Fluctuations in mental health in students accessing a university-wide online mental health promotion intervention before and during the COVID-19 pandemic

ABSTRACT

Objective

Growing evidence shows that online mental health interventions for university students are a cost-effective means for addressing mental health problems among university students. This study presents data on the numbers and characteristics of students who engaged in an online mental health promotion intervention called YOLO in the context of a university-wide rollout of the program before and after the onset of the COVID-19 pandemic.

Methods

Participants completed a questionnaire assessing socio-demographics and mental health (wellbeing and distress). A total of 240 students accessed the program over a 12-month period and of these, 164 provided reliable data.

Results

YOLO was accessed by students representing diverse socio-demographics including domestic vs. international student status, degree level, faculty, ethnicity, and age, although there was an over-representation of females (77.2%). Students who accessed YOLO pre-COVID-19 did not substantially differ across most mental health dimensions from those who accessed it during-COVID-19, although more students reported mild-moderate anxiety symptoms during-COVID than pre-COVID-19. The only socio-demographics significantly associated with mental health were age and degree level. Younger age was related to higher anxiety and 3^rd and 4^th year undergraduate students reported poorer mental health comparative to students in other years. Additionally, a higher proportion of these students accessed YOLO during-COVID-19 than at pre-COVID-19.

Conclusion

The present data along with findings from other studies that have evaluated YOLO, suggest that online mental health promotion programs that do not require in-person contact are an important and necessary resource for all university students.

KEY POINTS

What is already known about this topic:

• (1) University students report high levels of mental distress.

• (2) Universities often lack capacity to provide support to all students.

• (3) Online, mental health programs may be a valuable tool to provide support.

What this topic adds:

• (1) University students’ self-reported psychological distress did not increase significantly pre and during COVID-19, however third-year and fourth-year students reported the most psychological distress during COVID-19.

• (2) The YOLO program was used by students of multiple cultural backgrounds, ages, enrolment statuses, and disciplines.

• (3) The YOLO online program may be one useful program Universities can use to support students’ psychological wellbeing.

KEYWORDS:

• Acceptance and Commitment Therapy
• digital health promotion
• mental health promotion
• self-guided interventions
• university students
• web-based interventions

University students have an elevated risk for mental health problems relative to their peers (Blanco et al., 2008; Gallagher & Taylor, 2013; Hunt & Eisenberg, 2010; Stallman, 2010). Prevalence estimates suggest between 12% and 46% of university students will experience mental health difficulties within any given year (Auerbach et al., 2016, 2018; Blanco et al., 2008; Eisenberg et al., 2013; Verger et al., 2010). Within Australia, 83.9% of students report elevated psychological distress regardless of university location (Stallman, 2010). The onset of the COVID-19 pandemic and associated lockdown and social distancing measures have increased the risk of mental health problems among university students (Browning et al., 2021; Chen & Lucock, 2022; Křeménková et al., 2021), and intensified calls for cost-effective mental health promotion interventions for university students (Li et al., 2021). This focus is important given that mental health problems negatively impact students’ academic performance, work, daily activities, physical health, and quality of life (Andrews & Wilding, 2004; Stewart-Brown et al., 2000; Vaez & Laflamme, 2008), as well as causing long-term health and adjustment problems (Rickwood et al., 2005). The purpose of the present study is to present data on students who engaged in an online mental health promotion intervention following a university-wide rollout of the program prior to and during the pandemic.

Some contextual factors may help explain why university students show poorer mental health outcomes. Many university students are young adults, a life phase characterised by reconciling a multitude of important and potentially stressful milestones (Arnett et al., 2014). Other factors relevant to university itself are also implicated, such as increased workloads and academic pressure (Mofatteh, 2020). Evidence has shown that the COVID-19 pandemic may compound these developmental and educational stressors due to the frequent and notable disruptions in student’s lives and education (Appleby et al., 2022; Howard et al., 2021). These disruptions include prolonged partial or full closure of university campuses, subsequent shift to an emergency online learning format, and the reduction or elimination of social contacts and cultural activities (Kohls et al., 2021). Results of a systematic review and meta-analysis of 27 studies (n = 706,415) that investigated mental health among university students during the pandemic (December 2019 to October 2020) showed that the prevalence of depression (39%) or anxiety (36%) among students increased during the COVID-19 pandemic, particularly for students studying outside of China (depression 60% and anxiety 60%) (Li et al., 2021).

The onset of the pandemic resulted in Australian universities switching to a complete distance learning modus, generally using some form of online instruction, which presented some difficulties for students such as limited social interaction and challenges maintaining motivation and self-discipline, which in turn may have jeopardised the mental health of students. However, many students identified advantages of online learning, including flexibility and self-directed learning (Kovacs et al., 2022). In addition, although the social isolation of lockdowns presented many challenges, students have identified positives in the experience including personal growth and improved close relationships, goal setting, and self-care due to increased time to devote to these areas (Owens & Cassarino, 2022). In addition, the pandemic prompted some academics to introduce innovative curriculum adjustments that have improved the flexibility and delivery of courses as illustrated in the modification of social work field placements in one Australian university (Morley & Clarke, 2020). Hence, the pandemic-related forced changes on higher education present a mix of challenges and opportunities for students.

Despite the high rates of pre-pandemic mental health difficulties, university students have reported relatively low rates of help-seeking behaviours. Between 36% − 50% of students with mental health problems do not receive targeted mental health support (Eisenberg et al., 2011; Zivin et al., 2009). Common barriers to accessing mental health care before and during the pandemic include uncertainty in the effectiveness of treatment, a lack of perceived urgency, stigma associated with seeking mental health support, a preference for self-reliance, poor mental health literacy, and cost and time constraints (Eisenberg et al., 2011; Martin, 2010). Evidence suggests however that if all students requiring care sought support at the same time, services would be unable to meet demand (Stallman & Kavanagh, 2016), with Australian university counselling services showing counsellor to student ratios of 1:4,340 (Stallman, 2012).

One option for addressing the elevated mental health problems and barriers to mental health help seeking among university students is the development of transdiagnostic web-based interventions (Pedrelli et al., 2015). A meta-analysis of web-based mental health promotion approaches within university samples reported small-moderate effect sizes for improvements in mental health (Harrer et al., 2019). Web-based therapies are also beneficial as they allow for cost-effective, flexible, and discrete mental health support, which address some of the aforementioned barriers to mental health help-seeking (Amstadter et al., 2011; Craske, 2012a; Petersen et al., 2019). Given the wide variation in mental health problems among university students, transdiagnostic mental health promotion programs that target comprehensive skills training are required. Transdiagnostic approaches are likely to have greater reach than disorder-focused approaches in the context of a larger target population. Such approaches also fit well within the university setting as they are more likely to encompass the major areas of dysregulation present in some students (Craske, 2012b; Hayes et al., 2012)

Acceptance and Commitment Therapy (ACT; Hayes et al., 2011) is an empirically supported transdiagnostic approach which fosters mental health skills through the development of psychological flexibility (Biglan et al., 2008). According to the ACT framework, psychological flexibility is enhanced through six core processes (acceptance, cognitive defusion, present moment awareness, flexible perspective taking, values, and committed action), which constitute important skills for enhancing mental health (Hayes et al., 2011). A review of 20 meta-analyses that included 133 randomised control trials of ACT (n = 12,477) reported effect sizes ranging from small to medium for anxiety and depression (Gloster et al., 2020). ACT has been shown to be a useful transdiagnostic intervention for university students that can be delivered in a variety of formats (Muto et al., 2011; Pakenham, 2015; Stafford-Brown & Pakenham, 2012). In particular, web-based ACT interventions with varied levels of practitioner involvement for university students have shown improvements in many areas including education values, depression, social anxiety, academic concern, wellbeing, values obstruction, mindful acceptance, life satisfaction, and academic performance (Chase et al., 2013; Levin et al., 2014, 2016; Rasanen et al., 2016).

A relatively brief, online ACT-based mental health promotion intervention for university students called YOLO (You Only Live Once) has been shown to improve mental health outcomes and skills in pilot and randomised controlled trials in one Australian university (Viskovich & Pakenham, 2018, 2020). Consistent with the transdiagnostic ACT approach, the intervention effects were evident across the distress continuum. The intervention also had wide reach whereby over 2,000 students responded to the invitation to take part in the study. Connection to personal values played a key role in mediating improvements in depression, anxiety, wellbeing, self-compassion, and academic performance. Qualitative data (Viskovich et al., 2021) showed that the YOLO values experiential exercises helped student learning and career development (Economics, 2017; PricewaterhouseCoopers, 2016). Following these promising evaluations of YOLO, the program was made available to all students enrolled at The University of Queensland in June 2019.

In order to investigate the feasibility of this university-wide roll out of the YOLO program, the first aim of this study is to present data on the numbers and characteristics of students who engaged in the YOLO program. The program was made available prior to and during the pandemic, which provided the opportunity to investigate possible differences between students who accessed the intervention before and after the onset of the pandemic, which shaped the following three aims. The second aim is to examine the levels of mental health reported by students relative to norms and investigate whether they vary as a function of engaging in the program before and after the onset of the pandemic. The third aim is to examine differences in socio-demographics between students who engaged in YOLO prior to and during the pandemic. The fourth aim is to investigate differences between the two groups in mental health. The fifth aim is to identify socio-demographic (including higher education) variables that are associated with variations in mental health.

Method

This is an exploratory cross-sectional study which involves the analysis of data collected from university students prior to commencing the YOLO online mental health promotion program in the context of a university-wide rollout of the intervention. Data were collected on mental health (distress and wellbeing) and socio-demographics.

Context of YOLO program rollout

The YOLO program was made available for university students to access from 25th June 2019 and advertised widely to students through social media posters, flyers during orientation, and hosted on the university’s website for wellbeing resources. To be involved in the YOLO program, participants followed a link and registered online using their verified student information. Prior to commencing baseline measures, participants could opt-in to having their data collected for research evaluation granted they met eligibility criteria: a) enrolled at the university, b) fluent in English, and c) 18 years or older. Following this, participants provided their informed consent, completed baseline measures, and were then given access to the program. Participants described in this study are those who self-selected and consented to have their data used for research evaluation purposes.

Data reported in this study includes participants who enrolled into the YOLO program before and during the COVID-19 pandemic. The Queensland Government declared COVID-19 a health emergency on 29^th January 2020, in this institution’s first of two semesters of the year. By 23rd March 2020 all undergraduate classes transitioned to online learning. Hybrid (both online and offline) classes were offered institution wide from 8th August 2020 (semester two) and this format was retained for the duration of the data collection period.

YOLO program structure and procedure

YOLO is a 4-week self-guided, web-based program for university students consisting of four 30–45-minute modules, each targeting one or two of the six core ACT processes (Viskovich & Pakenham, 2020). Program exercises were 5–15 minutes in duration. Modules could only be completed in sequential order. Previous exercises could be repeated, but participants could not progress forward until completion of the previous module. The program consisted of animated presentations, video clips, audio files and written exercises. For a full description of the program, please see Viskovich and Pakenham (2020).

Following provision of consent and completion of baseline measures, participants were provided access to the full program. Participants were instructed to aim to complete one module per week over 4-weeks, although they had flexibility to complete as desired. Participants were monitored online through the program administrator portal during the intervention period and all email communication to participants was automated. Participants received standard generic non-personalised reminder emails every 3–7 days to prompt program engagement, until the 4-week completion window expired. Standard reminder emails for completion of the pre-intervention assessment were sent at 1-week intervals until the window for completion lapsed. Standard emails were sent at relevant points in each module to reinforce program content (one per module). Standard emails were also sent on module completion, providing a short video recap and instructions for the next module. Program access was not revoked after the 4-week period; however, engagement and reminder emails were discontinued. The study received ethics approval by the university’s internal review board (ethical approval number 2019/HE001551).

Measures

Distress

The widely used 21-item Depression Anxiety and Stress Scale Short Form (Lovibond & Lovibond, 1995) measured depression, anxiety and stress. Participants rate how much each statement applied to them over the past week on a 4-point scale (0 did not apply to me at all to 3 applied to me very much or most of the time) with higher scores indicating higher distress. Items are summed for each subscale and multiplied by two for comparison to the 42-item parent measure. Cronbach Alpha’s for this study were as follows: depression .86, anxiety .78, and stress .78.

Wellbeing

The widely used 14-item Mental Health Continuum Short Form (L. M. Keyes, 2009) measured wellbeing. Each item is rated on a 6-point scale (0 never to 5 every day) with higher mean scores indicating higher wellbeing. Mean scores were calculated for overall wellbeing, plus the subscales: psychological, emotional, and social wellbeing. Cronbach Alpha’s for this study were as follows: overall wellbeing .95, psychological wellbeing .90, emotional wellbeing .91, and social wellbeing .89.

Socio-demographics

Information was obtained for age, gender, student domestic/international status, ethnicity, degree type (undergraduate, postgraduate [e.g., master], and research higher degree [e.g., PhD]) and degree title (open-ended: coded into university faculty).

COVID-19 status

A dichotomous COVID-19 status variable was developed to indicate whether a participant undertook the program prior to the onset of the COVID-19 pandemic in Queensland or during the pandemic by using the date of the pre-intervention survey as a determining factor. Hence, the “pre-COVID-19” group referred to the period 25^th June 2019 when the YOLO program was made available on the university website up to 28^th January, 2020, and the “during-COVID-19” group spanned 29^th January, 2020 to 25^th February, 2021; from here on referred to as the pre-COVID and during-COVID groups.

Statistical analyses

All variables were examined for accuracy of data entry and missing values, and whether they met the assumptions of multivariate analyses. To address the first aim, descriptive data on socio-demographics are presented for three groups: total sample, pre-COVID-19 group, and during-COVID-19 group.

Regarding the second aim, descriptive analyses were undertaken to investigate levels of depression, anxiety, stress, and wellbeing. Depression Anxiety and Stress Scale scores were compared to severity cut-off scores established by normative data and categories were calculated to investigate range of severity. Categories of positive mental health were calculated for wellbeing scores. Chi-square analyses were undertaken to investigate differences between the pre-COVID-19 and during-COVID-19 groups on depression, anxiety, and stress normative categories.

To address aim three, differences between the pre-COVID-19 and during-COVID-19 groups on socio-demographic variables were explored using a one-way ANOVA for the only continuous variable (e.g., age), and chi-square analyses for categorical variables (e.g., gender, ethnicity, degree, faculty, and domestic/international status).

To investigate aim four, differences between the pre- and during-COVID groups on distress and wellbeing, were investigated using ANCOVAs and Quade’s ANOVAs (Conover & Iman, 1982) whilst controlling for relevant covariates, and taking into account uneven sample sizes and a mixture of normal and non-normal data.

Regarding the fifth and final aim, we examined associations between socio-demographics and both wellbeing and distress using correlations for continuous socio-demographics (e.g., age) and MANOVAs and ANOVAs for categorical demographics (e.g., gender, ethnicity, degree, faculty, and domestic/international status) in the total sample.

Results

Participant flow and characteristics

Regarding the first aim, Figure 1 provides an overview of participant flow through the program. In total, 357 participants accessed the program during our recruitment period (25th June 2019–25th February 2022), of which 240 opted-in to the research evaluation component. Of these, 76 were removed due to mostly or fully incomplete data. This left a final sample of 164 individuals whose data were used in this study. Socio-demographics are described in Table 1. A total of 75.0% were female, 62.2% domestic students, 68.9% undergraduate, 12.2% postgraduate (e.g., master), and 12.2% were research higher degree (e.g., PhD) students. The mean age was 27.62 years (SD = 9.16; range = 18–59). Ethnicity of participants were Australian (44.5%), Asian (39%), European (7.3%), South American (1.8%), Canadian (1.2%), Aboriginal or Torres Strait Islander (0.6%), and Other (3.7%). All faculties of the university were represented with 20.1% health and behavioural sciences, 18.3% humanities and social sciences, 15.2% science, 11.6% business, economics, and law, 11% engineering architecture and information technology, 4.9% medicine.

Figure 1. Participant flow through the YOLO program.

Table 1. Participant characteristics for the total sample, and the pre-COVID-19, and during-COVID-19 groups.

Characteristic	Total (n = 164)		Pre-COVID-19 (n = 45)		During-COVID-19 (n = 119)
	%	n	%	n	%	n
Gender
Male	20.1	33	20.5	9	20.2	24
Female	75.0	123	68.9	31	77.3	92
Transgender male	2.4	4	2.2	1	2.5	3
Gender fluid	0.6	1	2.2	1	-	-
Non-binary	0.6	1	2.2	1	-	-
Missing	1.2	2	4.4	2	-	-
Degree level
1^st − 2^nd year Undergraduate	24.4	40	28.9	13	22.7	27
3^rd − 4^th year Undergraduate	28.1	46	20.0	9	31	37
Postgraduate (e.g., Master)	31.7	52	24.4	11	34.5	41
Research Higher Degree (e.g., PhD)	12.2	20	17.8	8	10.1	12
Missing	3.7	6	8.9	4	1.7	2
Student Type
Domestic	62.2	102	64.4	29	61.3	73
International	34.8	57	28.9	13	37.0	44
Missing	3.0	5	6.7	3	1.7	2
School Faculty
Health & Behavioural Sciences	20.1	33	22.2	10	19.3	23
Humanities & Social Sciences	18.3	30	6.7	3	22.7	27
Science	15.2	25	15.6	7	15.1	18
Business, Economics, & Law	11.6	19	15.6	7	10.1	12
Engineering, Architecture & Information Technology	11.0	18	13.3	6	10.1	12
Medicine	4.9	8	-	-	6.7	8
Missing	18.9	31	26.7	12	16.0	19
Ethnicity
Australian	44.5	73	48.9	22	42.9	51
Asian	39.0	64	35.6	16	40.3	48
European	7.3	12	8.9	4	6.7	8
South American	1.8	3	-	-	2.5	3
Canadian	1.2	2	-	-	1.7	2
Aboriginal/Torres Strait	0.6	1	-	-	0.8	1
Islander
Other	3.7	6	2.2	1	4.2	5
Missing	1.8	3	4.4	2	0.8	1
Mean age in years	M(SD) range		M(SD) range		M(SD) range
	27.62 (9.16);		27.83 (8.38);		27.54 (9.47);
	18–59		18–43		18–59

“Other” responses for ethnicity were chosen by participants and a written response was not provided.

Distress and wellbeing compared to normative data

With respect to aim two, distress and wellbeing scores for the total sample and the pre- and during-COVID-19 groups were compared to normative data. The percentage and numbers of participants in each of the normative categories for each scale are presented in Table 2. Regarding the total sample, the proportion reporting mild and moderate distress was 42.1% (n = 69) depression, 47% (n = 77) anxiety, and 35.9% (n = 59) stress, and rates of severe and extremely severe symptoms were 15.2% (n = 25) depression, 22.5% (n = 37) anxiety, and 13.5% (n = 22) stress. With respect to wellbeing, 7.9% (n = 13) fell in the languishing category, 55.5% (n = 91) moderate wellbeing and 31.1% (n = 51) the flourishing category.

Table 2. Percentage of participants in the normative categories for depression, anxiety, stress and wellbeing for the total sample, and the pre- and during-COVID-19 groups.

Outcomes	Total (n = 164) % (n)	Pre-COVID-19 (n = 44) % (n)	During-COVID-19 (n = 119) % (n)
Depression
Normal	39.6 (65)	38.6 (17)	39.5 (47)
Mild	16.5 (27)	18.2 (8)	16.0 (19)
Moderate	25.6 (42)	20.5 (9)	27.7 (33)
Severe	6.7 (11)	4.5 (2)	7.6 (9)
Extremely Severe Missing	8.5 (14) 3.0 (5)	11.4 (5) 6.8 (3)	7.6 (9) 1.7 (2)
Anxiety
Normal	27.4 (45)	38.6 (17)	22.7 (27)
Mild	10.4 (17)	13.6 (6)	9.2 (11)
Moderate	36.6 (60)	15.9 (7)	44.5 (53)
Severe	9.1 (15)	9.1 (4)	9.2 (11)
Extremely Severe	13.4 (22)	15.9 (7)	12.6 (15)
Missing	3.0 (5)	6.8 (3)	1.7 (2)
Stress
Normal	47.6 (78)	54.5 (24)	44.5 (53)
Mild	20.7 (34)	15.9 (7)	22.7 (27)
Moderate	15.2 (25)	15.9 (7)	15.1 (18)
Severe	9.8 (16)	2.3 (1)	12.6 (15)
Extremely Severe	3.7 (6)	4.5 (2)	3.4 (4)
Missing	3.0 (5)	6.8 (3)	1.7 (2)
Wellbeing
Languishing	7.9 (13)	6.8 (3)	8.4 (10)
Moderate	55.5 (91)	54.5 (24)	56.3 (67)
Flourishing	31.1 (51)	27.3 (12)	31.9 (38)
Missing	5.5 (9)	11.4 (5)	3.4 (4)

Chi-square analyses were used to investigate whether the proportion of participants in each of the distress and wellbeing categories differed between the pre- and during-COVID-19 groups. Due to small numbers in some categories the following groups were combined for analysis in order to meet the expected count assumption (Kim, 2017). Distress categories were combined as follows: normal, mild/moderate, and severe/extremely severe and entered separately into the model together with the COVID-19 status variable. Results indicated one significant finding for anxiety with more participants in the during-COVID-19 group reporting mild/moderate anxiety (54.7%) and fewer reporting normal levels of anxiety (23.1%) compared to the pre-COVID-19 group (31.7% and 41.5% respectively; χ2 (2, N = 158) = 7.23, p > .03).

Differences in socio-demographics between the pre- and during-COVID-19 groups

To investigate aim 3, differences between the pre- and during-COVID-19 groups on socio-demographics were explored using a one-way ANOVA for the only continuous variable age, and chi-square analyses were used for the remaining socio-demographics. Regarding age, there was no statistically significant difference between pre- and during-COVID-19 groups (p = .95), For the chi-square tests, Fisher’s Exact Test was used for gender due to the expected cell count violation (Fisher, 1922), whereas Pearson Chi-Square statistic was used for the tests conducted on the remaining variables. Categorical variables that had small numbers in some sub-categories (e.g., gender, degree, ethnicity, faculty), were recategorised to meet the expected count assumption. Final groupings were as follows: gender (male n = 33, female n = 123, trans/non-binary/fluid n = 6); degree (1^st/2^nd year undergraduate n = 40, 3^rd/4^th year undergraduate n = 46, postgraduate = 52, research higher degree n = 20); ethnicity (Australian n = 73, Asian n = 64, Other n = 26); and school faculty (health and behavioural sciences/humanities, and social sciences n = 63; business, economics and law/engineering, architecture, and information technology n = 37; and medicine/science n = 33). The pre- and during-COVID-19 groups significantly differed on one variable, degree X² (3, N = 153) = 21.20, p = <.001. Specifically, more 1^st/2^nd year undergraduates (32.5%), postgraduate students (27.5%), and research higher degree students (20.0%) were in the pre-COVID-19 group compared to the during-COVID-19 group (22.1% 1^st/2^nd year undergraduates, 8.0% postgraduate students, and 10.6% higher research degree students) and there were fewer 3^rd/4^th year undergraduates in the pre-COVID-19 group (20.0%) compared to the during-COVID-19 group (59.3%).

Differences in distress and wellbeing between the pre- and during-COVID-19 groups

To investigate aim four, we first conducted preliminary analyses which revealed that psychological wellbeing, depression, anxiety, and stress violated the assumptions of normality and homogeneity of variance. Hence, the non-parametric analysis Quade’s ANOVA was conducted on these variables (Conover & Iman, 1982). ANCOVAs were conducted on total, emotional, and social wellbeing. In all analyses, the dependent variables were wellbeing and distress, the independent variable was pre- vs. during-COVID-19 group, and the covariate was degree level. Quade’s ANOVA was performed in SPSS. The results of these analyses are summarised in Table 3 and showed that the two groups did not differ on the mental health outcomes.

Table 3. Means and standard deviations for distress and wellbeing for the total sample, and the pre- and during-COVID-19 groups.

Outcomes	Total Sample	Pre-COVID-19	During- COVID-19	Statistics
	M (SD)	M (SD)	M (SD)	F	p
Distress
Depression	12.83 (9.12)	13.61 (10.16)	12.65 (8.74)	0.01^b	.92
Anxiety	11.21 (6.93)	10.98 (8.89)	11.38 (6.08)	0.77^b	.38
Stress	16.19 (7.92)	14.83 (8.33)	16.70 (7.77)	0.59^b	.45
Wellbeing
Total	2.90 (1.14)	2.78 (1.22)	2.94 (1.12)	0.05^a	.82
Psychological	2.98(1.24)	2.83 (1.34)	3.04 (1.21)	0.30^b	.58
Emotional	3.28 (1.21)	3.12 (1.24)	3.32 (1.20)	0.26^a	.61
Social	2.58 (1.27)	2.51 (1.32)	2.59 (1.26)	0.07^a	.79

Total sample n = 164, pre-COVID-19 group n = 39–41, during-COVID-19 group n = 114–117. Variations in sample sizes are due to missing data. ^aANCOVAs ^bQuade’s ANOVA.

Relationship between socio-demographics and mental health outcomes

To examine the relationships between socio-demographics, and both wellbeing and distress in the total sample, the fifth study aim, correlations were used for the only continuous socio-demographic age, and ANOVAs and MANOVAs were used for the remaining categorical socio-demographic variables. These results are summarised in Table 4. Only two socio-demographics were significantly related to one of the distress or wellbeing subscales – degree level and age. Regarding degree level and wellbeing subscales, all three subscales were significant: psychological wellbeing F (5, 146) = 2.84, p = .018; partial η² = .089; emotional wellbeing F (5, 146) = 2.34, p = .045; partial η² = .074; and social wellbeing F (5, 146) = 2.77, p = .022; partial η² = .085. Post hoc analysis with a Bonferroni correction revealed that 3^rd year undergraduate students had significantly lower wellbeing scores than postgraduate students for psychological wellbeing (M difference = −1.17, p = .010[95% CIs −2.1650, −.1679], emotional wellbeing (M difference = −1.11, p = .014[95% CIs −2.0814, −.1295], and social wellbeing (M difference = −1.17, p = .013[95% CIs −2.1881, −.1439].

Table 4. Relations between socio-demographics and distress and wellbeing variables.

Socio-Demographics	n	Total Distress		Depression		Anxiety		Stress		Total Wellbeing		Psychological Wellbeing		Emotional Wellbeing		Social Wellbeing
		Mean (SD)	F	Mean (SD)	F	Mean (SD)	F	Mean (SD)	F	Mean (SD)	F	Mean (SD)	F	Mean (SD)	F	Mean (SD)	F
Degree Level	138		3.35*		4.26*		2.06		1.86		3.10*		2.84*		2.34*		2.73*
1^st year UG	29	36.34 (15.78)		11.31 (7.77)		10.97 (4.77)		14.07 (6.27)		2.91 (1.08)		3.07 (1.20)		3.23 (1.18)		2.51 (1.28)
2^nd year UG	10	43.64 (25.54)		12.73 (8.87)		13.09 (8.46)		17.82 (9.69)		2.78 (1.15)		2.67 (1.29)		3.37 (1.08)		2.56 (1.27)
3^rd year UG	18	48.20 (22.14)		17.70 (11.13)		12.50 (8.20)		18.00 (7.79)		2.16 (0.90)		2.23 (1.07)		2.50 (1.09)		1.87 (1.06)
4^th year UG	26	50.50 (17.29)		17.67 (8.14)		13.67 (5.55)		19.17 (7.15)		2.74 (1.10)		2.78 (1.17)		3.21 (1.25)		2.40 (1.21)
Postgraduate	48	35.88 (21.31)		10.49 (8.39)		9.88 (6.42)		15.51 (8.50)		3.31 (1.18)		3.40 (1.23)		3.61 (1.26)		3.03 (1.32)
RHD	20	32.40 (18.14)		9.90 (6.97)		8.60 (7.14)		13.90 (7.50)		2.86 (1.12)		3.04 (1.32)		3.32 (1.04)		2.37 (1.16)
School Faculty	133		1.97		2.35*		0.64		1.96		0.52		0.28		1.39		0.74
HABS	32	47.63 (22.47)		16.38 (10.43)		12.13 (7.47)		19.13 (8.83)		2.64 (1.12)		2.80 (1.14)		2.89 (1.35)		2.30 (1.23)
HASS	29	35.93 (13.86)		10.14 (5.66)		10.34 (5.51)		15.45 (6.52)		2.93 (1.19)		2.93 (1.35)		3.41 (1.13)		2.65 (1.43)
BEL	19	46.00 (28.49)		14.95 (11.10)		12.84 (8.09)		18.21 (10.41)		2.86 (1.17)		2.91 (1.41)		3.07 (1.33)		2.67 (1.13)
EAIT	18	43.56 (21.66)		15.56 (10.90)		12.00 (6.96)		16.00 (6.44)		2.74 (1.16)		2.76 (1.10)		2.98 (1.27)		2.56 (1.40)
Medicine	7	28.57 (13.79)		8.00 (5.66)		8.86 (5.40)		11.71 (6.68)		3.31 (1.05)		3.27 (1.25)		3.88 (0.59)		3.00 (1.23)
Science	25	37.76 (16.45)		12.24 (7.73)		11.20 (5.54)		14.32 (6.34)		2.90 (1.12)		3.04 (1.28)		3.44 (1.18)		2.41 (1.14)
		Pearson Correlation	p	Pearson Correlation	p	Pearson Correlation	p	Pearson Correlation	p	Pearson Correlation	p	Pearson Correlation	p	Pearson Correlation	p	Pearson Correlation	p
Age	154	−.14	.091	−.06	.438	−.27*	<.001*	−.06	.47	.16	.06	.14	.09	.13	.12	.15	.06

*Indicates significance at p = .05. UG = undergraduate. RHD = research higher degree. ATSI = Aboriginal and Torres Strait Islander. HABS = Health and Behavioural Science; HASS = Humanities and Social Sciences; BEL = Business, Economics and Law; EAIT = Engineering, Architecture and Information Technology. Missing data not reported therefore totals may be < N = 164.

Regarding degree level and distress subscales, depression was the only significant finding F (5, 147) = 4.26, p = .001; partial η² = .126. Post hoc analysis with a Bonferroni correction revealed that 3^rd year undergraduate students had significantly higher depression scores than post-graduate students (M difference = 7.26, p = .023[95% CIs .5444, 13.9756]) and 4^th year undergraduate students had significantly higher depression scores than post-graduate students (M difference = 7.23, p = .012[95% CIs .9235, 13.5299] and research higher degree students (M difference = 7.78, p = .045[95% CIs .0817, 15.4516].

ANOVAs were conducted on total wellbeing and total distress scores. Only one socio-demographic was significantly related to both wellbeing and distress: degree level. Regarding degree level and total wellbeing, F (5, 146) = 3.10, p = .01; partial η² = .096, post hoc analysis with a Bonferroni correction revealed that 3^rd year undergraduate students had significantly lower wellbeing scores than postgraduate students (M difference = −1.15, p = .004[95% CIs −2.0663, −.2401]). Regarding degree level and total distress, F (5, 147) = 3.35, p = .007; partial η² = .102, post hoc analysis with a Bonferroni correction revealed that 4^th year undergraduate students had significantly higher distress scores than research higher degree students (M difference = 18.1, p = .043[95% CIs .2751, .35.9249]). The significant correlation between age and anxiety showed that younger age was associated with higher anxiety (r = −.27, p = <.001).

Discussion

The purpose of the present study was to present data on students who engaged in the YOLO mental health program following university-wide rollout prior to and during the pandemic. A total of 240 students accessed the program over a 12-month period and of these, 164 provided reliable data. Overall, a relatively small number of students accessed YOLO which may be explained by several factors. First, the program link was difficult to find on the university mental health website, potentially making independent access difficult. Second, well over 2,000 students had already accessed the program during the pilot and RCT studies in the years prior to the university-wide rollout. Finally, conversations in university social pages suggested that some students felt they needed more university-related structural changes (e.g., longer assignment extensions) to aid in the management of their distress, rather than a mental health program.

Uptake of YOLO was much stronger during the initial RCT comparative to the university rollout, likely due to the more comprehensive dissemination of program information to university students and staff (e.g., emails to large groups of students, class presentations, and meetings with student support staff) in the RCT. This likely made the program more visible to students, which in turn may signal the importance of mental health and work to de-stigmatise, consequentially helping students access mental health support (Giroux et al., 2019; Withers et al., 2021). In light of the public health impacts of COVID-19, a concentrated effort on behalf of universities to generate awareness around mental health problems and facilitate access to support is vital (Harris et al., 2022).

Consistent with prior literature on mental health help-seeking behaviour (Liddon et al., 2018; Parent et al., 2016), markedly more females (77.2%) than males accessed YOLO. The age of most participants fell in the emerging adult developmental phase (18–29 years; Arnett et al., 2014), which reflects the age range of most university students reported in recent investigations into university student wellbeing (Browning et al., 2021; Dodd et al., 2021). In view of data suggesting international students report higher levels of mental health problems (Skromanis et al., 2018), it is important to note that approximately half of the students who accessed YOLO reported international student status despite these students accounting for approximately 30% of higher education enrolments in Australia (Australian Institute of Health and Welfare, 2019). There was also wide ethnic diversity represented in the sample. There was a relatively even spread of participation across all faculties of the university except medicine, which had the lowest representation.

When examining mental health outcomes as normative categories, there was a higher proportion of students reporting mild-moderate levels of anxiety in the during-COVID-19 group compared to the pre-COVID-19 group. The lack of substantial differences in mental health between the two groups is in contrast to evidence from a meta-analysis that showed the prevalence of depression and anxiety among university students had increased during the COVID-19 pandemic (Li et al., 2021). However, no Australian study was incorporated in the review and relative to other countries included, Australia (and in particular Queensland) had been less severely affected by COVID-19 in the first year of the pandemic. This context may have made it easier for university students in YOLO to benefit from some of the positives associated with the move to online learning and the increased time afforded by lockdowns to attend to personal growth, close relationships, goal setting, and self-care reported by university students in other studies (Owens & Cassarino, 2022),

Of all the socio-demographics considered in this study, only two were related to wellbeing – age and degree level. This is in contrast to previous research on student wellbeing that has highlighted factors such as ethnicity, gender, and student status as being implicated in wellbeing (Lipson et al., 2018; Prowse et al., 2021; Skromanis et al., 2018). Regarding degree level, results demonstrated that 3^rd and 4^th year undergraduate students reported lower wellbeing, and higher depressive symptomology comparative to students in other years of study. Deficits on the two key mental health dimensions, wellbeing and distress, suggest that students within the 3^rd and 4^th years of study are at greater risk of psychosocial impairments (Keyes, 2002).

These findings call into question what may be unique about the 3^rd and 4^th years of undergraduate study which might drive these poorer mental health outcomes. One study found that later stage undergraduates demonstrate significantly higher distress comparative to those in the early or mid-progression stages (Elias et al., 2011), perhaps due to increased academic pressures at the later-stages of their degree and future-related anxiety. This proposition aligns with our data, as Australian undergraduate degrees are typically only 3–4 years in duration, therefore, many of the 3rd and 4th year students in our sample were likely to be grappling with pressures related to entering the workforce. Evidence from Australian universities suggests that during-COVID-19, undergraduate students had greater anxiety about the future relative to postgraduate students (Dodd et al., 2021). This is perhaps due to repercussions of COVID-19 public health measures forcing redundancies and business closures – of which Australian young Australians were most impacted by (Churchill, 2020).

Aligned with our finding that younger age was associated with greater anxiety, it may also be that comparative to postgraduate cohorts, 3^rd and 4^th year undergraduates have under-developed psychological resources to cope with this additional stress. This proposition aligns with findings showing greater use of adaptive coping styles (Monteiro et al., 2014), and higher wellbeing (Dodd et al., 2021), in postgraduate cohorts comparative to undergraduates. Therefore, academic stress, future anxiety, and underdeveloped coping resources during-COVID-19 may have propelled the elevated stress for 3^rd and 4^th year undergraduates in the present study.

The finding that there was an increase in 3^rd and 4^th year students who accessed the YOLO program during-COVID-19 suggests that these students were cognisant of their elevated stress and that they understood the importance of accessing support resources. This stands in contrast to previous findings which show that undergraduates, comparative to postgraduates, are less likely to access psychological support (Rafal et al., 2018). It is also important to acknowledge that whilst in this study 3^rd and 4^th year students appeared to demonstrate elevated risk, literature also posits first-year undergraduate students as having high mental health vulnerability (Duffy et al., 2020; Farrer et al., 2016). For example, one study found that being a first-year student was a predictor of depressive symptomology at an Australian university (Farrer et al., 2016). International longitudinal evidence from the United States of America also shows that this group had significantly worse mental health during-COVID-19 (Fruehwirth et al., 2021). Therefore, it is important to adopt contextually sensitive perspectives on the mental health of university students and our findings indicate that online, self-guided mental health programs such as YOLO should be made available to all students seeking support.

The data we have presented on the university-wide rollout of YOLO should be interpreted in the context of methodological limitations characteristic of descriptive studies reporting on the dissemination of interventions in “real world” settings. First, convenience non-random sampling from one university and the bias towards female participants limits the generalisability of findings. Second, of the 357 students who expressed interest in undertaking the program, only 46% logged into the program, consented to research participation, and provided useable data. Hence, there may have been a response bias related to the sample upon which analyses were performed. Third, examination of the effects of participation in YOLO pre-COVID-19 vs. during-COVID-19 was based on convenience sampling and was not an a priori study design element.

The extent to which YOLO is acceptable across culturally diverse audiences should be further explored given the large numbers of international students attending Australian universities and their vulnerability to mental health problems. However, it should be noted that there was good representation of international students and wide ethnic diversity represented in the present study, and in the pilot and RCT evaluations of YOLO (Viskovich & Pakenham, 2018, 2020). In addition, international vs. domestic student status did not predict completion of YOLO in the pilot or RCT studies (Viskovich & Pakenham, 2018, 2020). ACT has also been used in bibliotherapy to improve mental health and decrease anxiety and depression in international university students (Muto et al., 2011).

Conclusions

Descriptive data on the university-wide rollout of YOLO shows that it was accessed by students representing a diverse range of socio-demographic characteristics. The poorer intervention uptake rate of the university-wide rollout relative to that of the prior RCT suggests that online mental health promotion interventions need to be assertively and widely promoted using diverse mediums (e.g., social media, in-person, and hard copy print) directed at students and staff. In addition, access to the online intervention should be prominent, and quickly and easily accessible, particularly in light of social distancing measures used during the pandemic. Data from this study suggest that 3^rd and 4^th year students may be particularly vulnerable to mental health difficulties and should be targeted with respect to the dissemination of information on mental health promotion interventions. Students who accessed YOLO pre-COVID-19 did not substantially differ across most mental health dimensions from those who accessed it during-COVID-19, although there was a higher proportion of students who reported mild-moderate anxiety symptoms during-COVID-19 than pre-COVID-19. Given the social distancing measures used during the pandemic and the potential adverse mental health impacts of COVID-19, the provision of an online mental health promotion intervention that does not require in-person contact is an important advantage. Consideration of the present data along with findings from the pilot, RCT, and qualitative studies that have evaluated YOLO, suggests that online mental health promotion programs are an important and necessary resource for all university students.

Author contributions

Shelley Viskovich and Kenneth Pakenham conceptualised the study. Shelley Viskovich conducted statistical analyses. Shelley Viskovich, James Fowler, and Kenneth Pakenham all provided critical and conceptual contributions to the writing of the original manuscript. All authors provided critical revisions to form the finally submitted manuscript.

Research involving human participants or animals

This work was approved by The University of Queensland Human Research Ethics Committee approval number: 2019/HE001551.

Informed consent

All participants provided their informed consent to participate in this research and have their data used in such a way.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Due to the nature of this research, participants of this study did not agree for their data to be shared publicly, so supporting data is not available.