JMIR Publications

ABSTRACT

Background:

The mental health of university students is a widely recognized global concern. The transition to university life marks a crucial developmental phase characterized by individuation, the establishment of new social connections, and increased autonomy and responsibility. This period aligns with continued, rapid brain development at a time when university students are exposed to multiple risk factors known to affect mental health, including psychosocial stressors, recreational drug use, alcohol binging, and disruptions in sleep patterns. Mental health disorders typically present before or during young adulthood, often going unrecognized for years, resulting in significant delays in receiving treatment. Failing to adequately address mental health issues in a timely fashion can lead to the progression of more complex outcomes, such as school dropout, addiction, and self-harm. Indeed, the international prevalence estimates of mental health disorders reveal higher rates among college students compared to the general population. For example, a multicenter study involving 13,984 students from eight countries, led by the World Health Organization (WHO), demonstrated that one third met clinical criteria for a psychiatric disorder in the previous year (Auerbach et al., 2018). The most frequently reported were depressive disorders (18.5%) and anxiety disorders (16.7-18.6%), followed by alcohol (6.8%) and other substance use disorders (3.0%). Moreover, approximately 22.6% of university students reported experiencing suicidal thoughts. Tragically, suicide stands as the second leading cause of death within this population (Sivertsen et al., 2019). The presence of psychiatric disorders is linked to numerous detrimental consequences both in the short and long term. In the short term, individuals may experience a decline in their quality of life, poorer academic performance, increased absenteeism and a higher likelihood of course dropout. In the long term, individuals endure lower quality of life, higher unemployment rates, and socioeconomic impairment (Duffy et al., 2019); (Dyrbye et al., 2012). Though the alarming data presented above were derived from studies conducted in countries with diverse income levels, to date, there is a disproportionately limited body of evidence addressing the mental health issues of young adults attending universities in low income countries. In Brazil, for example, there are approximately eight million students enrolled in 2714 higher education institutions (INEP, 2021; (Power BI Report, n.d.). The few studies that have been conducted in Brazil have tended to focus on medical students; also unveiling concerns about rates of mental distress in this population (Miguel et al., 2021). For example, approximately 37% were under psychiatric treatment and 4.5% had attempted suicide. The most prevalent diagnoses were depression (39.1%) and anxiety disorders/phobias (33.2%), while severe mental health disorders like psychotic disorders (3.7%) and bipolar disorder (1.9%) were less common (Campos et al., 2017). Here we aim to address numerous gaps in the current national and international literature. First, we will address the mental health of university-aged students in a low income country, namely Brazil. In addition, given that previous studies have been plagued by selection biases owing to the challenge of engaging otherwise healthy young adults in research surveys, we will utilize an engaging web-based survey to mitigate this possibility. The survey will also address another largely unexplored facet of mental health issues, namely the longitudinal course of mental health and quality of life during college. To this end, we propose to measure the mental health of 8,028 undergraduate students of the UniEduK group. UniEduK is a private university located in two medium-sized cities in the southeast of Brazil (Indaiatuba and Jaguariúna), offering numerous undergraduate courses in technical fields (e.g., administration, accounting, architecture, law, engineering) and health sciences (e.g., medicine, nursing, psychology, biomedicine, veterinary medicine). To achieve this goal, we carefully designed an engaging web-based mental health survey suitable for Brazilian Students considering three main challenges: the large number of individuals we aim to screen (n = 8,028), the relatively low budget and resources available to us, and the expected low response rates for this kind of survey.

Objective:

The aim of this paper is to describe the methods used in the design of a web-based mental health survey to be administered in a private Brazilian University.

Methods:

3.1 Overall design The study will be based on validated self-report questionnaires, structured within an electronic web-based survey. The protocol will be repeated every year, evaluating mental health trajectories over time, and including new incoming students to account for any trends specific to the new cohorts of students. The data will be analyzed cross-sectionally (with descriptive and association analyses between the different constructs measured) and longitudinally (with trajectory analyses at the level of the individual, area of academic concentration, and institution). 3.2 Participants/Recruitment All undergraduate students from UniEduk (n = 8,028) will be invited to participate annually through an email message briefly explaining the research and encouraging participation via the electronic survey. The only exclusion criteria are related to limitations in accessing or responding to the survey (i.e., no access to electronic devices or internet connection; being illiterate [both are considered to be of negligible occurrence in this specific setting]). We have devised several strategies aimed at increasing participation. First, we will conduct university wide media campaigns to promote awareness concerning mental health and discuss the importance of the study. Second, professors of all disciplines will be encouraged to remind students to engage in the survey. Third, non-responders and survey non-completers will receive three invitation reminders via email, followed by three reminders via text message (WhatsApp). Importantly, students will be informed that the time spent with the survey will count as an equivalent complementary academic activity. All these steps will be executed automatically and anonymously to ensure students' privacy. When invited, students are required to read and sign an informed consent twice a year. They will be informed at this time that declining participation will not cause any harm to their academic standing or relationship with the University and that they will not receive any more invitations. 3.3 Instruments and Research Requirements We selected empirically validated psychometric self-report questionnaires, with a validated Brazilian Portuguese translation, that can be presented electronically. A typical dilemma encountered with a survey of this type, is the tension between having enough items to adequately assess the phenotypes we aimed to measure, and the impact that a time consuming survey can have on the engagement rate of participants. To address this dilemma, we propose a two-step strategy (Figure 1). In the first step, in addition to collecting sociodemographic and overall lifestyle/quality of life information (SHORT-SMILE) (De Boni et al., 2023), we will screen for an array of mental health conditions using the DSM-5 Cross-Sectional Adult Symptoms Scale DSM-5 Level 1 (DSM-5-TR Online Assessment Measures, n.d.) and a short version of the Adult Self Report Scale for Attention Deficit/Hyperactivity Disorder (ASRS-18) (Kessler et al., 2005) (Table 1). In the second step, participants scoring above an indicated threshold within a specific domain will be administered more detailed psychometric scales. These scales could relate to depression, mania, generalized anxiety, sleeping disorders, borderline personality disorder, obsessive-compulsive disorder, attention-deficit/hyperactivity disorder and/or substance use (Table 2). After administration of relevant scales, the participant will be given the option to respond to a questionnaire about personality traits (i.e, Big Five Inventory (Roiz Junior et al., 2023)). Finally, the total time to complete the questionnaire is estimated to be approximately 20 to 40 minutes, depending on the domains assessed. 3.4 Data Collection Instruments Data collection will be performed through Research Electronic Data Capture (REDCap), as it is a secure online platform for data management that allows the presentation of generalizable online instruments and the database storage. The platform is designed to support data capture for research studies while protecting participant anonymity (Tamuhla et al., 2022). This platform provides a convenient “Survey Queue” for participants to access the survey questionnaires and a “to-do list” so they can keep track of their progress. This allows for the tracking of initial participation, completeness status, and longitudinal data collection for all participants. The REDCap feature best suited to address automation of the communication process and, furthermore, better data collection is the Automated Invitations. The participants receive an individual link and we are able to choose how many reminders will be sent as well as their periodization. 3.5 Survey Distribution Tools We will upload student data straight from a file containing the emails of all enrolled students (provided by the university's administration sector), in order to generate a unique individual ID link for each student. This will allow us to track the participants’ survey engagement/completion rate. 3.6 Reports and Alerts “Reports and alerts” are research tools built into the REDCap platform to keep track of certain events and expected outcomes about the collected data. Our main use of this information will be to keep track of overall completion rates. To this end, we will first establish which information is needed for these reports, and apply the logic criteria that will filter our results. The report then queries all records and returns those that match the criteria. The same principle will also be applied to suicide risk monitoring (Figure 2), to ensure the security of the participants. This monitoring procedure is based on the answers given by the participants in the Cross-Sectional Adult Level 1 Symptoms and PHQ-9 Scales. Specifically, REDCap is set up such that participants answering in the affirmative to either of two questions directly addressing 1) suicidal ideation or 2) suicide plans, automatically generates an email alert notifying the study team. The study team then immediately notifies a relevant healthcare team associated with our research group. Upon receiving this notification, the health care team will contact the participant to offer them appropriate psychiatric care. If the participant does not answer the first contact, the health care team will call again every day for three days. If the research participant accepts the offered care, they will be evaluated by psychiatry professionals and given necessary referrals. If they do not agree, the study team will register this non-agreement and monitoring by the research team will end. 3.7 Statistics analyses As we will be assessing students longitudinally (annually) during their university journey, we will be able to apply numerous appropriate methodological designs in the statistical analyses including cross-sectional, accelerated cohorts, and traditional cohorts. 3.7.1 Simple Cross-Sectional Analyses Descriptive statistics will be performed on the collected variables, with reports divided by categories of interest (e.g., academic courses enrolled, sociodemographic data, etc.). Simple statistical associations will be conducted through comparisons of independent groups using the Student's t-test (for variables with a normal distribution), the Mann-Whitney test (for variables with a non-normal distribution), or ANOVA (depending on the number of groups studied). These analyses will associate quality of life and mental health indices with these categories of interest, as well as with university dropout rates and academic performance. Linear and logistic regressions will be conducted to control for confounding variables such as gender, age, and socioeconomic status. The anticipated analyses include, but are not necessarily limited to, the following: comparing quality of life, personality traits, and the prevalence of psychiatric symptoms among students in various academic courses offered by the university; and exploring associations between socioeconomic status, quality of life, and psychiatric symptoms to identify potential risk and protective factors. 3.7.2 Accelerated Cohort Cross-Sectional Analyses Taking advantage of the large sample size and the design tracking all enrolled students from the first to the last semester of each course, we can use cross-sectional data to compare data by course and semester (or year) for each student, inferring possible changes in these indices from the beginning to the end of the course. For this purpose, Generalized Estimation Equations (GEE) analyses will be designed, using the independent variable of semester/year as a substitute for time in these analyses, while controlling for confounds such as gender, age, and socioeconomic status. 3.7.3 Longitudinal Analyses/Cohort Analyses By following students longitudinally during their university journey, we will be able to analyze data using Generalized Estimation Equations; including time as an independent variable will allow for the detection of nonlinear variations in outcomes and, in turn, address the evolution of mental health symptomatology in individuals. 3.8 Ethical aspects This research was approved by the Research Ethics Committee of the UNIFAJ University Center from UniEduk (no. 6.153.870, 2023), and will be carried out in accordance with the recommendations of Resolution no. 466/2012 of the National Health Council (Conselho Nacional de Saúde, n.d.). Students will be invited to participate and informed about the objectives and methods. They will also be provided with a guarantee of confidentiality and anonymity. The Informed Consent Form (ICF) will be presented to all participants via an online form. Participants will be instructed to read the ICF and express their agreement or otherwise to the question: "Have you understood the guidelines and do you agree to participate freely, knowledgeably and spontaneously in this research?" If they agree, they are asked to enter their full name so that it can be attached to their acceptance to take part in the study. The participant will then be directed to a link to the online survey. If they do not agree, the participant will receive a thank you note and the contact will be closed.

Results:

Discussion: The psychological well-being of young adult college students is gaining significant attention. This is appropriate, as it is one of the major determinants of their overall academic success, personal development, and future prospects. Meanwhile, the existing literature suffers from a dearth of evidence concerning the mental health of university students in low income countries. The present study protocol describes a proposal to systematically and longitudinally survey the mental health of a large population of university students in Brazil. Our plan is based upon methodological decisions that balance quality, precision, scalability, cost, and the likelihood of participant engagement. Among the challenges encountered during the design of this web-based mental health survey, we have highlighted the need to assess a large number of potential participants (n = 8,028) while aiming for a maximum response rate to increase generalizability. Ethical considerations addressed here are related to the importance of ensuring data privacy and interaction with the local health care services to provide assistance in cases of psychiatric emergency (i.e., risk of suicide). We have also strived for this longitudinal study proposal to be as cost-efficient as possible considering financial constraints. Our proposed use of an internet portal will improve participant engagement, data integration and reduce longitudinal data collection time and expense (Donohoe et al., 2012)(Callegaro et al., 2015). Indeed, automated data capture minimizes the need for paid researchers to run participants and enter data, while also reducing data entry errors (Alessi & Martin, 2010). Finally, since our survey involves multiple questionnaires, the platform allows for easier and more engaging access for potential participants (Bernard, 2011). Limitations: Of course, there are disadvantages associated with online research. For example, open-ended questions cannot be explored with immediate follow-up questions and participants are unable to seek clarification of ambiguous items (Andrews et al., 2003). To address this issue, in the present survey, participants can easily email the project team with any questions they may have. Selection bias presents another challenge for online research. Internet access is affected by myriad variables including income, geographical location, mental health status and age. (Ball, 2019)(Gulliver et al., 2010). We do not anticipate serious issues with internet accessibility. UniEduk has free internet access for all students. In addition, there were 181.8 million internet users in Brazil at the beginning of 2023, which means 84.3% of the population (Kemp, 2023); and the southeast, where UniEduk is, has an even greater user concentration. Our plan to remind participants via WhatsApp to finish their questions is sound, as WhatsApp is the most used social media by Brazilians, around 169 million (Kemp, 2023). Thus, though we cannot fully account for all these possibilities, the present strategy maximizes accessibility in a way that will mitigate these potential confounds and, consequently, increase response rate. After identifying participants, the present protocol contains strategies to maximize the probability that participants will continue to engage with and complete the survey. For example, one way to increase engagement is to provide participants with information about the study that a) piques their interest, b) helps them understand the importance of their participation in mental health research, and c) increases their confidence about participating. To this end, the research team will hold meetings with UniEduk course coordinators in order to transmit detailed study information so that it reaches the students. Furthermore, a two-minute video will be sent to all participants. This video contains information about the research team, the reason for the research and its objectives, details about the questionnaires (application time, confidentiality, freedom to decline) and the opportunity to ask questions via email. Our intention is to demonstrate the importance of participation in promoting mental health among the university population. The security of collected data is ensured by the REDCap platform, a trusted and secure data collection and storage platform (Patridge & Bardyn, 2018) used throughout the scientific community. The platform allows for the long-term reduction of research costs, the possibility of utilization on many devices, and rapid data entry, review and analysis (Walther et al., 2011). The data analyses in this project will include cross-sectional, accelerated cohorts, and traditional cohorts. Each analysis aims to summarize the data collected and integrate information about the mental health of this population. The cross-sectional analysis will allow us to understand the Brazilian university population in order to think about intervention strategies. As noted in the Introduction, there is a paucity of data that has been collected from low income countries. Existing studies were carried out mainly with students from specific courses (Pacheco et al., 2017) – national studies covering a range of degrees are rare. This survey is a structured means for assessing this population and their quality of life, given that this period represents a hub that critically impacts their responsibilities, mental health and values (Aceijas et al., 2017). The main intention of the present survey methodology is to inform future preventative measures. It has been well established that preventive actions are feasible, cost effective and efficient in improving overall mental health (Arango et al., 2018). We believe that the results of this survey can guide the design and implementation of preventive programs destined to have a positive impact on the mental health of university students.

Conclusions:

College and university students have high rates of mental health issues. We have developed and described a web-based mental health survey that will allow us to evaluate and detect these issues with low cost and reasonable response rate in a University in Brazil. These efforts will allow us, in the near future, to monitor and test the efficiency and impact of mental health preventive programs. This model could be scaled up across other universities in Brazil to easily assess the mental health status of their students and have a significant impact on the mental health of our communities. Clinical Trial: Study registration: This study was registered in the Open Science Framework (OSF) under the https://doi.org/10.17605/OSF.IO/AM5WS