When a Client Says They Want to Die
A Practical Clinical Guide to Responding to Suicidal Ideation in the Therapy Room
Over the past several years, I have conducted hundreds of suicide risk assessments in a psychiatric emergency department. That experience made clear that graduate training rarely prepares clinicians for what it actually feels like to navigate these situations in real time. This guide is an attempt to address that gap.
This is not a rigid protocol. It is a clinical framework aimed at helping clinicians outside of hospital settings, including those in private practice, community mental health, and outpatient care, balance thorough risk assessment with maintaining the therapeutic relationship, while respecting patient rights throughout.
Clinician Response: What You Communicate Before You Speak
Clients presenting with suicidal ideation are often monitoring the clinician’s response closely. Body language, tone, and the clinician’s immediate reaction will influence whether the client continues to disclose or begins to minimize what they are sharing. A client who senses panic, alarm, or a shift into procedural mode may pull back, withhold information, or downplay the severity of what they are experiencing.
Remaining composed and non-reactive is a clinical skill, not simply a matter of demeanor. It signals to the client that the clinician is capable of holding what is being shared and that the session remains a safe space for disclosure.
What to say and what to avoid:
Helpful responses
“I can see how dire things feel for you right now. If you’re comfortable, I’d like us to talk more about this. I want to try to understand what you are going through.”
Responses to avoid
“You should really think about all the good in your life.” -- The client may not experience their situation that way and may feel dismissed. / “This is very serious and I need you to tell me more.” -- This introduces pressure that can shut down disclosure rather than encourage it.
The clinical priority at this stage is to keep the client talking. Shifting prematurely into crisis management risks cutting off information that is necessary for accurate assessment.
Clinical Assessment: What You Are Looking For
Before any clinical action can be determined, the clinician needs a clear picture of what the client is experiencing. The following domains guide that assessment.
1. Nature of the Ideation
Passive death wish and active suicidal intent are clinically distinct. A client expressing a wish that their pain or problems would end occupies a different risk category than one who has formed an intention to act. Both warrant clinical attention, but the assessment and response will differ.
Is there intent to act, or is the client expressing a wish that their suffering would stop?
Is there a plan?
Assess its specificity and feasibility. A client who identifies a method but lacks access to the means and has no intention of obtaining them is in a different risk category than one who does.
Is there a timeline? The identification of a specific date or triggering event, such as an anniversary, a legal proceeding, or the end of a relationship, represents an escalation in risk. A timeline indicates that ideation has moved toward planning.
2. Assessing Risk Level: The Scaling Question
A direct and reliable way to assess current risk is to ask the client to quantify it:
The Scaling Question
“On a scale of 1 to 10, how likely do you think you are to actually hurt yourself between now and when we next meet? Where 1 is not at all and 10 is certain.”
Once the client provides a number, follow up from both ends. The examples below illustrate how this plays out in practice.
If a client says they are a 4, ask what kept them from saying 5 or 6:
Example
“You said you’re a 4. What made you not say a 5 or a 6?”
This tends to surface protective factors the client may not have volunteered otherwise. Responses like “I would never put my kids through something so awful” or “I get angry in the moment but I know I would not follow through” identify what is currently keeping the client safer and become important clinical and therapeutic reference points.
If a client says they are a 9, ask what would have made that number lower:
Example
“You said you’re a 9, almost certain you would hurt yourself. What would have made you less certain? What would have made it a 7 or 8 instead?”
This can reveal either genuine intent or specific triggers that, if addressed, could reduce risk. A response like “because I have pills at home and I overdosed before, so I know I might not be able to stop myself” signals both access to means and a prior history that warrants immediate attention. It also opens a direct clinical conversation about means restriction: could the client agree to remove those medications from the home, perhaps giving them to their emergency contact or another trusted person until the risk subsides? A response like “because things feel different this time and that worries me” indicates the client themselves recognizes an escalation and that observation should be taken seriously.
The score provides a useful starting point, but the follow-up questions are what tell the clinical story. Listen for ambivalence, protective factors, and the degree of specificity in the client’s intent.
3. History of Self-Injurious Behavior
Self-injurious behavior should be assessed carefully, with particular attention to its function. The same behavior can carry very different clinical implications depending on why the client engaged in it.
Has the client previously or are they currently engaged in non-suicidal self-injury (NSSI)? Clarify the function: coping, emotional regulation, communicating distress, or self-punishment.
That distinction matters clinically. Self-harm in service of coping, however maladaptive, reflects an attempt to manage or survive a stressor. Self-harm as self-punishment reflects something directed against the self, which carries different implications for intent and risk. Understanding the function shapes both the risk assessment and the treatment focus.
Clinicians should also assess for accidental lethality. Clients who self-harm to manage distress or elicit a response, particularly those with poor impulse control or significant emotional dysregulation, may unintentionally inflict lethal harm. The absence of suicidal intent does not eliminate this risk.
4. History of Suicide Attempts
A history of prior attempts is among the strongest available predictors of future risk. Before accepting a reported attempt at face value, the clinician should explore it. Clients sometimes describe past self-injurious behavior as a suicide attempt when, on further discussion, the intent was to cope or communicate distress rather than to die. That distinction carries clinical weight.
How many prior attempts? Did any require medical intervention?
What was the method and the degree of lethality?
Were there any self-interrupted attempts? What stopped the client?
How does the client feel about having survived? Relieved, indifferent, or wishing the attempt had succeeded?
Were similar stressors present during prior episodes of ideation or attempts? Recurring themes, such as loss or rejection, are worth noting even when the specific circumstances differ.
Assess carefully for ambivalence. A client who is genuinely uncertain about whether they are glad to be alive presents a higher risk profile and warrants more thorough follow-up and closer monitoring.
5. Current Medication Compliance
For clients prescribed psychiatric medication, the following should be assessed:
Are they taking medications as prescribed?
Have there been any recent changes to medications or dosing?
Have there been changes in physical health, substance use, or sleep that may be interacting with their current regimen?
Recent discontinuation of psychiatric medication is an independent risk factor. It is also worth noting that if a client has stopped taking their medication, those pills may still be in the home. For a client at elevated risk, that represents an accessible means and should be addressed directly as part of means restriction.
Structured Assessment: The C-SSRS and Clinical Style
The Columbia Suicide Severity Rating Scale (C-SSRS) is a standardized instrument used widely in hospital and crisis settings. The full version includes 25 questions across domains covering ideation type and intensity, history of suicidal behavior, and deterrents. It provides a reliable structure and is a legitimate clinical tool, particularly for clinicians who prefer a formal framework or work in settings where standardized documentation is required.
The domains the C-SSRS covers, ideation, intent, plan, history of attempts, access to means, and protective factors, can also be addressed through a direct clinical conversation without a formal instrument. In my practice, I do not administer the C-SSRS as a form. I have found that clients are more forthcoming and the information gathered is more clinically useful when the assessment unfolds conversationally rather than as an intake procedure. Clients who feel they are being processed through a checklist are less likely to offer the kind of detail that informs good clinical judgment.
This is not an argument against structured assessment. It is a note that the format matters. Whether a clinician uses a formal instrument or a conversational approach, all relevant domains must be covered, and the client must feel that the interaction is directed at understanding them rather than completing a form.
Contracting for Safety: Why It Does Not Work
A no-suicide contract is an agreement, written or verbal, in which a client commits to not harming themselves before the next session. The practice has not been supported by research. These agreements have not been found to reduce suicidal behavior, and they can give clinicians a false sense of security. From the client’s perspective, being asked to make a formal promise during a crisis can feel coercive or dismissive of the severity of what they are experiencing.
The Joint Commission, the U.S. body responsible for accrediting healthcare facilities, moved away from recommending these contracts, and the broader clinical field has largely followed. Safety planning, described below, is the evidence-based alternative.
Safety Planning: What to Use Instead of Contracting
A safety plan is a personalized set of strategies and resources a client can draw on if suicidal ideation intensifies between sessions. Unlike a no-suicide contract, it is built collaboratively, focuses on actionable steps, and is grounded in the client’s actual circumstances rather than an abstract promise.
Formal safety planning tools are available, including the Stanley-Brown Safety Planning Intervention, and clinicians who prefer a structured format can use them as a guide. The same information can also be gathered conversationally. The goal is for the client to leave the session with a working understanding of their warning signs, their coping options, who they can contact, and what to do if those resources are not enough. The format is secondary to whether that goal is achieved.
In practice, I integrate safety planning into the natural flow of the session, typically toward the end, without signaling a formal transition. The questions below cover the necessary clinical ground while keeping the conversation focused on the client’s experience rather than a procedure:
What does the rest of your day look like after we finish here?
What do you have going on this week: work, people you will see, things that provide some structure?
If things get difficult before we meet again, who would you reach out to first?
Are there environments or situations you know tend to increase your distress? Are any of those avoidable for now?
If things get harder before they improve, what do you think will be driving that?
Means restriction should also be addressed as part of safety planning. This includes assessing whether firearms, medications, or other lethal means are accessible in the home and discussing practical steps to limit that access. Research consistently identifies means restriction as one of the most effective suicide risk reduction strategies available, and it is often underaddressed in outpatient settings.
Trusted Contacts and Collateral Information
Most clinicians collect emergency contact information at intake. That contact, and any other trusted person in the client’s life, is a resource that should not be overlooked when suicidal ideation is present. Clinicians who do not already collect this information in outpatient practice should strongly consider doing so.
When suicidal ideation is disclosed, it is worth exploring whether there is someone in the client’s life who is aware of what they are going through and what role that person might play in supporting them. This does not need to be limited to the named emergency contact. Relevant questions include:
Does your emergency contact know how you have been feeling?
Is that something you would be comfortable sharing with them, or would you want help thinking through how to approach that conversation?
Do you think involving them at this point would be helpful, or would it create additional stress?
Would it be useful for me to speak with them, either together or separately?
Involving a trusted contact should not be done without clear clinical justification and, wherever possible, the client’s agreement. Confidentiality is not broken simply because a clinician is concerned. The threshold is imminent risk to self or others. Short of that threshold, the preferred approach is to work with the client rather than around them. If reaching out to a contact is clinically warranted and the client is reluctant, the clinician can discuss the reasoning openly and, in many cases, encourage the client to make that call themselves or agree to be present when it is made. That approach preserves trust while still mobilizing support.
Collateral information from someone with regular contact with the client is also one of the most valuable inputs to a risk assessment. In psychiatric emergency settings, it is routinely gathered and frequently changes the clinical picture. A collateral contact can provide information about recent behavioral changes, history of dangerous behavior, and their own assessment of current risk that the client may not offer directly. A contact reporting that a client “always says that when he’s angry but has never acted on it” carries clinical weight. So does hearing the opposite.
Increasing Session Frequency
For clients presenting with significant suicidal ideation who are not at imminent risk, increasing session frequency is often the most appropriate clinical response. It maintains the therapeutic relationship, increases monitoring, and avoids the disruption and potential harm of unnecessary hospitalization.
The offer should be framed as support rather than a crisis response:
Example
“I can see you are really struggling right now and I do not want to leave a whole week between sessions. Would you be willing to come in again before the weekend?” Or, toward the end of the week: “Can we plan to meet again Monday before the week gets started?”
Clients agree to this more often than not. The offer itself communicates investment in the client’s wellbeing and continuity in the relationship, both of which have clinical value independent of the additional contact.
The client’s reaction to the offer is also worth noting as clinical data. A client who hesitates due to cost or scheduling constraints is demonstrating future-oriented thinking, which is clinically meaningful. Whatever the reaction, the clinician should discuss any observations directly with the client rather than drawing conclusions silently. Checking in on what is behind the response often yields information that is directly relevant to the assessment.
Communicating the Decision to Refer for Acute Crisis Intervention
The decision to initiate emergency intervention is reached when a client cannot provide reasonable assurance that they will not harm themselves before the next contact, and the clinician shares that uncertainty. Distress or ideation alone does not cross that line. What crosses it is the combination of a client who acknowledges they may not be able to stop themselves and a clinician who, having assessed the full clinical picture, cannot reasonably conclude otherwise.
Once that determination has been made, how it is communicated matters. A client who understands the reasoning is more likely to engage with the process cooperatively, and the clinician’s transparency at this moment is an expression of the same respect for the client that has guided the session throughout.
The language below illustrates how this might sound in practice:
Example: Communicating a Decision to Call EMS
“I want to be straightforward with you about where I’m at. You’ve told me that you’re not certain whether you’ll harm yourself before we meet again, and I’m not certain either. At this point, neither of us knows what you might do, and the risk of you taking a permanent, irreversible action is something I take seriously. Given that, I’m going to contact EMS. An ambulance will bring you to the hospital. I’m going to stay here until they arrive. I’ll speak with them directly, find out where you’ll be taken, and I’ll give them my contact information. I’d like to check on you. I plan on calling to see how you’re doing and to find out what I can do to help from here.”
This kind of statement names the shared uncertainty honestly, explains the clinical reasoning without being clinical in tone, and makes clear that the clinician’s involvement does not end with the call. Clients being referred for emergency care often feel abandoned. An explicit commitment to follow up can meaningfully change that experience.
A Note on Confidentiality When the Hospital Is Involved
If a client has not consented to the release of information, the hospital will not be able to share updates or clinical details with the referring clinician. This is a limitation clinicians should be aware of and address with the client.
The restriction runs in one direction only. Even when a hospital cannot release information to the clinician, they can still receive it. A clinician who has relevant information that could affect the hospital’s assessment should provide it. That information can directly shape the level of intervention the client receives, and the inability to receive updates is not a reason to withhold what is clinically useful to share.
When the Client Agrees to Go Voluntarily
In some situations, rather than initiating an emergency referral directly, the clinician may have a conversation with the client about going to the hospital themselves. This is appropriate when the clinician’s judgment supports it and the client is a credible candidate for voluntary follow-through.
In these cases, it is reasonable to establish a clear agreement: the client will go to the hospital within a defined timeframe, and if the clinician has not heard from them or confirmed that they went, there is a pre-agreed plan for the clinician to contact the client’s emergency contact by a specified time. This arrangement preserves the client’s autonomy and, when the clinical relationship supports it, can be the right call.
Clinicians should enter this arrangement with clear eyes about its limits. There is less certainty that the client will actually go, and that uncertainty is the tradeoff for a less disruptive intervention. This is not a fallback option for avoiding a difficult conversation. It is a deliberate clinical decision that should be documented as such.
Hospitalization: What Clinicians Should Understand
Hospitalization is not the appropriate response to suicidal ideation by default. It is a specific intervention for imminent risk, and understanding how that determination is made helps clinicians make better referral decisions.
In a psychiatric emergency setting, the assessment is focused on imminent threat to self, others, or property. A client presenting with suicidal ideation who has no history of self-harm or prior attempts, no expressed plan or intent, no collateral concerns, and who is engaged and able to participate in safety planning is generally not a candidate for involuntary commitment. An outpatient safety plan with increased session frequency is typically the more clinically appropriate response.
When voluntary admission is being considered, whether initiated by the clinician or agreed to by the client as described above, clinicians should be transparent about what inpatient admission actually involves: time away from work and personal life, agreement to psychiatric medication management, and a length of stay that varies based on clinical progress and typically ranges from a few days to several weeks.
Involuntary commitment is reserved for situations of imminent danger where the client cannot be safely maintained in a less restrictive setting. It requires a clinical and, in most jurisdictions, legal determination. The standard is imminent danger, not the presence of distress or ideation alone.
Documentation
Thorough documentation is part of good clinical care, not just a liability consideration. A well-maintained clinical record tells the story of what the clinician assessed, what was found, and what decisions were made and why. That narrative supports continuity across sessions, provides context if the client is ever referred to another provider, and becomes part of the clinical record if a higher level of care becomes necessary.
At minimum, when suicidal ideation is presented, the record should reflect:
Nature of ideation: passive wish versus active intent
Presence or absence of a plan, and its specificity and feasibility
Access to means and any discussion of means restriction
Protective factors identified
History of self-harm or prior attempts, including method and lethality
Rationale for outpatient management versus referral to a higher level of care
Safety planning steps taken
Any consultation obtained
Document the clinical reasoning, not just the conclusions. A note that reads only “client denied suicidal ideation” tells very little. A note that reflects what was assessed, what the client reported, what protective factors were present, and why outpatient management was determined appropriate tells the clinical story and supports continuity of care. When protective factors are strong and risk is being managed outpatient, explicitly note the client’s engagement in safety planning and any ambivalence toward self-harm as key supports for that clinical decision.
Clinical Judgment and the Limits of Certainty
When a client presents with suicidal ideation, the clinician faces a decision that carries real consequences in both directions. Acting when it was not necessary risks damaging trust and driving the client away from treatment. Failing to act when it was necessary can have irreversible consequences.
What reduces the risk of getting it wrong is thorough assessment, honest consultation when uncertainty is significant, and clear documentation of the reasoning behind the clinical decision. There is no way to eliminate the weight of these decisions. The work is to meet that weight with clarity, honesty, and as much care as the moment allows.
If you are a clinician interested in consultation or have thoughts on this piece, I would be glad to hear from you. If you found this helpful and are curious about the work I do, I practice in-person in downtown Jersey City and virtually throughout New Jersey. Feel free to get in touch.
