A quantitative research design was utilised employing a cross-sectional, descriptive design to establish and analyse the predictors and correlates of ADSI among university students in Namibia. The participants completed the self-administered questionnaire that gathered the demographic information and predictive factors, including academic, lifestyle, current mental illness, personality, alcohol and substance abuse, depression, and anxiety symptoms, including a binary enquiry of suicidal ideation. A sampling frame consisted of UNAM and NUST campuses, being the only two public universities in Namibia. Two NUST main and Eenhana campuses were purposively selected as they represent NUST sites nationwide. For UNAM, a systematic simple random sampling method was applied: from a list of all 12 campuses, every fourth campus was selected, ensuring equal probability of inclusion. This yielded four UNAM campuses: UNAM main and Hage Geingob (Khomas region, Central), Southern campus (Southern campus in //Karas region), and Oshakati campus (Oshana region, North). To test the research hypothesis, a structured questionnaire was distributed in person to respondents to collect relevant data on the main predictors of ADSI among undergraduate students in Namibia.

Data were collected from NUST and UNAM multiple campuses across Namibia, including in the Khomas, Oshana, Ohangwena, and Omusati regions. These regions were selected based on suicide prevalence: Oshana, Ohangwena, and Omusati ranked among the top five with the highest suicide rates, while the Khomas and //Karas regions have the lowest suicide rate in Namibia.¹⁵

The study population included undergraduate students aged 18–24 years, specifically those in their second and third academic years who volunteered to participate. The self-administered questionnaires were distributed to 780 participants recruited through a purposive sampling technique, which lasted for a period of 2 months.¹⁶ The response rate was 83%, representing 578 completed questionnaires from which the study findings were drawn. The sample size was calculated using Yamane’s formula (1973), n = N ÷ (1 + N × e²), where N = population size, 1 = a constant, and e = margin of error, set at 0.05 or the significance level.¹⁵

The self-administered questionnaires were designed in English and consisted of three sections. The first section, Section A, included demographic items targeting respondents aged 18–24 years who are current undergraduates in their second and third academic years from the participating campuses. Demographic data were collected from the respondents. Section B contained predictors associated with ADSI among respondents. The hypothesis included economic, family, and academic factors. Section C examined symptoms of anxiety and depression, as well as the evidence of suicidal ideation, which contributed to the prevalence of ADSI among public university students in Namibia. The questionnaires were handed to the participants in person.

The Patient Health Questionnaire-9 (PHQ-9) is a widely used, validated tool to assess depression, consisting of nine items that measure the frequency of depressive symptoms over a specific period.¹⁷ Patient Health Questionnaire-9 was used to screen for depression symptoms among the participants. Depression symptoms listed in DSM-5 include: (1) lack of interest, (2) depressed mood, (3) sleeping disturbances, (4) reduced energy levels, (5) eating disorders, (6) negative feelings about oneself, (7) psychomotor retardation, and (8) difficulty in concentrating.¹⁸ Respondents were instructed to rate symptoms on a 4-point Likert scale, indicating how often they experienced depression symptoms in the past two weeks (0 = not at all, 1 = several days, 2 = more than half the days, and 3 = every day). Scores for each category ranged from 0 (never) to 3 (every day). The PHQ-9 scores ranged from 0 to 27 and were categorised as follows: 0–7 indicating none to minimal depression symptoms, 8–16 mild depression, 17–24 moderate depression, and 25–27 severe depression. The Cronbach’s α coefficient was 0.874 for depression, indicating strong internal consistency.

The study modified Beck’s Anxiety Inventory (BAI) sheet, which included seven items used to determine the prevalence of anxiety symptoms among university students.¹⁹ The seven items used a four-point Likert scale, where 0–4 indicates minimal anxiety symptoms, 5–9 indicates mild anxiety symptoms, 10–14 indicates moderate anxiety, and 15–21 indicates severe anxiety. The internal consistency was confirmed by Cronbach’s α coefficient of 0.805, indicating strong internal consistency.

Suicidal ideation was assessed using the Columbia Suicide Severity Rating Scale Version 14 January 2009.²⁰ The scale included binary questions: ‘Yes’ for those who are suicidal, and ‘No’ for non-suicidal individuals. One score indicates the lowest level, while five represents severe suicidal ideation. The open-ended questionnaire asked participants to describe how they would like to die, to evaluate suicidal ideation in the past 2 weeks. The Cronbach’s alpha was 0.726, indicating an acceptable reliability.

Descriptive statistics, including frequencies and percentages, were used to summarise the demographic characteristics of the respondents. Descriptive statistics were analysed in SPSS version 30.²¹ To understand the correlation between anxiety, depression, and suicidal ideation, Pearson’s correlation coefficients were calculated. Hierarchical multiple regression analysis was performed to identify predictors of ADSI. Binary logistic regression was used to determine the prevalence of suicidal ideation categorised as ‘Yes’ or ‘No’. A p-value < 0.05 was considered statistically significant. All variables, including demographic characteristics, prevalence, and predictors, were analysed to identify independent predictors of ADSI among students.

The relevant institutions granted permission to conduct the study. Ethical approval was obtained from the Namibia University of Science and Technology Research Committee, with reference number FHNRAS:60/2023, and from the Namibian Ministry of Health and Social Services, with reference number 22/3/1/2. The six campuses where the study was conducted approved the data collection. The study adhered to ethical standards for health science research. In this quantitative study, steps were taken to ensure anonymity, voluntary participation, the right to withdraw, and respondent confidentiality. The research tool’s cover page included an informed consent form, which participants signed before the data collection began. The researcher collected data using self-administered questionnaires distributed to students who met the inclusion criteria, were available, and gave their consent to participate during scheduled campus visits. No emotional distress was observed among participants during the study. However, had any such responses emerged, appropriate referrals for counselling would have been initiated.

The data set included a total of 578 undergraduates from NUST and its Eenhana satellite campus, as well as the UNAM main campus, Hage Geingob campus, Oshakati campus, and Southern campus. Second- and third-year students enrolled at public universities were recruited for the study’s survey (Table 1). Most respondents were female, at 55.4%, while 44.6% were male. They consisted of 53.6% second-year students and 45.4% third-year students. Among the respondents, 25.3% lived in the campus hostel, and 72.9% lived off campus.

The prevalence of anxiety and depression symptoms was determined using frequencies and percentages to assess the extent of these conditions among respondents. As shown in Table 2, 43.3% experienced mild anxiety symptoms, 42.6% had moderate anxiety symptoms, and 8.1% exhibited severe anxiety symptoms. Regarding depression, 21.8% reported mild symptoms, 43.1% moderate symptoms, and 30.1% had severe depression symptoms.

The findings reveal various types of suicidal ideation among university students who answered ‘Yes’ (Table 3): passive suicidal ideation, 25.1%; suicidal ideation without a plan, 19.9%; active suicidal ideation, 15.2%; and suicidal thoughts in general, 13.5%. Alarmingly, active suicidal ideation with a plan was 9.7%. Although most participants responded ‘No’ across all variables, it is essential to recognise that ‘Yes’ indicates a significant mental health concern among students.

The research aimed to establish the correlation between ADSI and students. Pearson’s correlation coefficient was calculated to explore the relationships between anxiety, depression, and suicidal ideation (Table 4). There was a positive correlation between anxiety and depression (r = 284, p < 0.001), and between anxiety and suicidal ideation (r = 166, p < 0.001). Similarly, depression significantly correlated with suicidal ideation (r = 0.362, p < 0.001). Current mental illness was positively associated with anxiety (r = 0.162, p < 0.001), depression (r = 0.276, p < 0.001), and suicidal ideation (r = 0.261, p < 0.001). Family conflict showed a weaker but significant positive correlation with anxiety (r = 0.088, p < 0.034), depression (r = 111, p < 0.007), and suicidal ideation (r = 0.117, p = 0.004), while a weak but statistically significant negative correlation was observed with suicidal ideation (r = –0.111, p = 0.007) indicating that resilience has a protective influence.

A series of hierarchical regressions was conducted to identify the predictors of ADSI among students at public universities. Table 5 shows that the significant predictors of anxiety included current mental illness (β = –0.162, p < 0.001), personality (β = 0.131, p = 0.001), and faculty (β = 0.077, p = 0.031). Significant predictors of depression included current mental illness (β = –0.276, p < 0.001), family conflict (β = –0.144, p < 0.001), and personality (β = 0.087, p = 0.019). For suicidal ideation, significant predictors included current mental illness (β = –0.198, p < 0.001), faculty (β = –0.123, p < 0.001), family conflict (β = –0.171, p < 0.001), and age (β = –0.103, p = 0.007).