Barriers to accessing support for mental health issues at university

ABSTRACT Student mental health is an issue of great concern for universities, with rising numbers of mental health problems being reported and students reporting issues with accessing support. The current study, using a participatory research framework, investigated the possible barriers preventing students from accessing support, in terms of help-seeking intentions and actual help-seeking behaviour. Three hundred and seventy-six current UK students completed a questionnaire which measured help-seeking and possible barriers including perceived public stigma, self-stigma, educational impact, disclosure, coping behaviours and current mental health symptoms. Findings indicated that self-stigma, in particular, was a barrier to accessing support. Disclosure, educational impact, previous diagnosis, suspected diagnosis and mental health symptoms also interacted with help-seeking. These findings have implications for universities in tackling the barriers preventing students accessing support for their mental health.

4 reluctance to seek support was linked to perceptions of stigma. The experience of stigma likely exacerbates mental health problems and may prevent people from seeking help (Vidourek et al. 2014). Students may also experience self-stigma -an individual's own internalised attitudes, often adopting the stigmatising public discourse towards their own mental health (Corrigan, Watson and Barr 2006) and self-stigmatising the act of help-seeking (Tucker et al. 2013). Self-stigma and public stigma independently contribute to mental health help-seeking (Bathje and Pryor 2011). In one systematic review, both types of stigma were the biggest barrier to help-seeking for young people (Gulliver, Griffiths and Christensen 2010).
A systematic review conducted by Clement et al. (2015) found stigma was only the fourth most important barrier to help-seeking in the general population. Instead, concerns around disclosing mental health issues was the top barrier. Individuals may differ in their willingness to disclose mental health difficulties due to concerns over shame and embarrassment (Kahn and Hessling 2001). Gulliver et al. (2010) found that concerns around confidentiality after disclosure represented a major source of concern for many young people.
An individual may fear that a breach of confidentiality following disclosure may lead to stigma from peers or family (Gulliver et al. 2010). Consequently, this fear may prevent an individual from disclosing their mental health issues to a professional.
The ability to cope with psychological distress must also be consideredit may be that individuals who are good at coping are less likely to seek help. The act of copingefforts to reduce or prevent stress, harm or threat (Carver and Connor-Smith 2010) is often vital in determining an individuals' psychological adjustment and well-being (Monzani et al. 2015). Actively attempting to improve circumstances and solve problems is advantageous (Carver and Connor-Smith 2010), linking to higher rates of attained goals and positive affect (Mackay et al. 2011), as well as higher academic performance (Struthers, Perry and Menec need. Individual differences in help-seeking intentions should also be considered as a barrier to accessing support. With help-seeking intentions, people may report that they plan to seek help if they experience psychological distress . The theory of planned behaviour suggests that intention is the strongest predictor of actual behaviour (Abraham, Sheeran and Henderson 2011). The theory argues that intention is predicted by a person's attitude towards the specific behaviour, their subjective norm of significant others (Montano and Kasprzyk 2015) and perceived control in the ability to carry out the behaviour (Abraham et al. 2011). However, intentions do not necessarily translate into actual behaviour (de Bruijn, Out and Rhodes 2014), particularly in the domain of mental health (Li 2016). Rickwood et al. (2005) conducted a systematic review of adolescents' help-seeking intentions for mental health problems and found mixed evidence on the relationship between intentions and behaviour. The relationship between intention and actual help-seeking has been suggested to be dependent upon specific mental health problems and barriers people face .
As shown by the number of students with mental illness, and the noted challenges in receiving adequate support, there is an issue with students accessing and receiving treatment from mental health issues while at university. Highlighting the barriers that students face may be the first step to implementing strategies to overcome these barriers. The current study therefore aimed to examine the possible barriers preventing UK students from accessing support for mental health issues. Based on the discussed literature, the following variables were examined as barriers to both intentions to seek help and actual help-seeking behaviour: perceived public stigma, self-stigma, educational impact, disclosure, coping and current mental health functioning in terms of depression, anxiety and stress. We hypothesised that

Participants
Three hundred and seventy-six current students took part with a mean age of 20.73 (SD = 3.34). Most participants were female (83.5%, n = 314; male 15.7%, n = 59; other/prefer not to say .8%, n = 3). Eighty-two percent were born in the United Kingdom. Most were White British (76.9%) or from another White Caucasian background (12%); 5.6% were Asian/Asian British, 3.7% were multi-ethnic and .5% were Black/Black British. Most were heterosexual (76.8%) with 12.6% reporting that they were bisexual and 7.7% homosexual. Participants were recruited via word of mouth, social media and the UK student mental health charity 'Student Minds' website.
Information regarding university and student status is displayed in Table 1. This information shows that most participants were undergraduate, full-time Home students, recruited from the South East of England.
[Insert Table 1 here] Ethical approval was obtained via [blinded for peer review] and all participants gave full informed consent before participating.

Materials
Participants completed an online survey, with the following measures presented using the 'Qualtrics' survey software.

Help-seeking.
Help-seeking intentions for mental health problems was measured using the General Help Seeking Questionnaire . Participants rated their likelihood of seeking help for personal or emotional problems, from four informal sources (parents, other family member, friends and intimate partner) and three formal sources (GP, mental health professional and helpline), on a 7-point Likert scale ('extremely unlikely' (1) to 'extremely likely' (7)). Higher scores represented higher intentions to seek help. Participants also rated the item 'I would not seek help from anyone'. Responses were categorised into three subscales (informal or formal source, or no-one). Internal consistency was acceptable for intentions from informal sources (ω =.62) and from formal sources (ω =.69).
Actual help-seeking behaviour was measured by asking participants "which of the following services have you accessed with regards to your mental health since you have been at university?" Participants could select from nine sources, such as a mental health advisor, counsellor or GP. Participants could also indicate if they had not accessed any support.
Responses were coded into either 'accessed support of any form' or 'none'.

Perceived Public Stigma.
Perceived stigma for receiving psychological help was measured using the Adapted Stigma Scale (Golberstein, Eisenburg and Gollust 2008). Participants rated agreement with five statements, such as 'It is a sign of personal weakness or inadequacy to receive treatment for mental health problems'. Each item was scored on a 5-point scale ('strongly disagree' (1) to 'strongly agree' (5)). Scores were summed to derive a total ranging from 5 to 25, with higher scores representing higher stigma. Internal consistency was good (ω =.77).

Self-Stigma.
Self-stigma was measured using the Self-Stigma of Seeking Help Scale (Vogel, Wade and Haake 2006). Participants rated agreement with ten statements on a 5-point scale ('strongly disagree' (1) to 'strongly agree' (5)), such as'If I went to a therapist, I would be less satisfied with myself'. Scores were summed to obtain a total ranging between 10 and 50, with higher scores indicating a greater self-stigma. Internal consistency was very good (ω =.87).

Educational impact.
Impact on academic functioning was measured with two questions rated on a 5-point scale ('never' (1) to 'a great deal' (5)) from Einsberg et al. (2007): 'How often have you missed classes/other academic obligations in the past four weeks due to mental health difficulties?' and 'How often has your academic performance been affected in the past four weeks due to mental health difficulties?' Scores ranged from 2 to 10 with higher scores indicating greater difficulties with academic functioning. Internal consistency was very good (α =.83).

Coping behaviours.
Coping behaviours for mental health issues were measured using the Brief Cope Scale (Carver 1997). Participants were asked to consider 'the ways they had been coping with mental health problems in their life'. They rated how often they had been engaging in 21 positive coping behaviours, such as 'I've been getting comfort and understanding from someone', on a 4-point scale ('I haven't been doing this at all' (1) to 'I've been doing this a lot' (4)). Scores ranged between 21 to 84 with higher scores indicating greater use of positive coping behaviours when dealing with mental health issues. Internal consistency was very good (ω =.88).

Disclosure.
Disclosure was measured using the Distress Disclosure Index (Kahn and Hessling 2001).
Participants rated agreement with 12 statements such as 'When I feel upset, I usually confide in my friends' on a 5-point scale ('strongly disagree' (1) to 'strongly agree' (5)). Scores were summed to obtain a total ranging from 12 to 60, with higher scores indicating more disclosure. Internal consistency was excellent (ω =.94).

Current mental health symptoms.
Depression, anxiety and stress were measured using the short-form version of the DASS (DASS-21; Henry and Crawford 2005). Participants rated 21 statements on a 4-point scale ('did not apply to me at all' (0) to 'applied to me very much or most of the time' (3)). There were three subscales with seven items each, corresponding to depression, anxiety and stress.
Scores were summed for each subscale and multiplied by two to derive three totals ranging from 0 to 42. Higher scores represented higher symptoms of depression, anxiety and stress respectively. Internal consistency was excellent for the depression subscale (ω =.92), very good for the anxiety subscale (ω =.88) and stress scale (ω =.87).
Participants were asked to self-report previously diagnosed mental health conditions. Details were recorded on specific diagnoses and age of diagnosis. Participants were also asked whether they believed that they had an undiagnosed mental health condition, indicated which condition they suspected, and whether they were seeking diagnosis.

Procedure
The current study used a participatory approach (Cornwall and Jewkes 1995), putting students at the centre of this research. Four students were the experts on student mental health and worked with an academic as researchers to develop the aims of the project. First, the group discussed the problem that needed addressed. Based on lived and peer experience, the group identified that accessing mental health services and the need for support should be the research focus, given the high prevalence of mental health problems for students.
Specifically, the aim was to investigate the barriers that may prevent students from accessing support.
Second, the group investigated the pre-existing literature to establish current knowledge regarding the aim, and to identify how the current study could add to the literature. The group also researched potential barriers to accessing support. Next, the group decided how to address the aim. Quantitative questionnaire methods were deemed appropriate and could be used to reach a large sample of students.
In the survey, participants first completed demographic information (including about their studies) and information about their mental health conditions, if applicable. Participants were then asked whether they had received support for their mental health during their time at university. Participants then completed measures on educational impact, self-stigma, perceived stigma, current mental health, coping behaviour, disclosure and help-seeking intentions. The survey took approximately 30 minutes and was part of a larger study with some data not reported here. Upon completion, participants were debriefed, signposted to useful resources and had the option to enter a prize draw. Data was collected between December 2017 and February 2018.

Design
A cross-sectional survey was used to examine the barriers to students accessing support for mental health. Hierarchical linear regression was used to test whether the proposed variablesstigma, self-stigma, educational impact, coping behaviour, disclosure and depressive, anxiety and stress symptoms -could predict help-seeking. Binary logistic regression was used to test whether these variables could predict whether participants had sought help for mental health issues while at university.

Mental health conditions.
In the sample, 40.2% (n = 151) self-reported a diagnosed mental health condition. Specific conditions reported are detailed in Table 2. For undiagnosed conditions, 29 participants (7.7%) reported that they suspected a mental health condition and were currently seeking diagnosis. Further, 137 participants (36.4%) reported that they suspected a mental health condition but were not seeking diagnosis. 55.9% (n = 210) did not suspect an undiagnosed mental health condition. The most commonly suspected conditions were anxiety and mood disorders (Table 2). Overall, 55.9% (n = 210) had accessed support for their mental health since they had been at university.

Barriers to mental health support.
Intentions to seek help.
First, three separate hierarchical linear regressions were used, with help-seeking intention scores (formal, informal or no-one) as the outcome variables. In the first step for all regressions, age, gender, previous diagnosis (yes/no) and suspected diagnosis (yes/no) were entered. In the second step, educational impact, perceived stigma, self-stigma, disclosure, coping, depression, anxiety and stress scores were entered.
For help seeking intentions from formal sources, the first step explained 8.3% of the variance in intentions, and the model was significantly better at predicting the outcome than the mean alone (F(4, 341) = 7.65, p<.001). The final model explained 16.0% of the variance (a significant increase, p<.001) and the model was significant (F(12, 341) = 5.24, p<.001).
In the final model (Table 3), gender was a significant predictor of intentions, with male participants (M = 34.03, SD = 6.73) more likely to intend to seek help than female participants (M = 31.26, SD = 6.45). Those with a diagnosed mental health condition (M = 10.38, SD = 4.05) were more likely to intend to formally seek help than those who did not (M = 8.55, SD = 4.15). Self-stigma significantly predicted help-seeking intentions, such that with increasing self-stigma, help-seeking intentions decreased. With increasing educational impact and perceived public stigma, help-seeking intentions from formal sources significantly increased.
[Insert Table 3  In the final model (Table 3), age was a significant predictor of intentions to seek help from informal sources, such that with age participants were less likely to seek help from informal sources. Those who suspected an undiagnosed mental health condition (M = 15.75, SD = 4.87) were less likely to intend to seek help from informal sources than those who did not (M = 18.82, SD = 4.85). Disclosure significantly predicted help-seeking intentions, such that with increasing disclosure, help-seeking intentions increased. With increasing stress, help-seeking intentions from informal sources increased, but with increasing depression symptoms, intentions significantly decreased.
For help seeking intentions from no-one, the first step explained 4.3% of the variance in non-help-seeking intentions, and the model was significantly better at predicting the outcome than the mean alone (F(4, 338) = 3.79, p =.005). The final model explained 45.4% of the variance (significant increase, p<.001) and was significant (F(12, 338) = 22.57, p<.001).
Self-stigma was a significant predictor of intentions to not seek help in the final model (Table 3): with increasing self-stigma, the likelihood of not seeking help increased. With increasing depression symptoms, intentions to not seek help significantly increased. With increasing disclosure, intentions to not seek help decreased.
Actual help-seeking behaviour. The variables which could predict actual help-seeking behaviour were examined. Descriptive statistics for each variable according to whether the participants had accessed support, are shown in Table 4.
[Insert Table 4 here] Binary logistic regression was used to test whether the variables could predict whether or not participants had accessed support for their mental health while at university. The predictors entered were age, gender, previous diagnosis, suspected undiagnosed condition, educational impact, perceived public stigma, self-stigma, help-seeking intentions (formal, informal and no-one), disclosure, coping, depression, anxiety and stress. Results are shown in Table 5.
[Insert Table 5 here] Those with a previous diagnosis were 4.24 times more likely to access support since they had been at university, and those who suspected an undiagnosed mental health condition were 2.41 times more likely than those who did not. With increasing educational impact, help-seeking intentions from formal sources and stress symptoms, the odds of accessing support increased. With increasing self-stigma, the odds of accessing support decreased.

Discussion
This study examined potential barriers preventing students from accessing mental health support. The examined barriers included perceived public stigma, self-stigma, educational impact, disclosure, coping and current mental health symptoms. In the study, four in ten participants reported a diagnosed mental health condition, with anxiety and mood disorders commonly reported. Further, a third of participants suspected a mental health condition, but were not seeking a diagnosis, and only 8% were seeking a diagnosis. Over half of the sample (55%) had accessed support for their mental health since being at university. In analyses examining the barriers to accessing support, both in terms of intentions to seek help and actual help-seeking behaviour, self-stigma associated with seeking help was a significant barrier.
Self-stigma was a significant predictor of a lack of help-seeking intentions and actual help-seeking behaviour, suggesting that self-stigma attached to help-seeking affects both intentions and behaviour. This finding supports previous qualitative research (Quinn et al. 2009) and research with non-student populations (Bathje and Pryor 2011;Gulliver et al. 2010). Corrigan (2004) suggests that self-stigmatising mental health issues affects selfesteem, linking to feeling ashamed and demoralised. Tucker et al. (2013) noted that selfstigma around seeking help (rather than self-stigma of having a mental illness) relates to selfblame and explains more variance in help-seeking intentionssuch that the act of helpseeking is the barrier, rather than being labelled 'mentally ill'. The current study measured only self-stigma of seeking help, therefore future studies would benefit from also measuring the self-stigma of mental illness. Nonetheless, the current findings indicate that the selfstigma of seeking help acts as a barrier to accessing support.
Notably, self-stigma did not predict help seeking intentions from informal sources. This finding may relate to informal sources offering help only in the form of social support. Indeed, with greater disclosure, participants were more likely to seek help from informal sources, but disclosure did not predict help-seeking from formal sources. This supports prior research that family and friends are more often recipients of disclosure than professionals for those with suicidal ideation (Husky et al. 2016) and severe mental illnesses, with higher levels of disclosure associated with higher perceived social support (Pahwa et al. 2017).
Disclosure to family and friends is thought to be beneficial due to the empathy received upon disclosure (Bril-Barniv et al. 2017).
Notably, only intentions to seek help from formal sources predicted actual helpseeking behaviour. Together, these findings suggest that disclosing mental health issues to informal sources does not necessarily relate to actual help-seeking. Nonetheless, social support is still important in mental health recovery (Hendryx, Green and Perrin 2009). For young people, Gulliver et al. (2010) found a preference to seek help from family and friends rather than professionals. Gulliver et al. (2012) also found that people are more likely to seek help from informal sources who provide social support rather than professionals.
Nevertheless, help-seeking from professional services is a fundamental pathway towards support for mental health issues.
While self-stigma and public stigma are separate but interrelated concepts (Corrigan 2004), perceived public stigma only positively predicted help-seeking intentions from formal sources. Perceived public stigma did not predict intentions in other analyses, nor did it predict actual help-seeking behaviour. Previous studies have found the perceived public stigma does not predict help-seeking in students (Eisenberg et al. 2009). Further, Vogel et al. (2006) found that self-stigma fully mediates the relationship between public stigma and help-seeking intentions for students. Together, these findings suggest that self-stigmaalthough influenced by perceived public stigmais a significant barrier to support for students, and more research is needed to understand the relationship between stigma and mental health in students.
Educational impact was also a significant predictor of actual help-seeking behaviour and help-seeking intentions from formal sources. As mental health issues begin to affect academic performance, it makes sense that students would be more inclined to seek help.
This finding also demonstrates the impact mental health conditions could have on academic functioningsuch as reducing concentration and motivation and increasing fatigue and social withdrawal (Kitzrow 2003). Positively, with increasing educational impact participants were more likely to seek helpsuggesting that they feel able to seek help when their studies are affected. However, this finding may also indicate that more needs to be done to prevent mental health problems affecting academic studies in the first place. Interventions focusing on improving students' resilience to academic-related stress may be beneficial (Hartley 2011).
In terms of mental health symptoms, increasing depressive symptoms predicted decreased help-seeking from informal sources and increased likelihood of seeking help from no-one. With depression, symptoms such as feeling guilty may relate to decreased motivation to seek help (D'Avanzo et al., 2012). Barney et al. (2009) found that participants were reluctant to talk to informal sources about depression due to concerns over appearing 'weak' or being 'viewed less' as a person. For actual help-seeking, increasing stress symptoms predicted greater likelihood of accessing support. As stress levels rise, seeking support to manage stress may be less stigmatised than depression, however greater understanding of stress at university is needed (Robotham and Julian 2006).
In this study, regardless of whether they were accessing support or not, participants reported high rates of depression, anxiety and stress in comparison to general population means (Henry and Crawford 2005). Many were therefore experiencing elevated levels of mental health symptoms, yet may not perceive that they need help or believe that their experience is "normal" . These findings may reflect normalisation of the narrative around elevated mental health issues in students. In general, mental health conditions are more prevalent and have their onset in adolescence or young adulthood (Patel et al. 2007). However, circumstances at universities in the UK, such as academic pressures and financial issues, may exacerbate mental health issues (Brown 2016). An Australian study comparing students to age-matched non-students, found that mental health conditions were more prevalent in students (Stallman 2011). In UK universities, 21% of students rated themselves as having low anxiety, compared to 43% of age-matched non-students (Brown 2016). These findings suggest there are specific stressors within universities that aggravate mental health issues.
Finally, previous diagnosis predicted help-seeking intentions from formal sources and actual help-seeking behaviour. In Hartrey, Denieffe and Wells' (2017) systematic review, students with good previous experience of mental health services were more likely to seek help than those who were not familiar with these services. In the current study those who suspected a mental health condition were less likely to intend to seek help from informal sources, but they were more likely to have sought actual help. One explanation is that these individuals felt that formal support would be more beneficial to them than informal support.

Limitations and future research
One significant limitation of the current study is its generalisability. Most participants attended universities in South East England, reflecting the location of the study authors. Most were also White British. Mental health statistics in England vary by ethnicity: Black, Asian and mixed-race individuals report more mental health concerns than White people (Baker 2018). Another limitation is that most participants were female. Women are more likely to take part in survey research (Sax, Gilmartin and Bryant 2003) and the topic may have been of greater interest to females, who are more likely to experience internalising problems (Kuehner 2003). Despite the limited generalisability, the current study still has value in understanding the barriers some students face.
The study also assumes no change in mental health issues or help-seeking behaviours during the academic year. Knoesen and Naude (2012) found that first-year students initially 'languished' and struggled academically, however over the academic year developed skills that allowed them to 'flourish'.  suggests that longitudinal research would indicate such fluctuations and provide better understanding of mental health problems at university. Thus, future longitudinal research would be highly beneficial.
Future research should also examine ways of reducing the self-stigma. Those with mental health conditions who self-stigmatise commonly state that treatment provides little personal empowerment, linking to reduced motivation to seek treatment (Corrigan 2004). Therefore, interventions that target self-stigma and increase personal empowerment may increase help-seeking behaviour. Lucksted et al. (2011) piloted an intervention to reduce selfstigma in those with mental illnesses and found a significant decrease in self-stigma and increase in personal empowerment after the intervention. Future research should endeavour to test further interventions in this area.

Conclusion
Overall, the current study identified possible barriers preventing students from accessing support, both in terms of help-seeking intentions and actual help-seeking behaviour. Self-20 stigma was a significant barrier of help-seeking behaviour, and disclosure, educational impact, previous diagnosis, suspected diagnosis and current mental health symptomology also interacted with help-seeking. This study makes an important contribution in deepening our understanding of the barriers faced by UK students in relation to mental health. More work needs to be done, however, on taking down these barriers, and promoting better student wellbeing.