Columbia Suicide Severity Rating Scale: How the C-SSRS Works

The Columbia Suicide Severity Rating Scale (C-SSRS), also called the Columbia Protocol, is a structured set of plain-language questions that asks whether and when a person has thought about suicide, prepared for it or attempted it^[1]. It rates the severity of suicidal ideation and records suicidal behavior; it does not diagnose a mental disorder.

Researchers at Columbia University, the University of Pennsylvania and the University of Pittsburgh built it with National Institute of Mental Health support for a 2007 NIMH study of treatments for adolescents with depression^[1]. Its first published validation appeared in the American Journal of Psychiatry in 2011^[2].

The scale's main contribution was shared terminology. FDA's 2012 draft guidance adopted 11 categories (five levels of suicidal ideation, five levels of suicidal behavior, and self-injurious behavior without suicidal intent) as its standard for classifying these events in drug trials, and named the C-SSRS as an instrument that sorts answers directly into them^[3]. The same guidance notes that the CDC adopted the definitions of the five behavior levels^[3].

The Columbia Lighthouse Project, which distributes the scale, says the protocol and its training are free for community and healthcare settings and for federally funded or nonprofit research^[1]. It also says no mental health training is required to ask the questions and that versions exist in more than 150 country-specific languages^[1].

The full C-SSRS measures four constructs^[2]:

• Severity of ideation: a 5-point rating where 1 is wish to be dead, 2 nonspecific active suicidal thoughts, 3 suicidal thoughts with methods, 4 suicidal intent and 5 suicidal intent with plan^[2].
• Intensity of ideation: five items, each rated on a 5-point scale: frequency, duration, controllability, deterrents and reasons for ideation^[2].
• Suicidal behavior: actual, aborted and interrupted attempts, preparatory behavior and nonsuicidal self-injurious behavior^[2].
• Lethality: the actual medical damage of attempts on a 6-point scale, or potential lethality on a 3-point scale when actual damage is zero^[2].

Keeping nonsuicidal self-injury as its own category separates self-harm without intent to die from a suicide attempt. The developers describe the C-SSRS as the first scale to cover the full range of suicidal thoughts and behaviors, including preparation and aborted attempts^[1].

In clinical trials, the full scale is given at baseline as a lifetime assessment and again at each visit^[3]. FDA states that a C-SSRS rating is not complete until information from family, caregivers and medical records has been integrated^[3].

The C-SSRS Screen Version is the short form used for triage. The developers say their shortest screeners run from two to six questions depending on the answers^[1], and VA guidance notes that either 3 or 6 items are asked on the standard Screener^[8].

Questions 1 and 2 are always asked and cover the past month: “Have you wished you were dead or wished you could go to sleep and not wake up?” and “Have you actually had any thoughts of killing yourself?”^[4]. A yes to question 2 leads to questions 3, 4 and 5, which ask in turn about thinking of how, having some intention of acting, and having worked out a plan with intent to carry it out^[4].

A no to question 2 skips to question 6, which asks whether the person has ever done, started to do or prepared to do anything to end their life^[4]. If the answer is yes, the screener asks whether that happened within the past three months^[4].

The screener has no total score. The highest item answered yes sets the tier: a yes to item 1 or 2 indicates low risk, item 3 moderate risk, and items 4, 5 or 6 high risk under the screener's general guidelines^[5].

What follows a given tier is decided locally. The Lighthouse Project says organizations set criteria for hospitalization, counseling, referral and other actions, informed by the yes or no answers and by how recent the thoughts and behaviors were^[1].

The full scale produces subscale scores rather than one number. Severity is a 6-point ordinal rating from no ideation up to 5 for suicidal intent with plan, and the five intensity items sum to a total from 2 to 25^[7].

The intensity subscale held together well in testing. Its internal consistency (Cronbach's alpha) was 0.937 to 0.946 in the adolescent attempter study and 0.73 across the other two 2011 studies^[2]; in 178 adolescents seen for psychiatric emergencies, it was 0.81^[7].

The 2011 validation drew on three multisite studies: a treatment study of 124 adolescent suicide attempters, a medication trial with 312 depressed adolescents, and 237 adults seen in an emergency department for psychiatric reasons^[2]. The authors concluded the C-SSRS is suitable for assessing suicidal ideation and behavior in clinical and research settings^[2].

For classifying behavior, the scale matched an independent suicide evaluation board with 100% sensitivity and 96% specificity for actual and interrupted attempts combined^[2]. In the adult emergency sample it identified recorded lifetime actual attempts with 100% sensitivity and 100% specificity^[2].

It also carried predictive signal. In the adolescent attempter study, each one-standard-deviation increase in lifetime ideation severity at entry was associated with 45% higher odds of a suicide attempt during the study, and adolescents who reported intent or intent with a plan had 3.26 times the odds^[2].

Results in larger emergency cohorts

The largest independent test followed 18,684 consecutive patients screened at Stockholm's psychiatric emergency department between May 2016 and December 2017; 107 died by suicide within a year^[5]. A cutoff of 3 (thoughts with a method) identified 56.5% of those who died within a week, with 75.6% specificity^[5].

Accuracy fell as the window lengthened, with AUC values of 0.72 at one week, 0.69 at one month and 0.62 at one year^[5]. The positive predictive value was 1.2% or lower at every cutoff, which the authors attribute largely to how rare suicide is within any short period^[5].

Other cohorts point the same way. Among 804 Swedish adults assessed after self-harm, 20.5% made a nonfatal or fatal suicide attempt within six months, and C-SSRS scores were only somewhat better than chance at sorting them (AUC 0.62 to 0.65)^[6]. In 178 adolescents aged 13 to 17, the intensity score predicted a suicide attempt at a return emergency visit within a year, with the duration of suicidal thoughts standing out among its items^[7].

Drug development was the first regulatory use. FDA's 2012 draft guidance calls the C-SSRS acceptable for prospective assessment and recommends such assessment in trials of any drug for a psychiatric indication, antiepileptic drugs and other neurologic drugs with central nervous system activity^[3].

FDA reports that the screening questions typically take 1 to 2 minutes when answers are negative, and the full interview usually takes under 10 minutes even with several positive findings^[3]. In almost 15,000 administrations of the electronic self-report version (eC-SSRS), average completion time was 3.5 minutes without positive findings and about 7 to 8 minutes with them^[3].

The Department of Veterans Affairs uses the C-SSRS Screener as part of its Suicide Risk Identification Strategy^[8]. The 2024 VA/DoD clinical practice guideline suggests a validated screening measure, a “weak for” recommendation, and lists the C-SSRS Screener with the Suicide Cognitions Scale-Revised and the PHQ-9 as tools with evidence for the general population^[8]. Data on veterans are on the veterans and suicidal ideation page.

Screening as a population strategy is less settled. In 2023, the U.S. Preventive Services Task Force concluded that evidence is insufficient to assess the balance of benefits and harms of screening for suicide risk in adults, including pregnant and postpartum people and older adults, an “I” statement^[9]. That finding concerns screening programs, not one instrument; the tools the Task Force names as examples are the Beck Hopelessness Scale, the SAD PERSONS Scale and the SAFE-T^[9].

Outside medicine, the developers report use in schools, faith communities, hospitals and the military in 45 nations^[1]. That figure comes from the Lighthouse Project itself.

The full scale has a baseline version that captures lifetime history and is repeated at later visits for the period since the last assessment; FDA also notes a version designed for patients already enrolled in a trial^[3]. The Screen Version asks about ideation in the past month and behavior over a lifetime and the past three months^[4].

The eC-SSRS is a computer-administered self-report version that uses probe questions similar to those a human interviewer asks^[3]. The scale was validated in both adolescents and adults^[2], and the developers describe it as suitable for all ages^[1].

A yes on the C-SSRS records that suicidal ideation or behavior is present and how severe it is. It does not forecast with precision who will die by suicide, as the low positive predictive values in the Stockholm cohort show^[5].

The Stockholm authors caution that a low-risk rating should not override a patient's own stated concern about suicide, and that risk stratification needs other known risk factors^[5]. They found the screener most useful for starting a structured conversation and deciding who needs a fuller assessment^[5].

Missing data is a practical limit. In that cohort, 20.7% of patients lacked an ideation or a behavior rating, and those patients were more often male and more likely to have substance use or psychotic disorders^[5].

For how often these thoughts and behaviors occur nationally, see the data on suicidal ideation, suicide attempts and U.S. suicide statistics. Crisis services are covered in 988 hotline statistics and the explainer on crisis intervention teams.