The Mason: Health Starts Here Prospective Cohort Study

Background: Young adulthood is a period of increasing independence for the 40% of young adults enrolled in U.S. colleges. Previous research indicates differences in how students’ health behaviors develop and vary by gender, race, ethnicity, and socioeconomic status. George Mason University is a state institution that enrolls a highly diverse student population, making it an ideal setting to launch a longitudinal cohort study using multiple research methods to evaluate the effects of health behaviors on physical and psychological functioning, especially during the COVID-19 pandemic. Methods: Mason: Health Starts Here was developed as a longitudinal cohort study of successive waves of rst-time freshmen that aims to improve understanding of the natural history and determinants of young adults’ physical health, mental health, and their role in college completion. The study recruits rst-time freshmen who are 18 to 24 years old and able to read and understand English. All incoming freshman are recruited through various methods to participate in a longitudinal cohort for four years. Data collection occurs in fall and spring semesters, with online surveys conducted in both semesters and in-person clinic visits conducted in the fall. Students receive physical examinations during clinic visits and provide biospecimens (blood and saliva). Discussion: The study will produce new knowledge to help understand the development of health-related behaviors during young adulthood. A long-term goal of the cohort study is to support the design of effective, low-cost interventions to encourage young adults’ consistent performance of healthful behaviors, improve their mental health, and improve academic performance.


Background
The transition from adolescence to adulthood in postindustrial societies varies considerably in duration.
It may be an extended and "emerging" period characterized by exploration of identity and human capital acquisition for professional and technological careers (1), or it may be a brief and "accelerated" period in assuming adult roles (e.g., completing school, working fulltime, creating an independent household, or parenting) due to marginalized identities, such as gender, ethnic/racial minority, immigration status, or poverty status (2). The pace and processes of young adult development and maturation, notably socialcognitive and motivational systems, lay the foundation for self-e cacy, decision-making, and behavioral self-regulation (3). Further, the transition to college is a critical time for establishing and reinforcing healthy or unhealthy behaviors including those related to nutrition, exercise, sleep, substance use, and mental health (4). Approximately 40% of young adults enroll in college (5), making this a social context of considerable in uence during this period of development and maturation. Universities are uniquely positioned to study cohorts of young adults for the assessment and promotion of health. Such research has the potential to yield insights, and guide interventions and policies that improve students' health and educational outcomes.
Mental and physical health issues are leading factors interfering with students' academic performance (6,7). Approximately 32% of college students report signi cant mental health symptoms, including depression (17%), generalized anxiety (7%), and suicidal ideation (6%) (8). A recent study of 1,441 adults ages 18 to 39 found that the prevalence of depressive symptoms was more than 3-fold higher during the COVID-19 pandemic compared with before the pandemic (9). College students engage in heavier drinking than their non-college peers (10)(11)(12). Approximately 30% of college students ages 18 to 22 currently binge drink, and 20% used an illicit drug within the past month (13). Additionally, disordered eating behavior, such as binge eating, excessive exercise, and restriction is prevalent among college students (14,15) and linked to all of the mental health conditions described above (16).
By the time they enroll in college, most students have engaged in sexual behavior at some point in their lifetime (17,18). Rates of sexually transmitted infections (STIs) are highest among those in the traditional undergraduate college students age range (19). National samples of college students highlight inconsistent condom or other protective barrier use during sex (17), along with low rates of testing for STIs within this population (20). College students, particularly females, report very high rates of dating and intimate partner violence (21). Studies indicate that an average of 10% of college women experience at least one sexual assault during the academic year, and rates are especially high among rst year college women (22,23). One recent study of 155,026 students across 196 campuses found that the percent of students with lifetime mental health diagnoses had increased from 22% to 36% (24). Of importance to academic outcomes, college students with mental health conditions are at higher risk of dropping out than other students (25). All these factors underscore the need for improved understanding of emotional and behavioral health during the college years.
Unfortunately, higher rates of college enrollment have not uniformly translated into increased rates of graduation. Many students drop out or do not complete their degrees in four years, which creates a large economic burden nationally (26,27).
Several factors that promote college students' mental health and the likelihood of completing their degrees have previously been identi ed. For example, the presence of supportive social networks across multiple systems of relationships, including peers, friends, romantic partners and other adults can support personal and academic success (28). Recent research has linked supportive social networks to mental health (e.g., 29) as well as to educational perseverance and achievement (30). Other psychosocial factors, such as self-perceptions (self-esteem and self-e cacy), attributions (internal control and mastery), motivation (goals and values), and self-regulation (time management and study skills) predict persistence in staying in school and achievement (31). Further, a variety of young adults' personality strengths such as meaning in life, hope, grit, gratitude, curiosity, and knowledge of one's strengths have been correlated with reduced psychological symptoms and increased life satisfaction (32).
Other health behaviors known to shift substantially in young adulthood include diet and exercise, which can have considerable long-term implications. Excessive weight gain and higher rates of obesity are among the changes experienced by young adults. Obesity prevalence among young adults is 34 percent, compared to 20 percent among adolescents (33). Obesity has been found to be a substantial risk factor for, among many other conditions, type-2 diabetes, heart disease, severe osteoarthritis, and mortality (33). Moreover, this risk is disproportionately experienced by select racial and ethnic groups (30). While underresearched, ethnic disparities in the prevalence of obesity and excessive weight gain may begin during young adulthood (34)(35)(36). Particularly relevant for college campuses is that obesity has been found to spread along social and family networks (37). Fortunately, the risk of developing obesity, associated comorbidities, and their attendant costs may be substantially reduced through even modest (5-10%) weight loss. Modest weight loss can reduce diabetes risk up to 58% in individuals with pre-diabetes---a nding consistent for at least 15 years---as well as reduce diabetes-related medical complications (38)(39)(40).
Rates of physical activity fall sharply when young adults transition from high school to college (41). This decline is associated with a variety of factors, from psychosocial, to levels of self-e cacy, and to perceived health bene ts. Despite the health bene ts of physical activity, among them improved cognitive abilities and lower stress, college students face many demands on their time, including academics and work. Previous studies have found that students with higher tness levels have lower credit loads and study less than students with lower tness levels, underscoring the trade-offs that students may face (42). Students' physical activity is related to better academic performance. Data from college students has found that adequate physical activity, sleep, and diet were positive predictors of grade point averages (43), which opens the doors to a variety of novel studies and interventions that can target not just health but also academic and employment outcomes. Numerous questions remain among young adults, including the role of social media exposure in physical activity, differences in physical activity levels by socioeconomic status, ethnicity, culture, gender and other health factors, and whether low-cost, effective interventions can be designed to encourage participation in exercise, improved eating habits, improved academic performance and improved mental health.
Overall, previous evidence points to the importance of capturing the college experience early on, because experiences in freshman year are predictive of continued enrollment in post-secondary education (44). For example, excess stress in the rst semester has been found to be a predictor of poor academic performance, and further, poor academic performance in the rst year predicts attrition (45). Further, it has been observed that many maladaptive health behavior patterns begin in the freshman year and remain stable throughout later undergraduate years (e.g., 46,47). In response to this need to understand the health trajectories of young adults during college, Mason: Health Starts Here was developed as a longitudinal cohort study of successive waves of rst-time freshmen that aims to improve understanding of the natural history and determinants of young adult physical health, mental health, and college completion. The study will produce new knowledge to support the design of primarily low-cost, effective interventions to encourage participation in healthy behaviors, improve mental health, and improve academic performance among young adults.

Objectives and Study Design
The Mason: Health Starts Here is a longitudinal cohort study of rst-time freshmen, and the objectives of the study are: to examine health, health behavior, and mental health as predictors of college completion; to examine the applicability of emerging adulthood or accelerated adulthood theories to college students who differ by socioeconomic status, sexual orientation, ethnicity, culture, gender, and other demographic factors; to examine nutrition and physical activity levels during the college years, a time of increasing adult independence to examine differences in longitudinal associations among physical and psychological health factors by socioeconomic status, sexual orientation, ethnicity, culture, gender, and other demographic factors; to examine how social network in uences and social media modulate healthful and unhealthful behaviors; to examine how social connectedness and trauma exposures may impact mental and physical health; to examine how current and past risky behaviors, sexual risk, gun ownership, and substance use modulate health, mental health, and successful college completion; and to identify modi able risk and protective factors to inform new interventions to improve health and well-being.
Data collection occurs each fall and spring semester and each cohort is followed for four years. Each fall and spring semester, participants will be contacted to participate in online questionnaires. Each fall, they will also be asked to participate in an in-person physical exam in a research center which includes biospecimen collection. Saliva will be collected in all four years, and blood will only be collected during years 1 and 4. The longitudinal capture of data is illustrated in Figure 1, along with timing of incentives.

Study Population
The study includes rst-time freshmen who are 18 to 24 years old who are able to read and understand English. First-time freshman is de ned as students who are newly embarking on a four-year undergraduate degree at Mason. The goal is to recruit 300 participants each year.

Recruitment and Retention
In the pilot year (academic year 2019-2020), the study was advertised to students in selected courses with large numbers of freshmen enrollees through yers outside of classrooms, brief in-class presentations and to select student organization, online video, postcards distributed in class, and email.
Beginning in 2020, contact information of incoming freshman was obtained from University Life and recruitment was expanded to all incoming rst-time freshmen reached through invitation letters and emails, along with video and information during freshmen orientation sessions, and messages to followers on University social media. Additional recruitment occurs by asking faculty teaching courses with high proportions of freshman to share the recruitment video to their classes. Because of COVID-19, in 2020, all recruitment is virtual, but once a higher proportion of in-person classes and student organizations resume in future years, in-person presentations about the study will resume.
Once participants complete an online survey, they are invited to make an appointment for an in-person physical exam and biospecimen collection. Participants are offered $10 for completing each online survey and an additional $25 for completing each in-person visit. Participants receive personalized health feedback based on their survey responses and physical exam. To encourage study retention, participants receive survey reminder emails and birthday emails. Participants are asked to provide contact information of friends or family to facilitate follow-up. The study team also delivers health-related messages to the participant community, such as tips for stress management, and will share summary study ndings.

Data Collection
The study includes online and in-person data collection. Table 1 lists all study measures and the timing of data collection.

Online questionnaire
The online data collection focuses on physical and emotional health, nutrition, sleep, civic engagement, and academic success.

Demographics and Social Determinants of Health
These items include age, race, ethnicity, mother's and father's education, current employment status and hours worked, whether the participant lives on campus, romantic relationship status, religious preference, country of birth for participant and parents, sexual identity, current gender, and gender at birth. Adult food security was assessed using two items adapted from the Hunger Vital Sign screener (48): "1) I worried about not having enough to eat and 2) I tried not to eat a lot so that the food would last." The response options were: often true, sometimes true, never true. Housing insecurity was assessed using two items adapted from the Veterans Screener (49): "In the past 2 months, have you been living in stable housing that you own, rent, or stay in as part of a household?" and "Are you worried or concerned that in the next 2 months you may NOT have stable housing that you own, rent, or stay in as part of a household?" Both questions are answered with yes or no.

Health
Self-rated health (SRH) is measured with one item that asks, "Would you say your health is excellent, very good, good, fair, or poor?" Responses range from 1 (poor) to 5 (excellent). SRH is strongly correlated with objective health status (50).

Migraine
The presence of migraine and migraine severity is assessed using the American Migraine Prevalence and Prevention (AMPP) diagnostic module (7 items; 43) and the Migraine Disability Assessment (MIDAS) questionnaire (7 items; 44), respectively.

Nutrition, Physical Activity, and Sleep
Respondents are asked to self-report height and weight. Habitual dietary and supplement intake is assessed using the Diet History Questionnaire (DHQ)-III food frequency questionnaire, a 161-item tool which asks about food intake over the past month (53). Physical activity is measured with the International Physical Activity Questionnaire (IPAQ) -short version, a 7-item questionnaire (54). Sedentary behavior is assessed using the 18-item PACE questionnaire (55). Sleep quality is measured with 24 items from the Pittsburgh Sleep Quality Index or PSQI (56). The PSQI items form seven component scores that range from 0 to 3 points. Some items are ll-in-the blank and others have four choices to measure sleep quality, latency, duration, e ciency, disturbances, medication, and dysfunction during the past month. The Sexual Experiences Survey (SES; 57) asks participants whether or not they have experienced unwanted sexual contact via a number of different means. The questionnaire has been modi ed to ask participants if they have experienced each type of unwanted sexual assault since the age of 14 and in the past three months. Participants respond 'yes' or 'no' to a series of 10 questions describing sexual assault experiences (65).

Firearms
There are 30 questions assessing a variety of rearm related topics including: acquisition, ownership, access, proximity, safe storage, training, carrying behavior, reasons for carrying, and rearm type, as well as whether they knew someone who was unintentional or intentional shot. These items were adapted from several national surveys including the National Firearm Survey, Pew Research Center, National Comorbidity Survey -Adolescent Supplement (NCS-AS), as well as from previously conducted research (66,67).

Substance Use
Alcohol use is measured with the Alcohol Use Disorders Identi cation Test (AUDIT), a 10-item scale assessing alcohol consumption, drinking behaviors, and alcohol-related problems, which is used to identify individuals at risk of an alcohol use disorder (68). Degree of Problems Related to Drug Abuse is measured using the 10-item Drug Abuse Screening Test (DAST-10; Skinner, 1982). Nicotine use and dependence are measured with the 6-item Fagerström Test for Nicotine Dependence (FTND; 60). Ever smoking cigarettes, past 30-day cigarette smoking, and 30-day menthol cigarette smoking items were used/adapted from the Monitoring the Future (MTF) study and National Survey on Drug Use and Health (NSDUH). Vaping/electronic cigarette (e-cigarette) use was measured using items from the MTF study, modi ed with nicotine clari cation as part of the instructions per the Behavioral Risk Factor Surveillance System (BRFSS).
Past 30-day alcohol consumption items were measured using or adapting items from the NSDUH and BRFSS. Marijuana use was measured using three items from the BRFSS: ever use (yes/no), past 30-day use frequency, and manner of use (i.e., smoke, eat, drink, vaporize, dab, don't know, unsure). Risky drivingrelated behaviors (i.e., driving under the in uence of alcohol, riding in a vehicle driven by someone who had been drinking, driving under the in uence of marijuana, and texting or e-mailing while driving) were assessed using items from the NSDUH and YRBSS or adapting them (i.e., past 30-day rather than 12 month).

Access to Health Care
Health psychiatric medication, history of mental health treatment, health insurance, regular source of care and barriers to access. Measured with if you have ever been prescribed or currently prescribed psychiatric medicine with 1 "Yes" or 2 "No. Adverse Childhood Experiences assesses all types of abuse, neglect, and other potentially traumatic experiences that may be experienced by people under the age of 18 by asking participants to respond "yes" or "no" to the occurrence of 10 events in two categories: abuse (psychological, physical, and sexual) and household dysfunction (substance use, mental illness, physical violence, and criminal behavior). The prevalence of exposure is computed by summing across the number of abuse and household dysfunction categories (85).
Everyday Inequity Scale is a 5-item scale asking participants to report frequency, from 'never' to 'at least once a week', and reasons for experiencing racial and nonracial discrimination in their day-to-day life (86,87).

Peers and Social Support
Perceived Social Support is measured with 12 items from the Multidimensional Scale of Perceived Social Support (88). The items are rated on a 5-point scale from 1 (strongly disagree) to 5 (strongly agree) for relationships with family, friends, and a signi cant other.
Personal Network Composition and Structure. Individuals reported on up to 5 members of their personal networks with whom they discuss important matters and for each of these members they provided information on individual characteristics (e.g., gender, race/ethnicity, education, nature of relationship, frequency and mode of communication), rated positive and negative quality of their relationship with each network member, and estimated likelihood that network members know one another. Similar assessments have been used in large-scale aging surveys, such as the National Social Life, Health, and Aging Project (NSHAP; 78) and Chicago Health, Aging, and Social Relations Study (CHASRS; 79).
Personality Traits. The Mini-IPIP (91) is a 20-item short form of the 50-item International Personality Item Pool-Five-Factor Model Measure (92). The scale measures extraversion, agreeableness, conscientiousness, neuroticism, and intellect/imagination with 4 items for each trait. The items are rated on a 5-point scale from 1 (very inaccurate) to 5 (very accurate).

Environment and Sustainability
Environmental orientation is assessed with the 4-item revised New Environmental Paradigm Scale (93) and sustainable consumption is measured with the 18-item Sustainable Consumption Behavior Scale (94). The NAPS is used to uncover attitudes and beliefs in relation to the environment, while the SCBS measures what, if any, sustainable practices individuals are currently utilizing.

Environmental Exposures
Exposures related to food, employment and personal care products (94) includes 12 items related to purchases of food and personal care items and their packaging, 2 items related to second-hand smoke, 3 questions related to type of work, and 3 items related to products used in the home.

Assessment in the Research Center (In-Person Visit)
During the in person visit to the public health clinic, study participants are asked additional questions about their medical history, given a history and physical exam by a clinician, and asked to donate a venous blood sample and a saliva sample. Participants are requested to fast prior to their appointment and to avoid taking medications and engaging in rigorous physical activity within an hour of their appointment.
At intake, participants are asked for medical history and family history using a checklist of conditions, including indictors of cardiovascular-related health issues, such as high blood pressure, heart disease, stroke, high cholesterol, chest pain, and irregular heartbeat. Participants also complete a past hospitalizations list, a weight history by age, a current medications list, two items on perceived problematic gambling and shopping; six questions on meal timing and meal frequency; and a source and frequency of purchased meals checklist; and two questions on wake-up time that day and participants' usual wake-up time. Women are asked to report on a 15-item menstrual history. Participants are also asked to complete an interviewer-guided Self-Injurious Thoughts and Behaviors Interview (SITBI). This is a 25-item instrument that queries lifetime and past week self-injurious thoughts and behaviors (95). The physical exam includes body weight, height, waist circumference, blood pressure, pulse, respiratory rate, current oral temperature, heart and lung exam, abdominal exam, neck exam, and general physical appearance. Body composition is measured using a DXA scanner (DXA, Hologic Horizon densitometer Hologic, Boston MA). Resting metabolic rate is measured using indirect calorimeter (FitMatePro, Cosmed Inc, Rome, Italy).

Biospecimens
Participants are asked to ll two cryovials with approximately 1 mL of saliva using passive drool procedures. Blood samples include a venous blood draw of 36 mL. Samples are tested immediately for HbA1c, lipid panels, and blood chemistries. Abnormal results are communicated to participants by clinical staff. The remaining sample is processed, aliquoted, and frozen at -70 degrees C.

Discussion
The purpose of the Mason Cohort study is to assess and longitudinally follow, in a racially and ethnically diverse population of young adults, risk factors for the development of chronic disease conditions, as well as the presence and effect of health-promoting habits. Among the outcome measures of particular interest are those factors that predict failure to complete undergraduate degrees among participants.
While most longitudinal cohort studies are by design, purely observational, and thus, at best of indirect bene t to participants, even these may show expectancy effects. For example, participants may be triggered to answer survey questions "appropriately" (so-called "reactivity" effects, or to address unfavorable health-related behaviors through merely being asked about them (96). The Mason Cohort was envisioned as also being a tool to potentially bene t participants directly, in addition to bene tting our understanding as researchers in health elds. Thus, it is a hybrid design, part traditional cohort study and part intervention trial.
This has obvious bene ts to participants, as information and advice concerning their own health and risk factors are returned to them, and interest in improving those risks is potentially fostered. The study design also encourages researchers, stimulated by data gathered and associations discerned to propose sub-studies, which, upon review and approval by the study steering committee are submitted for IRB review and offered to all or subsets of cohort participants based on the speci c questions being asked in the sub-study. Preference will be given to proposed sub-studies which offer a direct potential bene t to study participants.
From an evaluation of theories on college student retention and success, interactionalist theory of social and academic integration suggests that college completion relies upon the extent to which students are committed to their institution, and this commitment level depends on the level of social and academic integration, determined by the quality of interactions (97). The student attrition model, focused more on cognitive and behavioral aspects of college completion, posits a signi cant role for perceptions of satisfaction and attitude impacting behavioral intentions to stay or leave (98). Besides, results from different interventions on college students suggest that low cost and brief interventions can have a meaningful impact on long-term student outcomes (99). Therefore, this study will follow a multi-stage developmental process for preventive interventions (following Institute of Medicine recommendations); the identi cation of a target problem, then the review and investigation of research that can identify protective factors and inform etiologic models, and nally, preventive interventions targeting these factors using experimental designs and intervention analyses (100).

Self-selected population pros and cons…
The self-selected population is a popular sampling technique in many health areas that require human subjects and is an effective sampling strategy, especially in experimental research settings (101). Since the potential research subjects contact study staff directly, this can reduce the amount of time necessary to search for appropriate participants. In addition, the potential participants are likely to be committed to taking part in the study, which can help improve attendance and a greater willingness to provide more insight into the phenomenon being studied. However, with self-selection potential research participants may exhibit an inherent bias in their ' characteristics and approach to the study (102).
To complete any study of living things, it is critical to retain study participants. Participants who complete their scheduled follow-up visits within prede ned visit windows will be considered as retained. Although every effort is being taken to facilitate completing each participant's entry, mid, and exit visits, attrition can and does happen; and the retention rate can uctuate over time and across visits (103). To ensure continued participation and minimize attrition, this study follows the key precepts of retention and practice: facilitating participation, communicating study progress, expressing appreciation, and informing participants of study results (104).
Assuming good rates of retention, we see the future of the study as promising-its breadth of coverage of health predictors and precursors, its diverse sample of young college students, and its exible implementation strategies enhance its capacity to inform emerging health problems and risk factors through examination of the wealth of data and specimens being gathered.
Also of note is the bene t of having baseline specimens and surveys in a diverse sample of young adults when new health problems like COVID-19 emerge.
More work needs to be done, though, including securing sources of ongoing funding, and developing collaborations with other studies of young adults/college students. Future enhancements being considered include following participants for a time frame well beyond the undergraduate years to track health trajectories across the lifespan, and gathering more family data, including adding parents and future offspring of current participants Declarations Ethics approval and consent to participate Study approval was given by George Mason University's Institutional Review Board. Informed consent was obtained. For the online survey portion of the study, potential participants were enrolled and consented online. A separate, written informed consent was obtained for the in-person physical examination. All methods were carried out in accordance with relevant guidelines and regulations, along with the IRB approval.