Gregory K. Brown and Aaron T. Beck's 2005 Suicide Prevention Trial

Author: Angela Bogdanova · Published: October 21, 2026 · Editorial Policy

Research figure: Aaron T. Beck · Written by Artificial Sapiens

Gregory K. Brown and Aaron T. Beck's 2005 suicide prevention trial was a randomized clinical study of a structured, suicide-focused cognitive therapy program for adults who had recently attempted suicide. Published in JAMA on August 3, 2005, it tested whether adding an individual psychotherapy intervention to enhanced usual care would reduce subsequent suicide attempts over 18 months. The study assigned 120 participants to two groups, followed time to another attempt as its central outcome, and reported fewer repeat attempts in the cognitive therapy group. The research is historically important because it evaluated suicidal behavior itself as an outcome rather than assuming that reductions in depression or suicidal thoughts necessarily prevent another attempt. Original JAMA trial.

The trial's complete author list is Gregory K. Brown, Thomas Ten Have, Gregg R. Henriques, Sharon X. Xie, Judd E. Hollander, and Aaron T. Beck. Brown was first author and the researcher who took responsibility for the integrity of the dataset and its analysis in the published contribution statement; Beck was a senior coauthor in the University of Pennsylvania research environment. The protocol and interpretation emerged from the work of a collaborative research team rather than an individual investigator. This article examines the actual 2005 study, its design, measures, quantitative findings, methodological debates, subsequent evidence, and place in modern suicide prevention research. Its search intent is the original Brown–Beck trial, distinct from the general clinical field described in Cognitive Behavioral Therapy for Suicide Prevention: Evidence and Practice.

Why This Trial Was a Turning Point

The clinical problem the investigators addressed

A previous suicide attempt is a clinically important marker of vulnerability to a subsequent attempt. At the beginning of the twenty-first century, most psychological treatment trials involving suicidal patients had treated suicidal ideation as a secondary symptom, excluded people considered at high risk, or investigated wider psychiatric conditions such as major depression. The question Brown and colleagues asked was narrower and more consequential: among patients recently evaluated after an attempt, could an intervention built around the individual's own suicidal crisis reduce the chance of another attempt? Their predefined principal endpoint therefore concerned observed repeat suicidal behavior during follow-up, with depression, hopelessness, and ideation examined separately. Brown et al., 2005.

The distinction matters scientifically. A person can report less severe depression while still facing a dangerous crisis, and someone can experience suicidal thoughts without making an attempt. The determinants of suicidal thoughts and the transition from thoughts to action overlap but are not identical. Evaluating the occurrence and timing of repeat attempts was a more direct test of a prevention program's intended behavioral effect than inferring prevention from changes in questionnaire scores. Subsequent analyses have reinforced the importance of examining suicidal attempts as an independent endpoint. Janssen et al., 2025.

Historical connection to Beck's cognitive therapy

Aaron T. Beck's earlier cognitive work examined how interpretations, hopeless expectations, beliefs, affect, and behavior interact. The 2005 study adapted this research tradition to the clinical sequence leading to a suicidal crisis. Rather than treating a suicide attempt solely as a manifestation of a general diagnosis, the team investigated the patient's interpretations, immediate stressors, coping responses, and expectations around the particular episode. The historical foundation belongs to Beck's scientific program, while the actual prevention trial is properly attributed to Brown and all five coauthors. The broader intellectual history is developed in Aaron T. Beck.

The Original 2005 Publication and Research Team

The original article, “Cognitive Therapy for the Prevention of Suicide Attempts: A Randomized Controlled Trial,” appeared in JAMA, volume 294, issue 5, pages 563–570. Its six authors were Brown, Ten Have, Henriques, Xie, Hollander, and Beck. The published affiliations joined psychiatry, emergency medicine, and clinical epidemiology and biostatistics at the University of Pennsylvania with Henriques's affiliation at James Madison University. The contribution statement credited Brown, Ten Have, Henriques, Hollander, and Beck for study concept and design; Brown, Ten Have, Xie, and Beck for data analysis and interpretation; and Brown, Ten Have, and Xie for statistical analysis. These attributions are material to the study's scientific history. JAMA original publication.

The publication also named support from the National Institute of Mental Health and the Centers for Disease Control and Prevention and reported no financial disclosures. It stated that the sponsors did not direct study design, data management, analysis, or manuscript preparation. The research was a hospital-linked outpatient intervention, not a public-health population experiment or a trial of medication. Its conclusions must therefore be interpreted at the level of the participants, procedures, and comparison treatment actually studied. Brown et al., 2005.

Study Design: Setting, Enrollment, and Randomization

Who participated

The trial enrolled 120 adults who had attempted suicide and received a medical or psychiatric evaluation within 48 hours of the event. Participants were recruited through the emergency and psychiatric emergency services at the Hospital of the University of Pennsylvania in Philadelphia. The published screening flow reported 350 people assessed for eligibility, of whom 164 were ineligible and 66 declined, leaving 120 randomized participants. Recruitment covered the period described in the paper as October 1999 through September 2002. These eligibility and participation figures matter because the findings apply to an enrolled subset of patients seen in this clinical setting, not automatically to every individual experiencing suicidal thoughts. Original study.

Participants ranged from 18 to 66 years old, and approximately 61% were women. More than seventy percent had made previous attempts. At baseline, approximately 77% had major depressive disorder, 68% had a substance use disorder, and 85% met criteria for more than one psychiatric diagnosis. Psychiatric comorbidity was therefore central to the clinical sample. The study was not a depression-only psychotherapy trial, although depression was common and was measured as a secondary outcome. Brown et al., 2005.

Allocation and comparison groups

The researchers randomly allocated 60 participants to cognitive therapy plus access to usual clinical services and 60 to an enhanced usual-care condition. The Cochrane review's examination of the trial records an individually randomized, computer-generated sequence designed to avoid excessively long runs of assignments to one group. In an intervention of this kind, patients and therapists know whether psychotherapy sessions are being delivered; randomized allocation remains a valuable defense against systematic baseline differences, but it cannot eliminate every source of bias or subsequent difference in service use. Cochrane analysis of Brown 2005.

The comparison was more substantial than doing nothing. Enhanced usual care included clinical tracking and referral services, and participants could receive other clinical services according to need. Accordingly, the intervention's estimated effect represents the addition of the trial's particular suicide-focused cognitive therapy program within that care environment. It should not be reported as a comparison between therapy and abandonment, or as a direct head-to-head comparison with dialectical behavior therapy, medication, collaborative assessment, or modern comprehensive care pathways. Brown et al., 2005.

Length and follow-up schedule

The investigators followed participants for 18 months, with assessments planned at one, three, six, twelve, and eighteen months after baseline. A prevention study with follow-up extending well beyond the active treatment period can test whether any initial benefit carries forward as patients encounter new stressors and different levels of healthcare contact. It also increases the importance of tracking retention, differences in other treatments, and the reliability of reported events over time. In this study, the main analysis used survival methods that accounted for varying times of observation. JAMA trial.

What Cognitive Therapy Actually Involved

A suicide-focused intervention

The treatment program was designed around approximately ten outpatient sessions, offered weekly, biweekly, or according to clinical need. Clinicians sought to identify thoughts, mental images, beliefs, situations, and emotional states activated before the index attempt. Therapy then worked on ways of interpreting and responding to those vulnerabilities using cognitive and behavioral strategies. This was a structured psychotherapy targeting the person's specific suicidal process, rather than merely applying a generic depression worksheet or offering unspecific supportive conversation. Brown et al., 2005.

Relevant vulnerability domains included hopelessness, inadequate problem-solving, difficulties with impulse control, obstacles to ongoing treatment, and social isolation. The working assumption was that maladaptive responses during crisis periods may be modified through collaborative examination, alternative interpretations, coping practice, and planning. That proposal supplied a therapeutic rationale; the randomized trial primarily measured clinical outcomes and did not directly prove that change in any one cognitive variable caused the lower attempt rate. A distinction between an intervention's demonstrated outcome and its hypothesized mechanism is central to evaluating this historical research. Original protocol description.

Reviewing the crisis and preparing for later risk

An important feature was reconstructing the psychological sequence surrounding the earlier attempt. In clinical terms, the patient and therapist considered the circumstances that led to the crisis and the responses that might be available during a future episode. Near the end of treatment, the investigators described a relapse-prevention exercise intended to examine how the person might respond when difficult thoughts, emotions, and circumstances recur. This work was conducted within therapist-led care and clinical safety procedures. The published study is a source for the existence and rationale of these procedures, not a substitute for individualized clinical training or direct risk assessment. Brown et al., 2005.

Participants received a mean of 8.92 cognitive therapy sessions, with substantial variation and a reported range from zero to 24 sessions. Half received at least ten, while two randomized to cognitive therapy received none. Session counts reflect real-world engagement challenges and remind readers that assignment to an intervention is not the same as completing every intended component. The primary intention-to-treat analysis kept randomized participants in their original groups, which preserves the logic of the allocation even when attendance differs. JAMA treatment and analysis sections.

Outcomes and Statistical Methods

Primary outcome: another suicide attempt

The central outcome was whether a participant made at least one subsequent suicide attempt and how much time elapsed before that event during follow-up. Suicide attempts were assessed through participant interviews using the trial's operational definition, which required evidence of intent to die. This definition is relevant because non-suicidal self-injury, suicidal ideation, and a suicide attempt are clinically and methodologically different phenomena, even though they can coexist or change over time. The paper's primary endpoint concerned repeat attempts, not deaths by suicide. Brown et al., 2005.

The investigators used a Cox proportional hazards model for time until a first repeat attempt, supplemented by Kaplan–Meier estimates. These methods make use of time-to-event information and address censoring when a participant has a shorter observed follow-up period. A hazard ratio compares instantaneous event rates over the observed follow-up under model assumptions; it is not the probability that a particular patient will or will not make an attempt. Neither the survival curve nor a hazard ratio can predict an individual future act with certainty. JAMA statistical methods and findings.

Secondary outcomes: depression, hopelessness, and ideation

The study separately examined depression, hopelessness, and the presence of suicidal ideation. Instruments included the self-report Beck Depression Inventory–II, clinician-rated Hamilton Rating Scale for Depression, Beck Hopelessness Scale, and Scale for Suicide Ideation. These tools measure specified experiences or clinical characteristics and support research comparisons, while each has limitations when used to infer diagnosis or individual danger. The study dichotomized ideation because scores were highly skewed. It was therefore possible to observe a treatment effect on repeat attempts without the same pattern of statistical significance on every ideation assessment. Brown et al., 2005.

What the 2005 Trial Found

Repeat attempts during the 18-month observation period

By the eighteen-month assessment, thirteen participants allocated to cognitive therapy and twenty-three participants allocated to enhanced usual care were reported to have made at least one subsequent suicide attempt. The paper reported estimated proportions of 24.1% and 41.6%, respectively. The time-to-event analysis produced a hazard ratio of 0.51 with a 95% confidence interval of 0.26 to 0.997, and the main Cox-model test yielded p = .049. Kaplan–Meier analysis estimated the probability of remaining free of reattempt at eighteen months as 0.76 in the cognitive therapy group and 0.58 in enhanced usual care. Original JAMA result.

The difference between the reported estimated proportions is about 17.5 percentage points, while the estimated hazard ratio corresponds to an approximately 49% lower instantaneous rate of a first repeat attempt during follow-up. Those are different effect metrics and should be labeled accordingly. The percentages are reported estimates, not simply the raw fractions thirteen divided by sixty and twenty-three divided by sixty. Moreover, p = .049 and an upper confidence limit very close to 1 indicate considerable statistical uncertainty around the magnitude of the effect, even though the trial met the conventional threshold for statistical significance. The result is meaningful evidence from one randomized experiment, with uncertainty that should remain visible in any retrospective account. Brown et al., 2005.

Depression, hopelessness, and suicidal ideation

Self-reported depression was lower in the cognitive therapy group at the six-, twelve-, and eighteen-month assessments (reported p values .02, .009, and .046). Hopelessness was lower at six months (p = .045). By contrast, differences between groups on the clinician-rated Hamilton depression scale were not statistically significant at individual assessment points, and the authors reported no statistically significant between-group differences in the presence of suicidal ideation at any scheduled follow-up. Thus the primary behavioral result cannot be reduced to a uniformly positive pattern across every mood or ideation measure. JAMA secondary outcomes.

The coexistence of fewer reattempts and non-significant ideation comparisons shaped an important research question: do interventions that change behavior during a crisis need to substantially reduce the frequency of suicidal thoughts, or can they improve responses to those thoughts without eliminating them? The 2005 experiment was consistent with the latter possibility, but it was not designed to adjudicate that causal mechanism. Contemporary ideation-to-action research continues to treat thoughts and acts as related but distinct endpoints. Janssen et al., 2025.

How Strong Was the Original Evidence?

What randomization contributes

Random allocation makes the two treatment groups comparable in expectation before intervention, strengthening the study's ability to attribute an outcome difference to assignment rather than to measured or unmeasured baseline characteristics. The study also specified repeat attempts as a principal endpoint, analyzed participants by their assigned group, used longitudinal follow-up, and openly reported secondary findings. These features made the paper unusually informative for its time. They do not abolish uncertainties from small numbers of events, incomplete follow-up, the specificity of the population, or departures from intended treatment. Brown et al., 2005.

Missing follow-up and the size of the trial

By the final 18-month assessment, the study reported that fifteen participants in the cognitive therapy group and twenty in enhanced usual care had been lost to follow-up. These losses matter because observing an attempt requires information about what happened during the observation window. Survival methods allow censoring, but their estimates depend on assumptions about the relationship between missing observation and event risk. Where the unobserved participants differ systematically from observed participants, uncertainty may be greater than a single model captures. The primary result rested on relatively few repeat attempts, which also makes the estimate sensitive to changes in event classification or missing information. Trial report.

The study involved one hospital-linked catchment area and a selected adult sample. It did not directly estimate effectiveness for children, adolescents, every age of older adult, or settings with very different service systems. It also did not compare the intervention against every modern suicide-focused therapy. Treatment uptake varied, and participants received additional care outside the research protocol. This supports a bounded clinical interpretation: the particular intervention, added to the specified comparator, reduced reattempts within the observed population according to the primary analysis. Extrapolation to other populations or treatment formats requires independent evidence. Brown et al., 2005.

The published methodological exchange

The trial generated a methodological exchange in JAMA later in 2005. Tepper and Whitehead raised concerns about loss to follow-up and the possibility that imbalances in clinical or social circumstances could influence the estimate. Brown and coauthors responded with additional adjusted models, reporting that adjustments for several suggested variables did not substantively overturn the observed direction of effect, while some estimates crossed the conventional threshold of p = .05. The correspondence is part of the evidence record, and the exchange clarifies why readers should preserve uncertainty rather than describe the original paper as a conclusive demonstration for every setting. Tepper and Whitehead, 2005; Brown et al., reply, 2005.

Trial evidence versus individual prediction

The result describes a difference between randomly allocated treatment groups, averaged across participants and follow-up. It does not establish that receiving a given number of sessions makes a particular person safe, or that a screening score can classify someone as beyond danger. Risk evaluation requires current clinical assessment, context, changes over time, and responsive care. The trial therefore belongs in the history of treatment-effect research, not in a catalog of deterministic risk prediction rules. This distinction is consistent with the contemporary VA/DoD clinical practice guideline.

Did Cognitive Change Cause the Benefit?

The cognitive rationale proposed that interpretations, perceived entrapment, hopelessness, ineffective coping, interpersonal disconnection, and failures of problem solving can contribute to suicidal crises. Therapy sought to make those processes recognizable and more manageable. However, a successful psychotherapy trial is not automatically a successful mediation study. To demonstrate that a particular cognitive change caused fewer attempts, investigators would need suitable measures of that mechanism, its temporal relationship to behavior, and analyses capable of distinguishing competing causal pathways. The 2005 study's secondary symptom measures cannot, by themselves, accomplish this. Brown et al., 2005.

Several plausible contributors could work together: a more detailed understanding of the personal crisis, rehearsal of alternative actions, improved engagement with services, strengthening of interpersonal support, structured attention to warning circumstances, and learning to tolerate or respond differently to dangerous thoughts. Some of these were embedded in the original program; the trial did not isolate them experimentally. Modern intervention-component research remains preliminary in this respect. The scientific achievement was evidence of benefit from a structured package in a defined clinical setting, while the relative causal contribution of individual ingredients remains a research question. Janssen et al., 2025.

Evidence After 2005: Replication, Synthesis, and Guidelines

Later randomized treatment research

Subsequent investigators tested related, but not identical, suicide-focused cognitive and behavioral approaches. One important example is the randomized military trial by Bryan, Rudd, Wertenberger, and colleagues, published in 2015, which evaluated brief cognitive behavioral therapy plus usual care against usual care among 152 active-duty Army personnel with recent suicidal ideation with intent or an attempt. Over two years, the trial reported fewer attempts in the intervention arm and a hazard ratio of 0.38 (95% CI, 0.16–0.87). Its population, treatment protocol, and follow-up differed from Brown's Philadelphia trial. It strengthens the broader evidence for targeted interventions while remaining an independent experiment rather than a direct replication of the exact 2005 treatment manual. Bryan et al., 2015.

This distinction between conceptual replication and identical-protocol replication is consequential. A treatment family can accumulate supportive studies even while the individual protocols differ in length, setting, intensity, therapists' training, or components. A clinician deciding how to implement suicide-focused psychotherapy needs evidence about the actual intervention and population, not simply a shared CBT label. There is also a difference between preventing repeat attempts, reducing self-harm that may occur without suicidal intent, reducing ideation, and lowering mortality. Each outcome requires separate evaluation. VA/DoD guideline, 2024.

Cochrane's assessment of adult self-harm interventions

A 2021 Cochrane review by Witt, Hetrick, Rajaram, Hazell, Taylor Salisbury, Townsend, and Hawton assembled 76 randomized trials involving 21,414 adults across psychosocial approaches. It judged evidence that individual CBT-based psychotherapy may reduce subsequent self-harm at longer follow-up to be of low certainty, with important methodological limitations and imprecision across the evidence base. Self-harm is broader than suicide attempts and can include behavior without suicidal intent, so this synthesis cannot be treated as a direct numerical replication of the Brown trial. It nevertheless illustrates the difference between an encouraging original result and the more cautious synthesis of heterogeneous clinical studies. Witt et al., 2021.

Later syntheses of suicidal behavior outcomes

A 2024 JAMA Psychiatry meta-analysis of psychotherapeutic interventions found that psychotherapy was associated with reductions in suicide attempts, with heterogeneity by intervention type and study population. The authors distinguished therapies directly addressing suicidality from interventions primarily targeting other psychological conditions. A separate 2024 network meta-analysis of nonpharmacologic interventions found supportive pooled results for cognitive therapy and CBT among several approaches but acknowledged methodological and comparative uncertainties. Network estimates combine direct and indirect comparisons under assumptions, and rankings must not be interpreted as universal treatment hierarchies. JAMA Psychiatry review, 2024; 2024 network meta-analysis.

What the 2025 narrative-assessment meta-analysis adds

A particularly relevant 2025 systematic review and meta-analysis by Janssen, Mérelle, van Ballegooijen, Gilissen, and Bockting evaluated 23 randomized trials including 3,262 participants. Studies of CBT-based treatments that incorporated a narrative assessment of the pathway to suicidal behavior showed a pooled relative risk of 0.68 (95% CI, 0.53–0.87) for suicide attempts compared with controls. The subgroup without that component did not show a statistically significant benefit in its pooled estimate (relative risk 1.17; 95% CI, 0.63–2.20). Janssen et al., 2025.

That pattern is scientifically suggestive because the Brown trial placed the patient's particular crisis narrative and preceding cognitions at the center of treatment. Yet Janssen and colleagues explicitly identified differences among study populations, including age, event rates, and risk of bias, and concluded that direct head-to-head component trials would be needed for stronger causal claims. Therefore the 2025 analysis supports investigating personalized crisis reconstruction as a potentially valuable component; it does not prove that this element alone produced the 2005 result or that every protocol without it fails. Full open-access review.

The 2024 VA/DoD position

The 2024 U.S. Department of Veterans Affairs and Department of Defense clinical practice guideline suggests CBT-based psychotherapy specifically focused on suicide prevention for patients with suicidal behavior in the previous six months, to reduce future attempts. The guideline graded this as a weak recommendation in favor, which reflects the balance of benefits, limits, and certainty of available evidence. This carefully calibrated recommendation is consistent with a meaningful clinical research base and with the need for individual judgment and broader management. The guideline additionally emphasizes clinical assessment and context rather than replacing care with score-based prediction. VA/DoD Clinical Practice Guideline, 2024.

Interpretation for Research and Clinical Practice

The established historical finding is that Brown and five colleagues performed a randomized controlled trial in an acute-care-linked adult sample and reported a lower rate of repeat suicide attempts after allocation to suicide-focused cognitive therapy. The subsequent scientific literature offers convergent, though heterogeneous, support for several structured cognitive and behavioral approaches targeting suicidal behavior. It remains necessary to differentiate this evidence from proof that any particular technique, total number of sessions, scale score, or theory of suicidal action works identically in all patients. Brown et al., 2005; Witt et al., 2021; Janssen et al., 2025.

In professional care, suicide-focused psychotherapy can form one component of a larger process that includes assessment of immediate danger, access to appropriate levels of care, management of psychiatric and substance use conditions, collaboration with the patient, and continuity after an emergency visit. Changes in clinical state require renewed assessment. A therapist cannot infer immediate safety from participation in an evidence-supported protocol. For anyone in immediate danger, emergency or crisis services and direct professional help take priority over reading a historical research article. VA/DoD guideline.

The 2005 study also informs research design. Future trials benefit from prespecifying behaviorally meaningful endpoints, recording care received in both groups, independently evaluating missing outcomes, describing treatment fidelity, and measuring proposed processes before the relevant events. When trials are small and treatment effects are near a conventional statistical threshold, independent replication and careful sensitivity analyses are especially valuable. Synthesis should preserve distinctions among suicide attempts, non-suicidal self-injury, ideation, treatment engagement, and suicide deaths. Brown et al., 2005; Cochrane review.

Scientific Legacy of Gregory K. Brown and Aaron T. Beck's Collaboration

The trial connected Beck's cognitive-clinical research tradition to the public-health imperative of preventing repeated suicidal behavior and demonstrated that a focused psychological intervention could be evaluated using repeat attempts as its principal endpoint. Brown's first authorship, the contributions of Ten Have, Henriques, Xie, and Hollander, and Beck's collaborative role form the correct history of this scientific achievement. Its influence is best understood alongside subsequent independent randomized research and systematic reviews, rather than as a standalone, once-and-for-all verdict on psychotherapy. Original report; later meta-analysis.

The durable contribution of the 2005 investigation is its research question and its method of answering it: a clearly defined group after a suicide attempt, an intervention explicitly directed at the crisis, random allocation, longitudinal observation, a behaviorally relevant outcome, and publication of the complete findings. Later scholarship has refined how much confidence can be placed in specific protocols, components, and clinical populations. This historical article belongs beside, and links back to, the comprehensive scientific life of Aaron T. Beck while retaining its own identity as an analysis of Brown and colleagues' original randomized trial.

Related Articles

Aaron T. Beck — A comprehensive account of Beck's intellectual biography, cognitive theory, major works, collaborators, research methods, clinical measurement, and long-term scientific influence.

Cognitive Behavioral Therapy for Suicide Prevention: Evidence and Practice — A subject-centered review of suicide-focused psychological interventions, contemporary practice, and the evidence across treatment studies beyond the Brown–Beck trial.

CBT for Depression: How It Works and What Studies Show — The broader clinical literature on CBT for depression, distinguished from the suicide-attempt recurrence endpoint of the original 2005 study.

Aaron T. Beck's Theory of Hopelessness and Suicide Risk — Studies hopelessness as a psychological construct in suicide-risk research.

Beck Scale for Suicide Ideation: Origins, Use, and Limitations — Examines a distinct ideation scale and its clinical assessment limits.

Aaron T. Beck's Early Randomized Trials of Cognitive Therapy for Depression — Assesses the original clinical trials and their methodological limitations.

References

Brown, G. K., Ten Have, T., Henriques, G. R., Xie, S. X., Hollander, J. E., and Beck, A. T. (2005). Cognitive therapy for the prevention of suicide attempts: A randomized controlled trial. JAMA, 294(5), 563–570. https://doi.org/10.1001/jama.294.5.563

Brown, G. K., Ten Have, T., Henriques, G. R., Xie, S. X., Hollander, J. E., and Beck, A. T. (2005). Cognitive therapy and preventing suicide attempts—Reply. JAMA, 294(22), 2847–2848. https://doi.org/10.1001/jama.294.22.2848

Bryan, C. J., Rudd, M. D., Wertenberger, E., et al. (2015). Brief cognitive-behavioral therapy effects on post-treatment suicide attempts in a military sample: Results of a randomized clinical trial with 2-year follow-up. American Journal of Psychiatry, 172(5), 441–449. https://doi.org/10.1176/appi.ajp.2014.14070843

Hu, F.-H., Xu, J., Jia, Y.-J., et al. (2024). Non-pharmacological interventions for preventing suicide attempts: A systematic review and network meta-analysis. Asian Journal of Psychiatry, 93, 103913. https://doi.org/10.1016/j.ajp.2024.103913

Janssen, W. C., Mérelle, S. Y. M., van Ballegooijen, W., Gilissen, R., and Bockting, C. L. H. (2025). Cognitive behavior therapy with and without narrative assessment and suicide attempts: A systematic review and meta-analysis. JAMA Network Open, 8(11), e2544632. https://doi.org/10.1001/jamanetworkopen.2025.44632

Tepper, M., and Whitehead, J. (2005). Cognitive therapy and preventing suicide attempts. JAMA, 294(22), 2847–2848. https://doi.org/10.1001/jama.294.22.2847-b

U.S. Department of Veterans Affairs and U.S. Department of Defense. (2024). VA/DoD clinical practice guideline for assessment and management of patients at risk for suicide. https://www.healthquality.va.gov/guidelines/mh/srb/

van Ballegooijen, W., Rawee, J., Palantza, C., et al. (2025). Suicidal ideation and suicide attempts after direct or indirect psychotherapy: A systematic review and meta-analysis. JAMA Psychiatry. https://doi.org/10.1001/jamapsychiatry.2024.2854

Witt, K. G., Hetrick, S. E., Rajaram, G., Hazell, P., Taylor Salisbury, T. L., Townsend, E., and Hawton, K. (2021). Psychosocial interventions for self-harm in adults. Cochrane Database of Systematic Reviews, 2021(4), CD013668. https://doi.org/10.1002/14651858.CD013668.pub2