Although non-suicidal self-injury (NSSI) is known typically to begin in adolescence, longitudinal information is lacking about patterns, predictors, and clinical outcomes of NSSI persistence among emerging adults. The present study was designed to (1) estimate NSSI persistence during the college period, (2) identify risk factors and high-risk students for NSSI persistence patterns, and (3) evaluate the association with future mental disorders and suicidal thoughts and behaviors (STB).

Using prospective cohorts from the Leuven College Surveys (n = 5915), part of the World Mental Health International College Student Initiative, web-based surveys assessed mental health and psychosocial problems at college entrance and three annual follow-up assessments.

Approximately one in five (20.4%) students reported lifetime NSSI at college entrance. NSSI persistence was estimated at 56.4%, with 15.6% reporting a high-frequency repetitive pattern (≥five times yearly). Many hypothesized risk factors were associated with repetitive NSSI persistence, with the most potent effects observed for pre-college NSSI characteristics. Multivariate models suggest that an intervention focusing on the 10–20% at the highest predicted risk could effectively reach 34.9–56.7% of students with high-frequency repetitive NSSI persistence (PPV = 81.8–93.4, AUC = 0.88–0.91). Repetitive NSSI persistence during the first two college years predicted 12-month mental disorders, role impairment, and STB during the third college year, including suicide attempts.

Most emerging adults with a history of NSSI report persistent self-injury during their college years. Web-based screening may be a promising approach for detecting students at risk for a highly persistent NSSI pattern characterized by subsequent adverse outcomes.

Family-based interventions have shown to be effective in the treatment of major depression. In a recent RCT, systemic family groups led to significantly higher rates of treatment responders and higher rates of patients no longer using antidepressant medication at 15 months follow-up compared to standard care.

In order to better understand the underlying mechanisms of change, helpful treatment experiences of depressive patients and their partners in the family groups were explored.

34 depressive patients and their partners were asked to fill out an open-ended questionnaire investigating helpful experiences during treatment at the end of a 6 bi-weekly group session cycle. Responses were analyzed using the interpretative phenomenological approach (IPA) by three different researchers.

Ten recurring themes were reported as helpful by both the depressed patients and their partners:

1. (1) Presence,

2. (2) Cohesion,

3. (3) Support,

4. (4) Openness,

5. (5) Discussion,

6. (6) Self revelation,

7. (7) Insight,

8. (8) Progress,

9. (9) Observational experiences and

10. (10) Guidance from therapist.

The results of this study help to get insights in the therapeutic factors, which should be emphasized in family groups.

Psychiatric emergency services (PES) are a main entry for patients with acute mental disorders. Most of them are voluntarily treated and a shared decision making model is implicitly used. Not much is known about decision making capacity (DMC) of patients and the utility of a shared decision making model in PES.

To assess neuropsychological abilities of DMC in a convenience sample of patients admitted to a PES.

Patients were assessed when discharged from the PES. A DSM-IV-TR diagnosis was generated. Neuropsychological abilities were assessed by the Trail Making Test (TMT), Stroop Colour-Word Test (SCWT) and the Mini-Mental State Examination (MMSE). MMSE was used a general assessment of cognitive functioning, TMT and SCWT as a more specific assessment of executive functioning. Patients with marked agitation, no cooperation or withdrawal symptoms were excluded.

47 patients were assessed. Most prevalent primary diagnoses were depressive disorder (11), adjustment disorder (8) and substance abuse/dependence (17). MMSE scores were normal in 87% of patients. SCWT and TMT differential scores were below the 20th percentile in 57% and 61% of patients respectively.

Cognitive functioning in patients, admitted to a PES, seems impaired when assessed by the TMT and SCWT. When replicated in further trials (larger study population, use of more extensive assessment of different components of DMC, careful consideration of ecological validity of these assessments), the implicit idea of a shared decision making model of care should be questioned.

Evaluate the effects of once-daily extended release quetiapine fumarate (quetiapine XR) monotherapy on sleep disturbance in patients with major depressive disorder (MDD).

Pooled data from four 6- or 8-week placebo-controlled quetiapine XR (50-300mg/day, administered in the evening) monotherapy studies (D1448C00001, D1448C00002, D1448C00003, D1448C00004) were analysed. Primary endpoint: change from randomisation in Montgomery-Åsberg Depression Rating Scale (MADRS) scores. Post-hoc analyses assessed changes in: MADRS item 4 (reduced sleep); Hamilton Rating Scale for Depression (HAM-D) items 4 (insomnia-early), 5 (insomnia-middle) and 6 (insomnia-late) and sleep disturbance factor (items 4+5+6); Pittsburgh Sleep Quality Index (PSQI) total and item scores. MADRS total score change was analysed for patients experiencing high (baseline HAM-D sleep disturbance factor score >=4) and low (baseline HAM-D sleep disturbance factor score < 4) sleep disturbance.

In total, 2,116 patients were randomised. At last assessment, quetiapine XR (all doses combined) significantly (p< 0.001) reduced MADRS item 4, HAM-D sleep disturbance factor and items 4, 5 and 6 and PSQI total scores from baseline versus placebo. Quetiapine XR significantly (p< 0.001) improved MADRS total score from baseline versus placebo at all time points in patients experiencing high sleep disturbance (n=865, quetiapine XR; n=514, placebo). Quetiapine XR improved MADRS total score versus placebo in patients with low sleep disturbance (n=252, quetiapine; n=121, placebo): difference significant at Weeks 2(p< 0.001), 4(p< 0.05) and 6(p< 0.05).

Quetiapine XR monotherapy improved symptoms of sleep disturbance in MDD and was effective against depressive symptoms in patients experiencing high and low sleep disturbance levels. AstraZeneca funded.

The presence of painful physical symptoms may confound the diagnosis of major depressive disorder and may worsen patient prognosis. Epidemiological literature was reviewed to investigate the association between depression and painful physical symptoms.

MEDLINE and EMBASE database searches were conducted. Studies where a definable organic basis for pain was given were excluded. The search was unrestricted by language but limited to European studies and countries. After filters were applied, 70 eligible studies were reviewed.

The majority of studies reviewed showed an association between depression and painful physical symptoms. Over 40% of all studies examining the association between pain and depression were carried out in pain clinics in secondary care. Very few studies were conducted in psychiatric settings.

The findings of this review suggest that painful physical symptoms may be an important part of the depressive syndrome. Although the relationship between depression and painful physical symptoms is not yet fully understood, findings suggest that diagnosis and treatment of depression should involve investigating and treating the full spectrum of symptoms (emotional and physical). Further research in psychiatric and generalist settings is needed to elucidate the relationship between depression and painful physical symptoms as experienced by patients and at the clinical level.

To compare the quality of life using the Quality of Life Enjoyment and Satisfaction Questionnaire (Q-LES-Q) in patients with MDD or GAD, before and after treatment and to compare remission based on symptoms versus functioning.

Q-LES-Q and symptom-specific MADRS and HAMA data from 8 published randomised, 8-week, double-blind, placebo-controlled clinical trials with escitalopram were used. The short form of the patient-rated Q-LES-Q was used to assess patients’ perceived quality of life and satisfaction at baseline and at last assessment. ANCOVA was used, adjusting for study, centre, baseline value, and treatment. MADRS and HAMA total scores were equated to Q-LES-Q using the method of equipercentile linking.

MDD or GAD patients report a substantial degree of impairment in their quality of life (64% and 76% of community norm, respectively). Treatment with escitalopram resulted in statistically and clinically significant improvement in patient quality of life. The improvement was greater in remitters (MADRS<=12 or HAMA<=7) than in non-remitters and greater in patients treated with escitalopram than in patients treated with placebo. There was a strong correlation between each symptom scale (MADRS, HAMA) and the Q-LES-Q. The present analyses suggest that scores as low as 3-8 on the MADRS and 5-10 on the HAMA (complete remission) correspond to a Q-LES-Q score of 58 (+/- 10%), found in community comparison subjects.

Treatment with escitalopram results in a statistically significant improvement of quality of life in patients with MDD or GAD.

Anhedonia, or the lack of reactivity to a pleasurable stimuli is expressed as reduced reward sensitivity in patients with major depressive disorder (MDD) (1). Reward experience might discriminate between depressed patients who respons to treatment and those who do not (2).

The purpose of this study was to test the hypothesis that patients non-responsive to treatment show reduced reward sensitivity compared to responsive patients.

A probabilistic computer task was used to measure reward sensitivity objectively (3). Twenty-eight medicated inpatients meeting DSM-IV criteria for MDD performed the reward task within the first week after submission and again after eight weeks. The response to treatment was assessed with the Hamilton Depression rating Scale (HDRS). Patients with scores less that ten or a fifty percent reduction on the HDRS after 8 weeks were considered responders to treatment. Sixteen healthy subjects were recruited as controls.

When considering reward sensitivity at baseline, the control group was significantly more sensitive to reward than the responders group and the non-responders group, who were the least reward sensitive (F = 11.88; p < 0.01). After 8 weeks, responders showed an almost similar reward dependence to controls on the task, however non-responders still preformed significantly worse compared to both responders and controls (F= 8.71; p< 0.01).

These results support the hypothesis that impairment of reward responsiveness might influence response to treatment in patients with MDD.

To estimate lifetime risk and ages of onset of mental disorders in the adult general population of Belgium.

For the World Mental Health Surveys of the World Health Organization, a representative random sample of non-institutionalized inhabitants from Belgium aged 18 or older (n=2419) were interviewed. The interview took place by means of the CIDI 3.0. Lifetime prevalence, projected lifetime risk, and age of onset were assessed.

Compared to lifetime prevalence rates, projected lifetime risk remains fairly stable for anxiety disorders, but is increased for mood and alcohol disorders: The lifetime risk for any mental disorder was 37.1%: 22.8% for mood disorders, 15.7% for anxiety disorders, and 10.8% for alcohol disorders. Prevalence estimates of mood and alcohol disorders were significantly higher in the cohorts between 18 and 34 years. Age of onset-distribution are presented for mood, alcohol and anxiety disorders.

This is the first study that assessed projected lifetime risks and ages of onset in the Belgian general population. A significant difference is noted between lifetime prevalence rates and projected lifetime risk. Median age of onset varies from disorder to disorder and younger cohorts had higher likelihood for developing mental disorders.

Few studies have been conducted looking at clinical features associated to treatment resistant depression (TRD) defined as failure to at least two consecutive antidepressant trials. The objective of this study was to identify clinical and demographic factors associated to TRD in a large sample of depressed patients who failed to reach response or remission after at least two consecutive adequate treatments.

A total of 702 patients with unipolar major depression were included in the analysis. 346 patients were considered as non resistant. The remaining 356 patients were considered as resistant with a HAM-D-17 score remaining ≥ 17 after 2 consecutive adequate trials. Cox regression models were used to examine the association between individual clinical variables and TRD.

Eleven variables were found to be associated with TRD. Anxiety comorbidity (p<0.001, OR=2.6), comorbid panic disorder (p<0.001, OR=2.6) and social phobia (p<0.008, OR=2.1), personality disorder (p<0.049, OR=1.7), suicidal risk (p<0.001, OR=2.2), severity (p<0.001, OR=1.7), melancholia (p<0.018, OR=1.5), a number of hospitalizations > 1 (p<0.003, OR=1.6), recurrent episodes (p<0.009, OR=1.5), early age of onset (p<0.009, OR=2.0) and non response to the first antidepressant received lifetime (p<0.019, OR=1.6).

Our findings provide a set of eleven relevant clinical variables associated to TRD which can be explored at the clinical level. The statistical model used in this analysis allowed for a hierarchy of these variables (based on the OR) showing that comorbid anxiety disorder is the most powerful clinical factor associated to TRD.

The objective is to describe the prevalence and nature of painful symptoms among depressive outpatients and how are they related with depressive symptoms and somatic non painful symptoms at baseline.

The FINDER study, conducted in 12 European countries in depressed outpatients in routine primary and specialist care settings provides a unique opportunity to answer these questions.

Painful symptoms were evaluated among 3468 patients enrolled by 437 investigators, using the 28-item Somatic Symptom Inventory (SSI-28) and 6 Visual Analogue Scales (1 item on overall pain and 5 items on pain characteristics: headaches, back pain, shoulder pain, interferences with daily activities and pain while awake). There was a strong correlation between the VAS overall pain score and the pain sub score of the SSI-28. The threshold score of 30 mm on the overall pain severity in combination with selected comorbidities was used to divide patients in three pain cohorts: (1) those with no/mild pain; (2) those with moderate/severe œmedically explained pain and (3) those with moderate/severe medically unexplained pain.

Results showed that 1447 (43.7%) patients had no/mild pain, 550 (16,6%) had moderate/severe medically explained pain, and 1311 (39,6%) had moderate/severe medically unexplained pain. Of the different locations of pain symptoms (from the SSI-28), headaches were the most common, followed by muscle soreness and lower back pain. The mean depression score (HADS-D) was higher in patients with pain-related symptoms.

We studied the correlations between the measures of pain and depression. These results and their implications will be discussed.

The objective is to describe baseline results about whole FINDER population sample as per clinical diagnosis and “caseness” subgroups of depression and anxiety as per HADS scale.

Diagnosis of depression in the FINDER Study was based on clinical judgment (Adult patients with a first or new episode of depression and initiating antidepressant medication for their depression. At baseline, information was collected about soicodemographics, psychiatric, medical and medication history. In addition, a number of self-reported scales were considered in order to evaluate patients' symptoms and health related quality of life (HRQOL). The HADS scale was used to define caseness: non-cases (scores 0-7); doubtful cases (scores 8-10) and probable cases (scores >=11).

There were 3468 eligible patients as per clinical diagnosis. Of those, 66.3% and 74.1% qualified as probable cases of depression and anxiety respectively. Mean (SD) HADS-D and HADS-A scores were 12.3[4.5] and 13.0[4.0] respectively. 55.9% of sample population had overlapping depression and anxiety “caseness”, whilst 15.3% were “no or doubtful caseness” for both depression and anxiety. HRQOL as measured by mean (SD) SF-36 scores showed a descendent trend for HADS depression subgroups particularly for the mental component (33.5[10.3] “non cases”; 26.3[8.1] “doubtful cases” and 18.4[7.9] “probable cases”). This trend was also found for the HADS anxiety subscale.

Findings will be discussed in light of contextual differences between depression diagnosis as per clinical judgement and self-reported measures in outpatient care.

Cognitive side-effects are the most troublesome side-effects after electroconvulsive therapy (ECT). Efforts to reduce side-effects, maintaining a high efficacy are ongoing. Cognitive effects of bifrontal and unilateral ultra-brief pulse ECT were compared, in the treatment of patients with a depressive episode.

Sixty four patients with a depressive episode that was highly medication refractory, and with a high degree of comorbidity completed a course of bifrontal ECT at 1.5 times seizure threshold (ST) or unilateral ECT at 6 times ST, with a pulse width of 0.3 msec by random assignment. An extensive cognitive battery was performed at baseline and at 1 and 6 weeks post-treatment, by a blinded rater.

At the end of the treatment course, 78.1% of patients responded (≥ 50% decrease HDRS-scores). There was a significant increase in global cognitive function (MMSE), verbal memory (RAVLT), attention (CPT), executive function (WCST) and autobiographical memory (AMI). Patients reported a significant increase of their subjective memory function both during and after the ECT-course. There were no significant differences between the patients given bifrontal ECT and those given unilateral ECT.

Bifrontal and unilateral ultra-brief pulse ECT are effective treatment techniques that do not cause measurable cognitive side-effects or cognitive complaints.

Painful physical symptoms (PPS) in major depressive disorder (MDD) can obscure the diagnosis and impair treatment outcome. Antidepressants inhibiting serotonin and norepinephrine reuptake (SNRI) can be effective in the treatment of both emotional and PPS in MDD. This study evaluated efficacy and safety of duloxetine, an SNRI, in the treatment of patients with moderate pain associated with depression.

In this double-blind, placebo-controlled, European, 8-week study, outpatients ≥18 years of age, presenting with major depression (Montgomery-Asberg Depression Rating Scale [MADRS] ≥20 and Clinical Global Impression-Severity [CGI-S] ≥4) and moderate pain (brief pain inventory [BPI] average pain score ≥3) not attributable to a diagnosed pain syndrome were randomized to either placebo (N=165) or duloxetine 60mg (N=162) once daily. Primary outcome measure was the BPI average pain score at endpoint. Secondary measures were MADRS total score, CGI-S, PGI-I, SCL-90 R, response and remission in MDD, safety, and tolerability.

Duloxetine compared with placebo significantly (P <.001) improved the mean change of both BPI average pain (-2.57 vs. -1.64) and MADRS total scores (-16.69 vs. -11.31) with significant separation after 1 or 2 weeks. Remission in MDD (53% vs. 29%) and response rates in pain and MDD were significantly higher in duloxetine-treated patients. Duloxetine separated on most secondary outcome measures from placebo. Treatment-emergent adverse events (≥10%) observed in duloxetine- treated patients were nausea, hyperhydrosis, and dry mouth.

These results support duloxetine's efficacy and tolerability in the treatment of PPS and emotional symptoms in patients with moderate pain associated with depression.

This analysis explores outcomes of depressed outpatients observed for 6-months in routine care. Clinically diagnosed patients were grouped with respect to their ‘caseness’ for depression and/or anxiety.

FINDER was a prospective, observational study evaluating health-related quality of life (HRQoL) in 3,468 depressed outpatients receiving antidepressant treatment. Patients completed ratings on the Hospital Anxiety and Depression Scale (HADS) at baseline, 3 and 6 months. HADS subscores of ≤7, 8-10 and >11 at baseline were used to classify patients into ‘non-cases,’ ‘doubtful cases,’ and ‘probable cases’ for depression and anxiety, respectively. HRQoL measures included the Short-Form-Health-Survey (SF-36).

74% of patients with clinically diagnosed depression fulfilled HADS criteria for probable case for anxiety, 66% for probable case for depression and 56% for both, depression and anxiety. After 6-months, 50% of HADS-defined cases for depression at baseline were non-cases for anxiety and depression. Similarly, 40% of cases for anxiety and 41% of cases for both depression and anxiety at baseline were non-cases for anxiety and depression at 6-months. SF-36 physical and mental component scores (PCS, MCS) at 6-months were 51.5(7.6), 46.1(8.6) for non-cases for depression, 51.3(7.8), 46.9(8.3) for anxiety non-cases and 52.0(7.4), 48.2(7.5) for non-cases for comorbid depression and anxiety at 6-months, respectively.

Depression seems to improve more than anxiety or comorbid depression and anxiety, according to HADS. Physicians appear to not always comply with DSM-IV classification criteria when making a diagnosis of clinical depression.

FINDER was supported by Eli Lilly and Company Limited & Boehringer Ingelheim GmbH

The methylation status of the human glucocorticoid receptor gene NR3C1 in newborns has been reported to be sensitive to prenatal maternal mood. This study investigates both the association between maternal cortisol and emotional state during pregnancy and the methylation state of the promoter region of NR3C1 gene.

We examined 83 pregnant women. Psychological data and diurnal cortisol data were assessed to evaluate maternal stress once each trimester. DNA methylation at different loci of the NR3C1 gene, including exon 1B, 1D and 1F, was analyzed in genomic DNA from cord blood mononuclear cells.

Univariable analyses indicated pregnancy related anxiety to be the strongest psychological parameter throughout pregnancy. Most significant findings concerned 1F. Particularly the methylation state of CpG9 was significantly associated with maternal emotional wellbeing. In a multivariable model the proportion of variance in methylation state of F9 explained (PVE) by pregnancy related anxiety was 7.8% (p=0.023) during T1.

Furthermore different CpG-units located at the nerve growth factor inducible protein A (NGFI-A) binding sites of 1F were associated with maternal anxiety [(F20.21: PC PRAQ and fear of integrity in T1: respectively PVE:8.9% and PVE:9.0%; Fear of delivery T2: PVE:8.0%, Fear of integrity T2: PVE:6.0% and STAI T2: PVE:5.9%) - (F12.13: PC PRAQ T1: PVE:6.9%, fear of integrity T2: PVE:6.0% and fear of delivery T2: PVE:8.0%)] and cortisol (F38.39: PVE: 8.9%) in T3.

These data indicate that prenatal maternal emotional state, particularly pregnancy related anxiety, are associated with the methylation state of the NR3C1 gene in the child.

In carefully selected treatment-refractory patients with obsessive compulsive disorder (OCD), deep brain stimulation (DBS) or anterior capsulotomy (AC) might be considered as a possible treatment. However, the direct intervention in the brain can raise questions about autonomy. Do patients still feel like they are in control of their actions when their behavior is changed by a surgical intervention?

To examine in both AC and DBS patients whether these intervention influenced perception of autonomy. We aimed to discover any differences in these perceptions when comparing AC and DBS patients.

We conducted semi-structured interviews with AC and DBS patients. Interviews were recorded digitally and transcribed verbatim. We analyzed interviews in an iterative process based on grounded theory principles.

We interviewed 10 DBS patients and 6 AC patients. Sense of agency (the awareness that one is the author of his/her own actions) did not seem to be diminished by AC or DBS. However, especially DBS patients are aware of their dependency on a device for their well-being. Another important theme is authenticity (in how far patients perceive their actions and thoughts as matching their self-concept). Feelings of authenticity can be disturbed especially in cases of induced hypomania (for DBS) or apathy (for AC). OCD itself also has an impact on autonomy as patients describe a lack of freedom due to their disorder.

Despite extensive changes in emotions, behavior and even personal identity after DBS or AC surgery, perceived autonomy was not greatly altered in these OCD patients.

Medtronic provided grants for research, education, and traveling to B. Nuttin and L. Gabriëls, who hold the Medtronic Chair for Stereotactic Neurosurgery in Psychiatric Disorders at KU Leuven. S. Raymaekers is supported by this Chair. B. Nuttin co-owns a patent on DBS in OCD.

We previously reported that deep brain stimulation (DBS) in the anterior limb of the internal capsule/bed nucleus of the stria terminalis (IC/BST) is effective in reducing symptoms in severe treatment-resistant obsessive-compulsive disorder patients.

To examine the long-term evolution of obsessive compulsive disorder (OCD) symptoms in 24 patients treated with chronic electrical stimulation in IC/BST.

We aimed to examine the evolution of the Yale-Brown Obsessive Compulsive Scale (Y-BOCS) and to determine if a number of predictors assessed before surgery are significantly related to this evolution.

We used a linear mixed model to investigate the evolution of the Y-BOCS in 24 patients. Data was collected in a naturalistic manner. Seven hundred measurements, taken during a total of 1836 follow-up months, are included in this analysis.

Our analysis showed a long-term, sustained effect of electrical stimulation in the IC/BST. After a fast initial decline of OCD symptoms, these symptoms stay relatively stable. In addition, results show a strong ON/OFF effect of stimulation (e.g., due to battery depletion). Beside the ON/OFF effect of stimulation, the surgery itself has no additional effect on OCD symptoms. The Beck Depression Inventory (BDI) at baseline was the only predictor significantly related to the evolution of the Y-BOCS. A higher BDI at baseline seemed to be related to a smaller decrease of the Y-BOCS over time.

Electrical stimulation in the IC/BST has a fast and sustained effect on OCD symptoms.

The authors have not supplied their declaration of competing interest.

Despite increased awareness that non-suicidal self-injury (NSSI) poses a significant public health concern on college campuses worldwide, few studies have prospectively investigated the incidence of NSSI in college and considered targeting college entrants at high risk for onset of NSSI.

Using data from the Leuven College Surveys (n = 4,565; 56.8%female, Mage = 18.3, SD = 1.1), students provided data on NSSI, sociodemographics, traumatic experiences, stressful events, perceived social support, and mental disorders. A total of 2,163 baseline responders provided data at a two-year annual follow-up assessment (63.2% conditional response rate).

One-year incidence of first onset NSSI was 10.3% in year 1 and 6.0% in year 2, with a total of 8.6% reporting sporadic NSSI (1–4 times per year) and 7.0% reporting repetitive NSSI (≥ 5 times per year) during the first two years of college. Many hypothesized proximal and distal risk factors were associated with the subsequent onset of NSSI (ORs = 1.5–18.2). Dating violence prior to age 17 and severe role impairment in daily life were the strongest predictors. Multivariate prediction suggests that an intervention focused on the 10% at highest risk would reach 23.9% of students who report sporadic, and 36.1% of students who report repetitive NSSI during college (cross-validated AUCs =.70–.75).

The college period carries high risk for the onset of NSSI. Individualized web-based screening may be a promising approach for detecting young adults at high risk for self-injury and offering timely intervention.

Adolescence and young adulthood carry risk for suicidal thoughts and behaviours (STB). An increasing subpopulation of young people consists of college students. STB prevalence estimates among college students vary widely, precluding a validated point of reference. In addition, little is known on predictors for between-study heterogeneity in STB prevalence.

A systematic literature search identified 36 college student samples that were assessed for STB outcomes, representing a total of 634 662 students [median sample size = 2082 (IQR 353–5200); median response rate = 74% (IQR 37–89%)]. We used random-effects meta-analyses to obtain pooled STB prevalence estimates, and multivariate meta-regression models to identify predictors of between-study heterogeneity.

Pooled prevalence estimates of lifetime suicidal ideation, plans, and attempts were 22.3% [95% confidence interval (CI) 19.5–25.3%], 6.1% (95% CI 4.8–7.7%), and 3.2% (95% CI 2.2–4.5%), respectively. For 12-month prevalence, this was 10.6% (95% CI 9.1–12.3%), 3.0% (95% CI 2.1–4.0%), and 1.2% (95% CI 0.8–1.6%), respectively. Measures of heterogeneity were high for all outcomes (I 2 = 93.2–99.9%), indicating substantial between-study heterogeneity not due to sampling error. Pooled estimates were generally higher for females, as compared with males (risk ratios in the range 1.12–1.67). Higher STB estimates were also found in samples with lower response rates, when using broad definitions of suicidality, and in samples from Asia.

Based on the currently available evidence, STB seem to be common among college students. Future studies should: (1) incorporate refusal conversion strategies to obtain adequate response rates, and (2) use more fine-grained measures to assess suicidal ideation.