (28.7% vs. 4.5% (PR, 95% CI: 6.43, 3.89–10.61)).
      Figure thumbnail gr1
      Figure 1Estimated associations between SARS-CoV-2 seropositivity and demographic/geographic characteristics in the weighted sample. CH: Chapel Hill; Weighted prevalence ratio (PR) estimates and 95% CIs were calculated using log-binomial regression with a robust error variance; PR presented on the log scale; Reference group remains constant until noted; PR not estimated if group had five or fewer individuals and/or zero outcomes (subgroups: Black/African-American, students who were in CH off campus on Aug 10 and then left CH).
      Table 3Estimated associations between SARS-CoV-2 seropositivity and COVID-19-related exposures, perceptions, and risk-related behaviors in the weighted sample
      Heelcheck survey questionSeropositivity
      %PR (95% CI)
      What do you think your risk or chance of getting COVID-19 is in the next year?
       Almost no chance12.2%1.00 (ref)
       Slight chance5.9%0.49 (0.12, 2.03)
       Moderate chance6.6%0.54 (0.13, 2.22)
       High chance10.8%0.88 (0.19, 4.21)
      Have you ever had or been suspected of having Coronavirus/COVID-19?
       Yes, positive blood test31.4%9.72 (2.25, 41.93)
       Yes, a positive nasal swab or saliva test73.4%22.76 (13.81, 37.51)
       Yes, a medical diagnosis, but no test
      Insufficient sample size for estimation (value suppressed); PR not estimated if group had five or fewer individuals and/or zero outcomes.
       Yes, possible symptoms, but no diagnosis by a test5.4%1.66 (0.68, 4.07)
       No, not to my knowledge3.2%1.00 (ref)
      Was anyone in your household ever diagnosed with Coronavirus/COVID-19 while you were living together?
       Yes28.7%6.43 (3.89, 10.61)
       No4.5%1.00 (ref)
      Have you lived with any essential workers since COVID-19 began?
       Yes8.1%1.25 (0.74, 2.12)
       No6.5%1.00 (ref)
      Once the first coronavirus vaccine becomes widely available, how likely are you to get vaccinated?
       Very or somewhat unlikely1.8%0.25 (0.03, 1.83)
       Unsure8.1%1.11 (0.44, 2.79)
       Very or somewhat likely7.3%1.00 (ref)
      Within the past month, how often have you done the following when leaving your home:
       Worn a mask inside when you're around other people not in your household
      Never or rarely8.4%1.18 (0.35, 3.99)
      Occasionally or often8.9%1.25 (0.63, 2.49)
      Most or all of the time7.1%ref (1.00)
       Maintained physical distancing (at least 6 feet of distance) between yourself and others not in your household
      Never or rarely17.9%3.15 (0.96, 10.31)
      Occasionally or often11.5%2.03 (1.17, 3.52)
      Most or all of the time5.7%ref (1.00)
       Only met with others in a group of 10 people or less
      Never or rarely26.8%4.19 (1.97, 8.92)
      Occasionally or often7.8%1.22 (0.55, 2.70)
      Most or all of the time6.4%ref (1.00)
      Please indicate your level of agreement with the following statements:
       I am not worried about getting COVID-19 because I'm not living with or interacting with high risk people
      Disagree or strongly disagree7.0%ref (1.00)
      Neither agree nor disagree7.2%1.04 (0.47, 2.27)
      Agree or strongly agree6.3%0.91 (0.38, 2.20)
       It is important to take part in contact tracing and quarantine as an essential part of controlling the spread of COVID-19
      Disagree or strongly disagree20.9%3.03 (0.81, 11.25)
      Neither agree nor disagree6.9%1.00 (0.24, 4.18)
      Agree or strongly agree6.9%ref (1.00)
       Young people should get COVID-19 so that they get immunity to the virus and then can get on with their normal activities
      Disagree or strongly disagree6.2%ref (1.00)
      Neither agree nor disagree8.3%1.34 (0.63, 2.85)
      Agree or strongly agree26.6%4.31 (1.96, 9.47)
       I am not worried about getting COVID-19 because I think COVID-19 is a hoax
      Disagree or strongly disagree7.1%ref (1.00)
      Neither agree nor disagree10.9%1.53 (0.23, 10.04)
      Agree or strongly agree
      Insufficient sample size for estimation (value suppressed); PR not estimated if group had five or fewer individuals and/or zero outcomes.
       College only happens one time and I am not going to let COVID-19 stop me from having fun or from doing the social activities that are part of the college experience
      Disagree or strongly disagree5.7%ref (1.00)
      Neither agree nor disagree11.7%2.05 (1.08, 3.90)
      Agree or strongly agree14.0%2.45 (1.13, 5.32)
      Please rate the following concerns about COVID-19 contact tracing programs:
       My friends being angry at me for sharing their names and contact information
      Not at all concerned4.7%ref (1.00)
      Slightly or somewhat concerned8.4%1.79 (0.95, 3.37)
      Moderately or extremely concerned18.8%3.99 (2.11, 7.55)
       My friends having to be in quarantine if I report them as my contact
      Not at all concerned5.1%ref (1.00)
      Slightly or somewhat concerned7.2%1.40 (0.74, 2.67)
      Moderately or extremely concerned11.4%2.22 (1.15, 4.30)
       Getting in trouble with UNC if I report being at a gathering or some other event I am not supposed to be at
      Not at all concerned3.8%ref (1.00)
      Slightly or somewhat concerned9.6%2.50 (1.28, 4.88)
      Moderately or extremely concerned13.9%3.61 (1.89, 6.90)
       My health and personal information being shared with the University
      Not at all concerned5.8%ref (1.00)
      Slightly or somewhat concerned9.4%1.62 (0.91, 2.88)
      Moderately or extremely concerned7.9%1.37 (0.62, 3.02)
       Having to be in isolation/quarantine if I test positive or report being a close contact to someone who has tested positive
      Not at all concerned6.3%ref (1.00)
      Slightly or somewhat concerned6.7%1.06 (0.55, 2.05)
      Moderately or extremely concerned9.1%1.44 (0.76, 2.74)
       My organization/club/team/fraternity/sorority getting in trouble or being closed as a result of something I report
      Not at all concerned4.9%ref (1.00)
      Slightly or somewhat concerned12.4%2.54 (1.26, 5.11)
      Moderately or extremely concerned18.9%3.86 (2.01, 7.44)
      In the past month, have you:
       Been in a car with people you do not live with, no masks
      Yes8.2%1.49 (0.82, 2.70)
      No5.5%ref (1.00)
       Been to an indoor party with 10 or more people, no masks
      Yes13.0%2.31 (1.34, 3.98)
      No5.6%ref (1.00)
       Been to a restaurant or bar indoors
      Yes9.1%1.95 (1.07, 3.55)
      No4.7%ref (1.00)
       Been to an athletic event such as a football or soccer game
      Yes11.8%1.83 (0.99, 3.38)
      No6.4%ref (1.00)
      Scenarios
       You and two roommates decide to go out on a Friday night. You all leave the house and are wearing masks. You go to a friend's house who invited you over for a socially distanced, outdoor party. When you get there everyone is inside, no one is wearing a mask. Do you:
      Go inside with your mask on9.0%1.74 (0.76, 3.98)
      Go inside with your mask on, but eventually take if off9.2%1.78 (0.79, 4.03)
      Go inside and take your mask off19.6%3.80 (1.58, 9.16)
      Talk to the host and see if you could move the party outside8.1%1.57 (0.67, 3.66)
      Leave because there are too many people and no masks inside5.2%ref (1.00)
       You're studying late in the common space of your house. One of your roommates comes home with three friends who do not live in your household. They all start hanging out in the room you are in. None of them are wearing masks. Do you:
      Hang out with them with; no mask on13.3%1.85 (0.99, 3.45)
      Hang out with them, but put a mask on3.4%0.47 (0.07, 3.34)
      Ask the friends to put on masks3.7%0.51 (0.22, 1.22)
      Ask the friends to leave6.0%0.83 (0.32, 2.15)
      Leave the room to go somewhere else7.2%ref (1.00)
      Weighted prevalence ratio estimates and 95% CIs were calculated using log-binomial regression with a robust error variance; Reference category is subgroup with least theoretical risk of SARS-CoV-2 acquisition; % = prevalence; PR = prevalence ratio.
      a Insufficient sample size for estimation (value suppressed); PR not estimated if group had five or fewer individuals and/or zero outcomes.
      In the scenario of the indoor party with no masking, those who replied that they would go inside and join the party without a mask had the highest seroprevalence, 19.6% -- 3.8 times the seroprevalence of those who would leave the party (PR, 95% CI: 3.80, 1.58–9.16). Compared to students who disagreed/strongly disagreed with the statement “College only happens one time and I am not going to let COVID-19 stop me from having fun…”, those who agreed/strongly agreed had roughly 2.5 times the prevalence of COVID-19 antibodies (14.0% vs. 5.7%; (PR, 95% CI: 2.45, 1.13–5.32)). Students who were moderately or extremely concerned about “My friends being angry at me for sharing their names and contact information” had four times the seroprevalence compared to those who were not at all concerned (18.8% vs. 4.7% (PR, 95% CI: 3.99, 2.11–7.55)). Those who recently went to an indoor party, unmasked, had 13% seroprevalence (vs. 5.6% (PR, 95% CI: 2.31, 1.34–3.98)); those who recently went to an indoor restaurant or bar had 9.1% seroprevalence (vs. 4.7% (PR, 95% CI: 1.95, 1.07–3.55)); and those who had recently been to an athletic event had 11.8% seroprevalence (vs. 6.4% (PR, 95% CI: 1.83, 0.99–3.38)).

      Discussion

      In this analysis, 7.3% of undergraduates were estimated to be seropositive for SARS-COV-2 at the end of the Fall 2020 semester, eventhough 23% self-reported a possible history of SARS-CoV-2 infection. While prevention measures were supported by the vast majority of students, COVID-19-related risk behaviors and attitudes/beliefs supportive of these behaviors were also prevalent. Importantly, many of these behaviors and attitudes were associated with SARS-COV-2 seroprevalence.
      Many institutions, including UNC-CH, returned to some degree of in-person instruction for the Fall 2020 semester [
      • Walke H.T.
      • Honein M.A.
      • Redfield R.R.
      Preventing and responding to COVID-19 on college campuses.
      ]. Even with the implementation of an array of COVID-19 mitigation strategies, rapid transmission of SARS-CoV-2 resulted in a shift to solely online instruction for UNC-CH on August 19, 2020 [
      • Walke H.T.
      • Honein M.A.
      • Redfield R.R.
      Preventing and responding to COVID-19 on college campuses.
      ,
      • Wilson E.
      • Donovan C.V.
      • Campbell M.
      • et al.
      Multiple COVID-19 clusters on a university campus - North Carolina, August 2020.
      ]. The impact of increased SARS-CoV-2 transmission on college campuses is not restricted to the campus setting itself. Leidner et al. [
      • Leidner A.J.
      • Barry V.
      • Bowen V.B.
      • et al.
      Opening of large institutions of higher education and county-level COVID-19 incidence - United States, July 6-September 17, 2020.
      ] compared county-level SARS-CoV-2 incidence during two periods, one before and one after the Fall 2020 semester started, and found decreased incidence in counties containing no university or a large university with remote-only instruction, but a relative increase of 56% in counties containing a large university that initiated in-person instruction. Furthermore, campus closures following a period of in-person learning may inadvertently seed communities far from campus with infected (possibly asymptomatic) students as they relocate post-closure [
      • Walke H.T.
      • Honein M.A.
      • Redfield R.R.
      Preventing and responding to COVID-19 on college campuses.
      ,
      • Fox M.D.
      • Bailey D.C.
      • Seamon M.D.
      • Miranda M.L.
      Response to a COVID-19 outbreak on a university campus - Indiana, August 2020.
      ,
      • Murakami K.
      ]. In our analysis, we found that one-quarter of UNC-CH students reported living on campus at the beginning of the semester. However, the majority (70%) of these students reported having left CH by the end of the semester. SARS-COV-2 seroprevalence was high (18.9%) among students who started on campus and transitioned to CH off-campus housing while it was lower (6.7%) among students who started on campus and transitioned to living at home or outside of CH and was lowest, just 1.9%, among students who never returned to UNC-CH for the fall semester. Our survey highlights the possible COVID-19 risk related to congregate living settings and suggests that individuals who stayed in the community surrounding campus experienced higher risk than those who left or never returned, likely due to continued socialization and behaviors that put them at risk for COVID-19 acquisition. University policies that detail the approach for a rapid transition to remote instruction, such as testing for COVID-19 before leaving the campus community or recommending that students self-isolate and practice safer behaviors in their new location when relocating, should be in place to minimize the community-level impact of future pandemic-related university closures.
      The COVID-19 pandemic disrupted the typical U.S. college experience for millions of young adults [
      World Health Organization
      Behavioural considerations for promoting safe behaviours policy brief.
      ]. Engagement in campus life is often an experience of unprecedented independence for students, and social connectivity supports their well-being during this time [
      • Peltier G.L.
      • Laden R.
      • Matranga M.
      Student persistence in college: A review of research.
      ,
      • Arnett J.J.
      Emerging adulthood. A theory of development from the late teens through the twenties.
      ]. There is clear evidence of an association between social isolation or loneliness and negative mental health outcomes, such as depression and anxiety, among young adults [
      • Loades M.E.
      • Chatburn E.
      • Higson-Sweeney N.
      • et al.
      Rapid systematic review: The impact of social isolation and loneliness on the mental health of children and adolescents in the context of COVID-19.
      ,
      • Beam C.R.
      • Kim A.J.
      Psychological sequelae of social isolation and loneliness might be a larger problem in young adults than older adults.
      ,
      • Giovenco D.
      • Shook-Sa B.E.
      • Hutson B.
      • et al.
      Social isolation and psychological distress among southern US college students in the era of COVID-19.
      ]. In our study, nearly every respondent (99%) disagreed with the statement that “…COVID-19 is a hoax”, but only 80% disagreed with the statement: “College only happens one time and I am not going to let COVID-19 stop me from having fun or from doing the social activities that are part of the college experience”. This finding highlights the importance placed on the college experience—a large part of which includes social events and activities—not just the academic classroom learning. Students in this study reported good adherence to COVID-19 prevention measures including masking, physical distancing, and limiting the size of gatherings. However, many had concerns about participating in contact-tracing programs due to concern around how peers would react. Furthermore, in the scenarios we posed regarding social situations with friends, many reported participating in the situation even if it posed risk of COVID-19 rather than leaving the situation or confronting peers with requests to make the situation safer. Interestingly, very few participants perceived themselves as having a “high chance” of getting COVID-19 in the next year despite 10% of the participants being sero-positive—the discrepancy between actual risk and perceived risk, especially among young adults, has been described in the context of infectious and noninfectious disease [
      • Antwi J.
      • Lavin R.
      • Sullivan S.
      • Bellavia M.
      Perception of and risk factors for type 2 diabetes among students attending an upstate New York college: A pilot study.
      ,
      • Hickey M.T.
      • Cleland C.
      Sexually transmitted infection risk perception among female college students.
      ]. While undergraduate students may strongly support COVID-19 prevention measures, there may be a discrepancy between support of prevention methods and their actual implementation—the desire to experience “normal” social lives and the potential to underestimate risk must be recognized and incorporated into realistic mitigation efforts [
      • Slovic P.
      Perception of risk.
      ]. As opposed to completely restricting social activities, colleges should create and encourage safe opportunities to socialize and experience “college life” while minimizing COVID-19 transmission.
      We identified several activities, attitudes, and settings that were associated with SARS-COV-2 seroprevalence. Greek life membership was associated with four times the seroprevalence of SARS-COV-2 and living in Sorority/Fraternity housing (vs. CH off-campus) was associated with over five times the seroprevalence. Previous studies have highlighted the COVID-19 risk associated with Greek life event participation [
      • Vang K.E.
      • Krow-Lucal E.R.
      • James A.E.
      • et al.
      Participation in fraternity and sorority activities and the spread of COVID-19 among residential university communities - Arkansas, August 21-September 5, 2020.
      ,
      • Segaloff H.E.
      • Cole D.
      • Rosenblum H.G.
      • et al.
      Risk factors for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and presence of anti–SARS-CoV-2 antibodies among university student dormitory residents, September–November 2020.
      ]. In fact, Segaloff et al. [
      • Segaloff H.E.
      • Cole D.
      • Rosenblum H.G.
      • et al.
      Risk factors for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and presence of anti–SARS-CoV-2 antibodies among university student dormitory residents, September–November 2020.
      ] identified just planning to attend a Greek life event as a predictor of SARS-CoV-2 infection. We also observed higher SARS-COV-2 seroprevalence among individuals who reported attending an indoor party, indoor bar or restaurant, or an athletic event. Many college campuses have implemented strict standards regarding masking and physical distancing on campus, however much of the COVID-19 risk is likely encountered in off-campus social settings [
      • Fox M.D.
      • Bailey D.C.
      • Seamon M.D.
      • Miranda M.L.
      Response to a COVID-19 outbreak on a university campus - Indiana, August 2020.
      ]. Partnering with influential student organizations (e.g., Greek life organizations) and community partners to promote and uphold COVID-19 prevention measures (while allowing for safer socializing) may help reduce the impact of future pandemics or new waves of COVID-19 in university settings [
      • Walke H.T.
      • Honein M.A.
      • Redfield R.R.
      Preventing and responding to COVID-19 on college campuses.
      ,
      • Vang K.E.
      • Krow-Lucal E.R.
      • James A.E.
      • et al.
      Participation in fraternity and sorority activities and the spread of COVID-19 among residential university communities - Arkansas, August 21-September 5, 2020.
      ].
      Several limitations should be noted. First, we inquired about behavior at a single time point and determined seroprevalence using an antibody test; therefore, we were unable to determine the temporal order of SARS-CoV-2 infection and self-reported exposures/behaviors. Future work should use additional data sources and multivariable methods to estimate causal effects for the associations we identified. Next, due to waning antibody levels, our estimate of seroprevalence may underestimate the true history of SARS-CoV-2 infection in the UNC-CH undergraduate population. Additionally, an unknown number of Heelcheck serology specimens were lost in the mail or destroyed in transit before arriving at UNC-CH, however we think this would be independent of serology result. Finally, our choice of raking variables for the weighting procedure was limited to the variables available from the Registrar's office. Non-response bias may still be affecting our estimates due to differential response rates for variables that predict seropositivity (and other characteristics we studied) and could not be accounted for using the Registrar's data.
      In conclusion, those who returned to UNC-CH, lived in congregate housing, had concerns related to prevention strategies/contact-tracing, and attended or condoned attending large social events without masks were more likely to have a history of SARS-CoV-2 infection. Universities must acknowledge students' need to socialize, discuss and alleviate student concerns about participating in public health prevention methods, and support and participate in the creation of safer activities that allow for socializing while minimizing COVID-19 risk.

      Funding Sources

      This work was supported by the North Carolina Policy Collaboratory through appropriation from the North Carolina General Assembly (NCGA) in support of research on treatment, community testing, and prevention of COVID-19 (as mandated by the NCGA in subdivision (23) of Section 3.3 of Session Law 2020-4).

      Supplementary Data

      References

        • Reese H.
        • Iuliano A.D.
        • Patel N.N.
        • et al.
        Estimated incidence of coronavirus disease 2019 (COVID-19) illness and hospitalization-United States, February-September 2020.
        Clin Infect Dis. 2021; 72: e1010-e1017
        • Boehmer T.K.
        • DeVies J.
        • Caruso E.
        • et al.
        Changing age distribution of the COVID-19 pandemic - United States, May-August 2020.
        MMWR Morb Mortal Wkly Rep. 2020; 69: 1404-1409
        • Salvatore P.P.
        • Sula E.
        • Coyle J.P.
        • et al.
        Recent increase in COVID-19 cases reported among adults aged 18-22 years - United States, May 31-September 5, 2020.
        MMWR Morb Mortal Wkly Rep. 2020; 69: 1419-1424
        • Walke H.T.
        • Honein M.A.
        • Redfield R.R.
        Preventing and responding to COVID-19 on college campuses.
        JAMA. 2020; 324: 1727-1728
        • Baden L.R.
        • El Sahly H.M.
        • Essink B.
        • et al.
        Efficacy and safety of the mRNA-1273 SARS-CoV-2 vaccine.
        N Engl J Med. 2020; 384: 403-416
        • Shah A.S.V.
        • Gribben C.
        • Bishop J.
        • et al.
        Effect of vaccination on transmission of SARS-CoV-2.
        N Engl J Med. 2021; 385: 1718-1720
        • Harvey W.T.
        • Carabelli A.M.
        • Jackson B.
        • et al.
        SARS-CoV-2 variants, spike mutations and immune escape.
        Nat Rev Microbiol. 2021; 19: 409-424
      1. Does the public want to get a COVID-19 vaccine? When? Kaiser Family Foundation.
        (Available at:)
        • Baack B.N.
        • Abad N.
        • Yankey D.
        • et al.
        COVID-19 vaccination coverage and intent among adults aged 18-39 years - United States, March-May 2021.
        MMWR Morb Mortal Wkly Rep. 2021; 70: 928-933
        • Losina E.
        • Leifer V.
        • Millham L.
        • et al.
        College campuses and COVID-19 mitigation: Clinical and economic value.
        Ann Intern Med. 2021; 174: 472-483
        • Cohen A.K.
        • Hoyt L.T.
        • Dull B.
        A descriptive study of COVID-19-related experiences and perspectives of a National Sample of College students in Spring 2020.
        J Adolesc Health. 2020; 67: 369-375
        • Currie D.
        • Moreno G.
        • Delahoy M.
        • et al.
        Interventions to disrupt coronavirus disease transmission at a university, Wisconsin, USA, August–October 2020.
        Emerg Infect Dis J. 2021; 27: 2776
        • Fox M.D.
        • Bailey D.C.
        • Seamon M.D.
        • Miranda M.L.
        Response to a COVID-19 outbreak on a university campus - Indiana, August 2020.
        MMWR Morb Mortal Wkly Rep. 2021; 70: 118-122
        • Vang K.E.
        • Krow-Lucal E.R.
        • James A.E.
        • et al.
        Participation in fraternity and sorority activities and the spread of COVID-19 among residential university communities - Arkansas, August 21-September 5, 2020.
        MMWR Morb Mortal Wkly Rep. 2021; 70: 20-23
        • Segaloff H.E.
        • Cole D.
        • Rosenblum H.G.
        • et al.
        Risk factors for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and presence of anti–SARS-CoV-2 antibodies among university student dormitory residents, September–November 2020.
        Open Forum Infect Dis. 2021; 8: ofab405
        • Wilson E.
        • Donovan C.V.
        • Campbell M.
        • et al.
        Multiple COVID-19 clusters on a university campus - North Carolina, August 2020.
        MMWR Morb Mortal Wkly Rep. 2020; 69: 1416-1418
        • Gruber J.
        • Prinstein M.J.
        • Clark L.A.
        • et al.
        Mental health and clinical psychological science in the time of COVID-19: Challenges, opportunities, and a call to action.
        Am Psychol. 2021; 76: 409-426
        • Morris M.E.
        • Kuehn K.S.
        • Brown J.
        • et al.
        College from home during COVID-19: A mixed-methods study of heterogeneous experiences.
        PLoS One. 2021; 16e0251580
        • Lederer A.M.
        • Hoban M.T.
        • Lipson S.K.
        • et al.
        More than inconvenienced: The unique needs of U.S. college students during the COVID-19 pandemic.
        Health Educ Behav. 2021; 48: 14-19
        • Morens D.M.
        • Fauci A.S.
        Emerging pandemic diseases: How we got to COVID-19.
        Cell. 2020; 182: 1077-1092
        • Icenogle G.
        • Cauffman E.
        Adolescent decision making: A decade in review.
        J Res Adolesc. 2021; 31: 1006-1022
        • Hendelman T.
        • Chaudhary A.
        • LeClair A.C.
        • et al.
        Self-collection of capillary blood using Tasso-SST devices for anti-SARS-CoV-2 IgG antibody testing.
        PLoS One. 2021; 16e0255841
        • Narowski T.M.
        • Raphel K.
        • Adams L.E.
        • et al.
        SARS-CoV-2 mRNA vaccine induces robust specific and cross-reactive IgG and unequal strain-specific neutralizing antibodies in naïve and previously infected recipients.
        bioRxiv. 2021; https://doi.org/10.1101/2021.06.19.449100
        • Premkumar L.
        • Segovia-Chumbez B.
        • Jadi R.
        • et al.
        The receptor binding domain of the viral spike protein is an immunodominant and highly specific target of antibodies in SARS-CoV-2 patients.
        Sci Immunol. 2020; 5eabc8413
        • Izrael D.
        • Hoaglin D.
        • Battaglia M.
        A SAS macro for balancing a weighted sample. Abt Associates Inc.
        (Available at:)
        • Williams D.R.
        • Wyatt R.
        Racial bias in health care and health: Challenges and opportunities.
        JAMA. 2015; 314: 555-556
        • Williams D.R.
        • Collins C.
        Racial residential segregation: A fundamental cause of racial disparities in health.
        Public Health Rep. 2001; 116: 404-416
        • Fox M.P.
        • Murray E.J.
        • Lesko C.R.
        • Sealy-Jefferson S.
        On the need to revitalize descriptive epidemiology.
        Am J Epidemiol. 2022; 191: 1174-1179
        • Conroy S.
        • Murray E.J.
        Let the question determine the methods: Descriptive epidemiology done right.
        Br J Cancer. 2020; 123: 1351-1352
        • Leidner A.J.
        • Barry V.
        • Bowen V.B.
        • et al.
        Opening of large institutions of higher education and county-level COVID-19 incidence - United States, July 6-September 17, 2020.
        MMWR Morb Mortal Wkly Rep. 2021; 70: 14-19
        • Murakami K.
        Fauci urges colleges not to send students home. Inside Higher Ed, Washington, DC2020
        • World Health Organization
        Behavioural considerations for promoting safe behaviours policy brief.
        (Available at:)
        http://apps.who.int/bookorders
        Date: 2021
        Date accessed: October 25, 2021
        • Peltier G.L.
        • Laden R.
        • Matranga M.
        Student persistence in college: A review of research.
        J Coll Student Retent Res Theory Pract. 2000; 1: 357-375
        • Arnett J.J.
        Emerging adulthood. A theory of development from the late teens through the twenties.
        Am Psychol. 2000; 55: 469-480
        • Loades M.E.
        • Chatburn E.
        • Higson-Sweeney N.
        • et al.
        Rapid systematic review: The impact of social isolation and loneliness on the mental health of children and adolescents in the context of COVID-19.
        J Am Acad Child Adolesc Psychiatry. 2020; 59: 1218-1239.e3
        • Beam C.R.
        • Kim A.J.
        Psychological sequelae of social isolation and loneliness might be a larger problem in young adults than older adults.
        Psychol Trauma. 2020; 12: S58-S60
        • Giovenco D.
        • Shook-Sa B.E.
        • Hutson B.
        • et al.
        Social isolation and psychological distress among southern US college students in the era of COVID-19.
        medRxiv. 2022; https://doi.org/10.1101/2021.12.31.21268596
        • Antwi J.
        • Lavin R.
        • Sullivan S.
        • Bellavia M.
        Perception of and risk factors for type 2 diabetes among students attending an upstate New York college: A pilot study.
        Diabetol Metab Syndr. 2020; 12: 25
        • Hickey M.T.
        • Cleland C.
        Sexually transmitted infection risk perception among female college students.
        J Am Assoc Nurse Pract. 2013; 25: 377-384
        • Slovic P.
        Perception of risk.
        Science. 1987; 236: 280-285