In-Session Safety: Therapists Share the One Question That Clarifies Risk Without Rupture
When a client hints at suicidal thoughts, knowing what to say next can mean the difference between connection and crisis. Experienced therapists reveal a single, carefully worded question that opens the door to honest conversation about safety without damaging trust. This approach, backed by clinicians who handle high-risk cases daily, offers a practical alternative to standard protocols that often feel clinical and disconnected.
Ask What Has Kept You Here
This sits at the centre of my clinical work, particularly with trauma clients where passive suicidal ideation often lives just beneath the surface of what's being directly said.
The framework I use internally is: *can this person leave the session and stay safe tonight?* Not next week - tonight. When someone hints at not wanting to continue but pulls back from disclosing a plan, I'm tracking the body as much as the words. Dissociation, a sudden flattening of affect, or a strange lightness after heaviness - these shift my assessment more than verbal content alone. EMDR work has sharpened this for me considerably; you become attuned to when someone's window of tolerance has genuinely collapsed.
The single question that has consistently clarified things without triggering withdrawal: *"What's kept you here until now?"* It does two things simultaneously - it opens the door to protective factors without demanding they justify their pain, and it tells me immediately how thin that scaffolding currently is. A client who answers with specific names, routines, or future moments is in a different position than someone who goes quiet or says "I don't know anymore."
What can wait is the full exploration of *why* they feel this way. What can't wait is whether they have something concrete enough to act on and whether they're alone tonight. Those two things I always establish before the session closes, and I do it directly - most clients, especially those who've experienced trauma, feel more contained when you name the elephant rather than both of you pretending it isn't there.
Conduct a Compassionate Risk Assessment
If a patient indicates that they do not wish to continue, regardless of whether they express an intent to kill themselves, I evaluate that as part of a comprehensive, compassionately conducted suicide risk assessment. The primary focus of my evaluation is to determine the intensity of the patient's suicidal ideation, level of suicidal intent, degree to which the patient has access to lethal means, presence of protective factors, and capacity to remain safe.
One single item that I have found particularly useful in this regard is the following: "What was it like for you when you said you didn't want to continue?" This allows the patient to provide additional information about their feelings without being evaluated, inquired into, or judged. Often, this provides the necessary background to assess the immediacy of the risk, while maintaining the relationship with the patient.
If the results of the risk assessment suggest that the patient is at imminent risk of harming themself, or indicates that the patient is incapable of developing a safety plan, immediate action will be taken by way of intervening on behalf of the patient. If the assessment does not indicate an imminent risk to the patient's health and safety, then a collaborative effort will occur to develop a safety plan with the patient and support system, enhance available supports, schedule close follow-ups, and educate the patient regarding how to obtain assistance should their risk increase.
Build a Clear Escalation Path
I am Elijah Fernandez, co-founder and CTO of CEREVITY. I am not the clinician making the in-session call, I am the person who built the system around it, so that is what I can speak to.
The mistake at the system level is leaving a clinician alone to improvise the most consequential moment of their week. My job is to make sure that when a clinician has a concern about a client's safety, they are never deciding in isolation what happens next. So the protocol I built removes the guesswork from everything except the clinical judgment itself: there is a defined escalation path, clear criteria for when it triggers, immediate access to support so the clinician is not the only set of eyes, and a documentation and handoff flow that closes the loop rather than leaving it dangling.
The principle is that the clinician's job is to read the person in front of them, and the system's job is to make the right next step obvious and frictionless once they do. When safety depends on whether an individual clinician happens to remember the right move under pressure, the design has failed. Build the protocol so the correct response is the path of least resistance, and you turn the hardest moment in the work into something the whole system carries, not one person.
Request a Call Before Decisions
This is something I navigate regularly in my work at Grace Recovery Services, where so many clients carry trauma underneath their addiction—and that same trauma often sits underneath suicidal ideation they're not ready to name directly.
When someone hints but denies a plan, I stop trying to assess and start trying to connect. The clinical instinct is to run through a checklist, but that's exactly what ruptures rapport. Instead, I'll say something like: "Help me understand what 'not wanting to go on' feels like for you right now." That open door lets them define it—and how they define it tells me almost everything about urgency.
The shift I watch for isn't in their words, it's in their body. Are they leaning in or pulling away? Is there relief when the topic surfaces, or shutdown? In our trauma-informed model, we treat the body as part of the story. A client who visibly exhales when the topic comes up is usually asking to be seen—that's a very different clinical moment than one who goes flat and distant, which tells me the risk is closer than their denial suggests.
What's consistently worked for me: once I have a read on their body language and affect, I ask them directly to make one agreement with me before the session ends—not a safety contract in the formal sense, but a relational one. "Can you call us before you make any decision alone?" That question keeps the relationship as the safety net, which is exactly what our whole model at Grace is built on—no one recovers, or stays safe, in isolation.
Clarify Pain Versus Death Intent
The fastest way to lose a client's honesty about safety is to make them feel assessed. When someone hints they don't want to go on but denies a plan, I stay close and curious instead of clinical and interrogating, because the moment it feels like a checklist or a tripped protocol, they shut the door and tell you what gets them out of the room. I want to understand the weight they're carrying and, just as much, what's kept them here so far, since a person's own reasons for staying are often the most stabilizing thing available. And I'm tracking safety directly the whole time, underneath the warmth, not instead of it.
The grounding question I rely on separates exhaustion from intent without any alarm in my voice: "When you say you don't want to go on, do you mean you want the pain to stop, or that you want to stop being here?" It's gentle, it's direct, and it tells me what I actually need to know while the person still feels met rather than managed.
