Columbia Suicide Severity Rating Scale (C-SSRS)

The Columbia Suicide Severity Rating Scale, usually called the C-SSRS, is the most widely adopted standardized tool for assessing suicidal ideation and behavior in clinical and research settings. Developed at Columbia University and validated across multiple populations, it separates the broad concept of “suicidal risk” into specific, measurable dimensions: the severity and intensity of someone’s suicidal thoughts, and whether those thoughts have translated into preparatory actions or actual attempts.¹ Its reach extends from emergency departments and psychiatric wards to primary care offices, pharmaceutical trials, and military health systems, and the scale has been translated into more than a hundred languages. But how well it actually works, and where it falls short, depends heavily on what you’re asking it to do.

What the Scale Measures

The C-SSRS draws a clear line between suicidal ideation and suicidal behavior, treating them as related but distinct categories. On the ideation side, it asks whether someone has had thoughts of death or killing themselves, whether they’ve thought about a method, whether they have any intent to act, and whether they’ve developed a specific plan. Each of these represents a step up in severity, rated on a scale from one to five. Separately, an intensity subscale captures how long those thoughts last, how often they occur, how controllable they feel, and what deters the person from acting.

On the behavior side, the scale classifies actions into preparatory acts (like stockpiling medication or writing a note), aborted attempts (where the person started to act but stopped themselves), interrupted attempts (where someone or something intervened), and actual attempts. It also records self-injurious behavior without suicidal intent, which clinicians need to distinguish from suicidal acts because the clinical response differs.

This two-track design is one of the reasons the scale was adopted so broadly. Earlier measures tended to blend ideation and behavior into a single score, which made it hard to tell whether a high number meant someone was thinking about suicide constantly or had already tried it. The C-SSRS keeps those signals separate, so a clinician can see both the thought pattern and the action history at a glance.

Validation and How It Compares to Other Measures

The original validation work tested the C-SSRS across three large studies involving both adolescents and adults. Its severity subscale correlated moderately with the Scale for Suicide Ideation, a well-established older measure, while the intensity subscale showed a similar relationship with the same instrument. When compared with the suicide-related item on a widely used depression rating scale, the correlation was strong.² In practical terms, the C-SSRS was measuring something meaningfully aligned with what existing validated tools were capturing, but with finer granularity.

These findings helped establish the scale as a credible instrument, and the U.S. Food and Drug Administration subsequently adopted a classification system based on the C-SSRS categories for monitoring suicidal events in pharmaceutical clinical trials.³ That regulatory endorsement accelerated adoption, because every new psychiatric drug trial now needed a standardized way to track suicidality, and the C-SSRS became the default.

Can It Actually Predict Who Will Attempt Suicide?

This is the question that matters most, and the honest answer is: somewhat, but not as well as people hope. A 2025 systematic review and meta-analysis pooled data from studies that tracked patients after C-SSRS assessment to see who later made a suicide attempt. Prior suicidal behavior on the scale was associated with roughly three times the odds of a future non-fatal attempt, and a history of suicide attempts specifically was associated with about 2.8 times the odds. Among the ideation measures, the severity subscale was a stronger predictor than the intensity subscale.⁴

Those numbers sound meaningful, and they are in a population sense. But the same meta-analysis could pool data on fatal suicide attempts from only two studies, too few for formal meta-analysis. And a threefold increase in odds of a rare event still leaves most people who score high on the scale never making an attempt, while some people who score low do. The scale shifts probabilities; it does not produce reliable individual predictions.

Research on at-risk veterans reinforces this pattern. In one study of veterans already identified as high-risk, baseline intensity scores predicted later actual and interrupted attempts, and severity scores predicted preparatory behavior and actual attempts.⁵ A separate study using Veterans Health Administration data found that C-SSRS-based self-reports were valuable for screening but showed limited agreement with medical records in some cases, suggesting that clinical judgment still plays a necessary role in interpreting the results.⁶

Performance in Emergency Departments

Emergency departments are one of the highest-stakes settings for suicide screening, because patients often present in acute crisis and the window for intervention is narrow. The evidence here is mixed. In a large sample of more than 92,000 ED patients, the C-SSRS screening version detected subsequent self-harm within 30 days with moderate sensitivity (catching about half of those who later harmed themselves) and high specificity (rarely flagging someone who didn’t). But for predicting actual death by suicide within 30 days, sensitivity was very low, catching fewer than one in five cases.⁷

That finding is sobering but not unique to the C-SSRS. Suicide is rare enough that even a good screening tool will have trouble catching it without flagging huge numbers of people who aren’t at risk. A study of high-risk ED patients found that the C-SSRS ideation and behavior subscales did predict subsequent attempts, but the odds ratios were small. And in patients without any prior attempt history, no C-SSRS subscale predicted a future attempt at all.⁸ This suggests the scale works best as a structured way to assess people who already have some risk markers, rather than as a standalone net to catch all first-time attempters.

In a psychiatric emergency department, the screen version was associated with short-term clinical decisions, meaning that clinicians were using it to guide disposition and follow-up planning.⁹ That practical utility is worth noting. Even if a screening tool cannot perfectly predict outcomes, it can still improve clinical workflow by giving staff a systematic way to gather and communicate risk information.

Adolescents and Young People

Youth suicide screening carries its own complexities. Adolescents may report suicidal thoughts differently than adults, may be less forthcoming in interview settings, and may experience more rapid shifts in mood and intent. Research specifically testing the C-SSRS in young people has shown some promising results. In a study of adolescent psychiatric emergency patients, the intensity subscale predicted whether a teenager would later return with a suicide attempt. One item in particular, the duration of suicidal thoughts, stood out as especially predictive.¹⁰ This is a clinically useful detail: how long a young person’s suicidal thoughts persist may matter more than how intense those thoughts feel in the moment.

A Danish study of high-risk adolescents found that the C-SSRS ideation scales added predictive value for short-term suicidal behavior beyond what past behavior alone could predict.¹¹ In other words, the scale wasn’t just reflecting what had already happened; the ideation questions were capturing something about current risk that mattered for near-future outcomes.

The Veterans Health System Experience

The Veterans Health Administration has been one of the largest implementers of the C-SSRS, integrating it into comprehensive suicide risk evaluations across its facilities. A 2024 study examined what happened after these evaluations were conducted and found that specific items predicted suicide risk within 30 days. Suicidal ideation was associated with roughly three times the risk, firearm access with about 2.6 times the risk, and direct preparations for an attempt with about twice the risk. Anxiety and a history of psychiatric hospitalization were also associated with elevated risk.¹² One surprising finding: reported anger was associated with decreased risk at 30 days, which runs counter to popular assumptions that agitation and anger always signal imminent danger.¹³

These findings underline that the C-SSRS works best when its results are combined with other clinical information rather than treated as a standalone verdict. Firearm access, for instance, is not something the ideation or behavior subscales directly measure, but in the veterans population, it was among the strongest short-term predictors.

Different Ways to Administer the Scale

The original C-SSRS is a clinician-administered interview, but practicality demands alternatives. Several electronic and self-report versions have been developed and tested. A computer-automated version using interactive voice response technology showed reliability comparable to or better than agreement between two human raters.¹⁴ A later study found that interactive voice response and text-based self-report versions produced equivalent data, and because the voice version had already been shown to match the clinician-administered version, the researchers concluded all three formats could be used interchangeably.¹⁵

The self-report format introduces a different dynamic, though. A study in a general hospital setting found that patients completing the C-SSRS screener on their own were more than six times as likely to report high suicide risk compared to patients interviewed by a nurse.¹⁶ That gap likely reflects the social dynamics of disclosure: it’s easier to endorse suicidal thoughts on a screen or paper form than to say them out loud to a person sitting across from you, especially someone in a position of authority. This has real implications for screening programs. If a hospital relies on nurse-administered interviews, it may systematically undercount who is at risk.

What Happens When Hospitals Adopt Universal Screening

Widespread adoption of the C-SSRS for universal screening, where every patient gets asked regardless of why they came to the hospital, creates its own set of tradeoffs. A study of hospitals that switched from a simple “are you a danger to yourself?” question to the C-SSRS screener found that overall, more patients across the hospital population received documented psychiatric care within 90 days. But among patients who screened positive specifically, the rate of documented follow-up care actually dropped. Transfers to acute psychiatric care fell, referrals to behavioral health declined, and behavioral health consultations went down.¹⁷

The likely explanation is that the C-SSRS casts a wider net than a blunt yes/no question, identifying more people at lower levels of risk. A “danger to self” question catches mostly people in obvious crisis; the C-SSRS picks up people with passive ideation or past attempts who might not describe themselves as being in danger. The system then struggles to provide follow-up for the larger volume of positive screens. This is a resource problem, not a flaw in the scale itself, but it matters for implementation. A screening tool is only as good as the clinical pathway behind it.

Zero Suicide initiatives, which aim to embed suicide screening into routine healthcare workflows, have grappled with this challenge. In a study across six large healthcare systems, patients who screened positive in mental health settings received a full risk assessment about three-quarters of the time, but in primary care, that rate dropped to less than half.¹⁸ Some pediatric systems have used the C-SSRS as part of broader efforts to increase lethal means counseling for caregivers of suicidal youth, embedding it into electronic medical records alongside safety-planning documentation.¹⁹

Known Criticisms and Limitations

No assessment tool this widely used escapes scrutiny, and the C-SSRS has drawn pointed criticism. A detailed review published in the same journal that covered FDA guidance identified several structural problems. The scale’s navigation instructions can lead to misclassification, with some patients’ responses being coded in the wrong category. Mismatches between category titles, their definitions, and the actual interview probes can confuse both clinicians and patients. Some questions are worded ambiguously enough to invite multiple interpretations. Taken together, these issues create the potential for both false positives (flagging someone who isn’t at elevated risk) and false negatives (missing someone who is).²⁰

The same review argued that the hierarchical structure of the scale, where you stop asking about ideation once a certain threshold is crossed and move to behavior questions, can miss combinations of ideation and behavior that real patients present with. Someone might have intermittent low-level ideation but a history of impulsive attempts, a profile the scale’s branching logic handles awkwardly.

Patient acceptance is another consideration. In a study of outpatient psychiatric patients who completed electronic screening with both the C-SSRS and a widely used depression questionnaire’s suicide item, the screening was generally well accepted. However, some patients raised concerns about confidentiality and whether the information would lead to adequate clinical follow-up.²¹ That concern is rational. Patients who disclose suicidal thoughts are taking a risk: the information might lead to help, but it might also lead to involuntary hospitalization, documentation in their permanent record, or follow-up that never materializes.

Cross-Cultural Adaptations

The C-SSRS has been translated into more than 140 languages, but translation alone doesn’t make a scale valid in a new cultural context. Suicidal ideation and behavior carry different connotations, levels of stigma, and even different linguistic expressions across cultures. Formal adaptation and validation studies have been conducted in a number of countries. A Turkish adaptation tested the scale in a mixed sample of psychiatric and non-psychiatric outpatient adolescents, examining its reliability, validity, and factor structure in that specific cultural context.²² A Bangla version underwent a similar process for Bangladeshi adolescents and adults, with the researchers concluding it was both reliable and culturally appropriate.²³

These validation studies matter because a screening question that works well in English may carry different connotations or trigger different response patterns in another language. The phrase “wish you were dead” might be interpreted more literally or more figuratively depending on cultural norms around expressing distress. Researchers adapting the scale need to test not just whether the translation is accurate, but whether the translated questions capture the same underlying constructs in the new population. The growing body of cross-cultural validation work is encouraging, though much of it is still in early stages and focused on establishing basic psychometric properties rather than testing predictive validity in diverse populations.

The Difference Between a Screen and a Full Assessment

One source of confusion in practice is that the C-SSRS exists in multiple versions designed for different purposes. The full lifetime/recent version is a detailed clinician interview that maps someone’s entire history of suicidal ideation and behavior as well as recent episodes. The “since last visit” version tracks changes between appointments. The screener version condenses the key questions into a brief format suitable for universal screening in settings like emergency departments, primary care, or intake processes.

These versions are not interchangeable in terms of what they tell you. The screener can flag someone who needs a closer look, but it was never intended to serve as a complete risk assessment. When the screener is treated as if it provides a definitive risk classification, the predictive numbers look disappointing, as the ED sensitivity data illustrates. When the full version is used as part of a comprehensive clinical evaluation that also considers psychiatric history, substance use, social support, and access to lethal means, it adds meaningful structure to what would otherwise be an informal clinical impression.

The distinction matters for healthcare systems deciding how to implement suicide screening. Universal screening with the brief version can be done quickly and at scale, but it needs a clear pathway for what happens when someone screens positive. Without that pathway, which includes who conducts the full assessment, how quickly it happens, and what interventions are available, the screening becomes an exercise in documentation rather than a clinical tool that changes outcomes.