Suicide among university students: prevalence, risks and protective factors

ABSTRACT Background: Research evidence on suicide in Ghana so far has focused mostly on suicide in the adult population and less work on the younger population such as university students and other vulnerable groups such as children, youth and the aged. Aims: This study was conducted to determine lifetime and current prevalence of suicidal ideation or attempts and identify the associated risks and protective factors among university students in Ghana. Methods: Using a cross-sectional design, 1003 university students (507 males and 496 females) with a mean age of 20.5 years (SD = 5.95) were administered questionnaires that measured suicide, psychological distress, self-esteem and subjective wellbeing. Results: We found the following prevalence rates of suicidal behaviours: ideations 15.2%, attempted 6.3%, death wishes 24.3% and suicidal plan 6.8%. Psychological distress was a risk factor for both suicidal ideation and suicidal attempt. Subjective wellbeing was protective of suicide attempt while self-esteem was protective of suicidal ideation. Conclusions: These findings underscore the need for school-based mental health-promoting programmes that enhance young people’s self-esteem, reduce psychological distress and boost subjective wellbeing.


Introduction
Suicide is a public health concern of great importance given the enduring devastating effects it has on families, friends and communities. Annually, it claims over 800,000 lives (WHO, 2014) and many of these are young people within the 15-29 age bracket (Nock et al., 2013;WHO, 2012); a situation with grave economic implications for our world today (Kinyanda, Hoskins, Nakku, Nawaz, & Patel, 2012). Hence, the World Health Organization's Gap Action Programme of 2008 included suicide as one of the priorities for serious attention and action.
Statistics from census report in Ghana (Ghana Statistical Service, GSS, 2013) shows that over 30% of recorded deaths categorized under deaths by accident, violence, homicide or suicide, occurred among the youth within the age range of pre-university and university populace. Existing evidence suggests that suicide among youth is becoming a daily reality in Ghana (Quarshie, Osafo, Akotia, & Peprah, 2015). Yet the extent to which reported risk factors reflect and affect suicidality in university students is yet to be examined. In Ghana there is still much to be understood in order to effectively design contextspecific interventions.
Suicidal ideation among university students is undoubtedly a multifaceted phenomenon with globally acknowledged negative and deleterious effects on families, friends and even the socio-economic development of countries within which it occurs. Among young people, especially university students, suicide is the second leading cause of death besides self-inflicted injuries, making this population an at-risk group (Nock et al., 2013;Taliaferro et al., 2009aTaliaferro et al., , 2009bWHO, 2012). Therefore, importance of research for greater understanding of this menace cannot be overemphasized.
Theoretically, the present study examines suicide from a human ecological model perspective in a bid to understand the occurrence of suicide among university students (Bronfenbrenner, 1977(Bronfenbrenner, , 1979Hawton & van Heeringen, 2002;van Orden et al., 2010). Just as suicide risks, as well as buffers, may be complex to unravel the model offers a conceptualization of suicide from a multifaceted approach integrating the individual, social and familial factors that may be associated with suicide occurrence (e.g. Ayyash-Abdo, 2002;Henry, Stephenson, Hanson, & Hargett, 1993); and thus afford a broader environmental context interpretation for a better understanding and elucidation of risk factors.
Scrutiny of the research on suicide in Ghana shows a gap as far as information on suicide among university students is concerned. This study aims to address the gap by focusing the research on university students and so add to existing information for the design of effective interventions for this at-risk group. Therefore, the twofold-specific objectives of the study were to determine lifetime and current prevalence of suicidal ideation or attempts; and to identify associated risks and protective factors. Based on existing literature, the study predicts that students who report higher levels of psychological distress but have lower levels of self-esteem and subjective wellbeing would be more likely to exhibit suicidal behaviours (herein ideations and/or attempts). In other words, a negative relationship is expected between suicidal ideations or attempts and these variables. Likewise, students with poorer self-esteem are expected to be less satisfied with their lives and be more psychologically distressed. Since greater subjective wellbeing has been associated with greater perceptions of control over important life choices (Owusu-Ansah, 2008;Owusu-Ansah, Agyei-Baffour, & Edusei, 2012), a greater subjective wellbeing is also expected to protect against suicidal ideations and/or attempts.

Study design and participants
The study was a cross-sectional design as it aimed to establish a possible relationship among various variables. This method of study allows the researchers to establish and measure relationships among study variables. This design was used because it entails surveys or other pre-structured means that helped to obtain a common dataset on some pre-selected variables. One thousand and three (1003) students purposively selected participated in the study. Participants were recruited through the Know Your Status' screening exercise: a screening exercise annually organized by the KNUST Counselling Centre (KCC) during which students voluntarily test for HIV/AIDS and Hepatitis B status, blood group and undergo breast examinations. Participation in the study was limited to students who attended the KCC programme. The sample size was determined a priori using GPower software version 3.1 (Faul, Erdfelder, Lang, & Buchner, 2007). With a power value of 0.80, a significant level of 0.10, and a minimum effect size of 0.10, the minimum sample size was calculated to be 448. Thus, a sample size of 1003 was found to be adequate.

Setting
The study was conducted among students of the Kwame Nkrumah University of Science and Technology (KNUST). The KNUST is one of the public and government-funded universities in Ghana located in Kumasi, in the Ashanti Region. It has a student population of over 40,000 and staff strength of approximately 3610. The university comprises of six colleges, namely College of Agriculture and Natural Resources (CANR), College of Art and Built Environment (CABE), College of Engineering (CoE), College of Humanities and Social Sciences (CoHSS), College of Health Sciences (CHS) and College of Science (CoS).

Procedure
Participants were provided adequate information about the study and had their questions about the study answered. Having understood what was required of them, those we voluntarily choose to participate were given questionnaires to complete. All key ethical principles of informed consent, voluntary participation and confidentiality were strictly adhered to throughout the study. The study protocol was reviewed and approved by the Committee on Human Research, Publication and Ethics (CHRPE/RC/253/17) of the School of Medicine and Dentistry (SMD), Kwame Nkrumah University of Science and Technology.

Demographics
Demographic characteristics assessed included sex, marital status, religion, year of study, college, residence and nationality.
Suicide-related items. Lifetime suicidal attempt, for instance, was measured by asking 'have you personally attempted to end your life?' to which the participants answered 'yes' or 'no'. Current suicidal attempt was elicited with the question 'If yes, how many times has this happened in the last 12months?' Lifetime suicidal ideation was measured by asking participants 'Have you ever thought about ending your own life? Lifetime death wishes was measure by asking participants 'Is there a time in your life when you wished were dead?' Lifetime suicide plan was measured by asking participants 'Have you ever had a clear plan for killing yourself'.
Self-esteem. Participants' self-esteem was measured with the Rosenberg Self-Esteem Scale (RSE). This is a 10-item questionnaire that asks respondents to indicate how they feel about themselves. Some of the items include 'On the whole, I am satisfied with myself', 'At times I think I am no good at all'. Participants indicate the extent to which they agree with these statements. Responses are anchored on a four-point Likert scale ranging from strongly agree (1) to strongly disagree (4). Five of the items are negatively worded while five are positively worded. Negatively worded items are reversed scored such that higher scores indicate more positive self-esteem. An individual's score is obtained by summing his/her responses. Scores ranged from 10 to 40. The validity of the Rosenberg Self-Esteem Scale is well established within the Ghanaian context (Nyarko, 2017).
Subjective wellbeing. The Ryff's Psychological Wellbeing (PWB) Scale an 18-item questionnaire was used to measure psychological wellbeing. Some of the items are 'When I look at the story of my life, I am pleased about how things have turned out' and 'I don't have a good sense of what it is I am trying to accomplish in life'. Responses are scored on a five-point Likert scale ranging from strongly disagree (1) to strongly agree (5). Negatively worded items are reversed scored and an individual's score is obtained by summing responses. Scores range from 18 to 90 with higher scores indicating higher or better subjective wellbeing and satisfaction with one's life. The PWB Scale has shown good psychometric properties in adult samples of all ages and backgrounds (Curhan et al., 2014;Ryff & Keyes, 1995) as well as college students (Gloria, Castellanos, Scull, & Villegas, 2009;Ryff, Keyes, & Hughes, 2003).
Psychological distress. This variables was assessed using the Depression Anxiety Stress Scale (DASS-21) (Lovibond & Lovibond, 1995) The DASS-21 is a well-structured questionnaire consisting of three subscales of seven items each measuring depression, anxiety and stress. The depression subscale assessed the presence of self-blame, pessimism and loss of enjoyment. The anxiety subscale assessed the state of persistent apprehension and worry, accompanied by physical symptoms of sympathetic activation. The stress subscale assessed the state of over-arousal, tenseness and the inability to relax. The internal consistencies of the three subscales were .71, .79 and .81 for depression, stress and anxiety, respectively. Items on the DASS-21 are scored on a three-point Likert scale ranging from 0 = does not apply to me at all to 3 = applies to me very much or most of the time. The reliability of the DASS-21 has been established within the Ghanaian context (Arhin, Oppong Asante, Kugbey, & Oti-Boadi, 2019;Asante, 2012). Since, the DASS-21 distinguishes between depression, anxiety and stress as distinct manifestations of psychological distress, the total of the subscales were used in this study.

Data analysis
The Statistical Package for Social Sciences (SPSS) version 24.00 was used to analyse the data. Descriptive statistics was used to describe the nature of the data through the generation of frequencies and percentages. Pearson's correlation analysis was used to test the relationships among the study variables. As the outcomes variables were categorical in nature (i.e. suicide ideation and attempts), logistic regression analyses were conducted to determine the predictors of both suicidal ideations and attempts. The results from the logistic regression analyses are presented as odds ratio (OR) and 95% confidence interval (CI). All statistical tests were performed using a two-tailed examination, and a p value of 0.05 or less was considered statistically significant.

Ethics statement
The study protocol was reviewed and approved by the Committee on Human Research, Publication and Ethics (CHRPE/RC/253/17) of the School of Medicine and Dentistry (SMD), Kwame Nkrumah University of Science and Technology.

Demographic characteristics of the sample
The demographic characteristics of the sample are presented in Table 1. The sample included 1003 university students (507 males and 496 females) with a mean age of 20.53 years (SD = 5.95). The majority (86.4%) of the students were single and half of them (50.3%) were first-year students. Approximately 94% of the students were Christians, 5.8% were Muslim and the remaining 0.5% belonged to other religious affiliations. The same number of students (23%) were from both the Humanities and Health Sciences colleges, while 21.7% of the students were from the Physical Science College. The majority of the students were Ghanaians (97.9%) and residents on campus (74.9%).

Prevalence of suicide
The prevalence of current and lifetime prevalence of suicide among students are presented in Table 2. Lifetime prevalence of suicidal attempts among the participants was 6.3%. About 15.2% reported suicidal ideation and almost a quarter of the participants had had death wishes. Approximately 7% have had clear suicide plans in their lifetime. With regard to lifetime prevalence, 23.8% had attempted suicide more than once while 25.0% had suicidal ideation more than once in the past year. Approximately 19% reported that they had more than once made plans for suicide in the past year.

Factors associated with suicide ideation and suicide attempt
The results of the logistic regression analysis of the predictors of suicide ideation and attempt among students are presented in Table 3. Suicidal attempt was significantly predicted by psychological wellbeing and psychological distress. Positive subjective or psychological wellbeing buffers against suicidal attempt (AOR = 0.95; CI 0.95, 1.05) while excessive psychological distress increases the likelihood of attempting suicide (AOR = 1.07; CI 1.04, 1.09). Positive self-esteem served as a buffer against suicidal ideation (AOR = 1.05; CI 0.90, 0.99) and the presence of psychological distress increased the likelihood of having suicidal ideations (AOR = 1.07; CI 1.04, 1.09). In suicidal ideation, selfesteem and psychological distress matter; while in suicidal attempts, subjective wellbeing and psychological distress are the risk factors. In other words, self-esteem seems to play a significant role in suicidal ideation as subjective wellbeing is in suicidal attempt aside the influence of psychological distress in both instances.

Discussion
This study was conducted to determine lifetime and current prevalence of suicidal ideation or attempts and identify the associated risks and protective factors among university students in Ghana. We found the following prevalence rates of suicidal behaviours: ideations 15.2%, attempted 6.3%, death wishes 24.3% and suicidal plan 6.8%. Psychological distress was a risk factor to both suicidal ideation and suicidal attempt. Subjective wellbeing was protective of suicide attempt while self-esteem was protective of suicidal ideation.

Suicide prevalence rates
The current prevalence rates of suicidal behaviours: ideations 15.2%, attempted 6.3%, death wishes 24.3% and suicidal plan 6.8% are fairly consistent with a recent study of nursing students on suicide in Ghana (Quarshie, Cheataa-Plange, Annor, Asare-Doku, & Lartey, 2019) as well as secondary school students (Oppong Asante et al., 2017;Baiden et al., 2019). In 2008, Hjelmeland and colleagues reported that only 4% of Ghanaian college students compared to 7% Ugandans and 8% Norwegians reported that they might consider suicide under some circumstances. Relatively lower percentages (2% and 4%) of Ghanaian college students reported own suicidal acts last year and earlier in life respectively in the current study. Eshun (2003) indicated that the low scores on suicide ideation by Ghanaian college students, compared with their counterparts elsewhere, are influenced by religion and negative attitudes towards suicide. The scores on suicide, although comparable to the global situation of suicide thoughts and behaviours of college students (Mortier et al., 2018) are slightly lower compared to samples in South Africa (e.g. Bantjes et al., 2016;van Niekerk, Scribante, & Raubenheimer, 2012). These low scores, more than a decade ago, are changing.

Psychological distress and self-esteem predicting suicidal ideation
Our findings showed that psychological distress predicted suicidal ideation among university students. The presence of psychological distress increased the likelihood of suicidal ideations, consistent with general findings of a strong link of suicide to psychological distress (Tanji, Tomata, Zhang, Otsuka, & Tsuji, 2018;Walker, McGee, & Druss, 2015). It was also observed that low self-esteem is associated with suicide. Self-esteem has been identified in the WHO (2014) report on suicide prevention as an important protective buffer against stressors and suicidal behaviours for young people. A strong association between self-esteem difficulties and suicide ideation among young adults has been documented (Mirsu-Paun, 2015). The present findings corroborate these earlier reports.
Psychological distress and wellbeing predicting suicidal attempt We expected higher suicide attempts among students with greater psychological distress and higher subjective wellbeing to protect against such behaviours. This prediction was confirmed. Psychological distress positively correlated with suicide attempt, but subjective wellbeing was inversely related. Consistent with our finding, some reports show that subjective wellbeing is related with low rates in suicide (Hsu, Chang, & Yip, 2019) and correlates negatively with suicide intention (Sisask, Värnik, Kolves, Konstabel, & Wasserman, 2008). In one study among students, results indicated that the effects of all types of bullying on suicidal ideation were mediated by subjective wellbeing (Lucas-Molinaa, Pérez-Albénizb, & Fonseca-Pedrerob, 2018). In an ecological study analysing the geography of suicide in New Zealand among young people, Snider (2011) reported a correlation between low levels of psychological wellbeing and increased suicidal behaviours and self-harm for both males and females. Other studies have consistently shown that subjective wellbeing is generally protective of young people's mental health in Eastern Europe (Helliwell, 2007;Lim, Cappa, & Patton, 2017); and in this study, protective of suicidality, in particular. Thus, what we have found in the present study reflects the critical role of subjective wellbeing in the mental health of Ghanaian university students.

Implications for interventions
These findings have implications for scaling up mental health services in this population. Frist, mental health in Ghana is already underfunded and in many universities policies to improve the mental health of students are barely existent. As a country, Ghana has no clearly delineated policies for improving adolescent and young adults' mental health. A major implication of this study is a clarion call to intensity efforts to promote mental health in universities in Ghana. For example, in Australia, the health-promoting schools approach model has been adopted (Wallace, Holloway, Woods, Malloy, & Rose, 2011) as recommended by the WHO to develop students' social, emotional and behavioural competencies; identify at-risk students who might have such problems, treat and prevent further disorders among such individuals. Such school-based mental health-promoting approaches could pay attention to building young people's self-esteem, reduce their psychological distress and boost subjective wellbeing. Secondly as recommended in a recent study which examined suicidal behaviour among nursing trainees (Quarshie et al., 2019), the importance of periodic screening of students and the strengthening of onsite mental health counselling services for them is equally important. Since reports indicate only a moderate level of psychological help-seeking behaviours among university students in Ghana , mental health services for this population may not be successful if presented in the traditional waiting, reactive mode but rather in the seeking mode (Maseko, Maunganidze, Mambende, & Maphosa, 2017;Orford, 2008). That is, the practitioner need not wait for individuals to initiate contact, but rather makes efforts to bring services to clients, understand presenting problems and develop preventive intervention response programmes.

Limitations of the study
This study is not without some limitations. First, there is a cultural revulsion against suicide (Adinkrah, 2011(Adinkrah, , 2012Osafo, 2016) and thus hampers accurate reporting on such behaviours. The current prevalence rates might simply be a tip of the iceberg and may not reflect the true state of suicidal behaviour among university students in Ghana. Secondly, causal inferences cannot be made from cross-sectional data, though the direction and strength of associations give us clues. Thirdly, caution should be exercised in the generalization of findings to other populations. These limitations notwithstanding, the present study, to the best of our knowledge, is the first attempt to estimate suicidality among a large sample of university students in Ghana. The relatively large sample size enhances the generalizability of the findings to students in institutions of higher learning in Ghana. It prepares the grounds for future works to further study this underserved population and engender policies to address their mental health needs in Ghana.

Conclusions
Findings from this study indicate that suicidal behavioursideations, plans, attempts, death wishesare prevalent among university students in Ghana. The results also show that psychological distress and low self-esteem are predictors of suicidal ideation. Whereas psychological distress predicted suicide attempt, subjective wellbeing attenuated it. These findings underscore the need for the development of mental health services in this population in Ghana.

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