The Warm Handoff: Suicide Risk

This is a post in the Warm Handoffs series, an ongoing series on what the warm handoff looks like for specific clinical presentations in the Primary Care Behavioral Health (PCBH) model. The previous post in this series covered PHQ-9 screening, the warm handoff structure for depression, and the shared responsibility framework for suicide risk—and deliberately deferred a full treatment of suicide risk assessment to this post. A companion post, Suicide Risk Assessment in Primary Care: A Clinical Reference, covers the Safety Planning Intervention, means restriction counseling, chronic suicidality, welfare checks, and professional accountability in greater depth.

A note before reading: This post discusses suicide, suicidal ideation, and clinical approaches to suicide risk assessment in detail. It is written primarily for behavioral health clinicians and primary care providers. If you are personally struggling with thoughts of suicide or self-harm, please reach out for support—resources are listed at the end of this post.

---

In the depression post that opens this series, Herman Melville’s Bartleby served as a portrait of what depression looks like when no one has a name for it—the pale, motionless copyist whose gray refusal his employer cannot diagnose because he has no framework to work from. Primary care behavioral health, the argument went, is where that framework finally exists.

But Melville did not end his story with the lawyer’s dawning comprehension. He ended it with a death.

By the story’s close, the lawyer understands, after a fashion, that Bartleby is suffering. He even attempts, fitfully, to intervene—offering him alternative employment, inviting him into his own home. None of it reaches. And when the discomfort of proximity becomes too great to manage, the lawyer does not escalate—he does not call for help, does not bring in anyone more capable of reaching Bartleby, does not make the referral that might have mattered. He vacates the building. He moves his entire law practice to escape the problem, and Bartleby—still present, still refusing, still dying by degrees—is eventually removed by the new tenants and sent to the Tombs, where he stops eating and dies.

What failed Bartleby was not the absence of a framework. It was the lawyer’s inability to stay inside his own discomfort long enough to act from clinical reasoning rather than anxiety. The escape was not a referral. It was a flight.

That failure has a clinical name. It has a literature. And it happens in primary care more often than we tend to acknowledge.

A note on scope before going further: suicide risk assessment is a topic large enough to fill several textbooks—and has. This post focuses on the acute warm handoff encounter: what the BHC needs to know, do, and communicate when suicidal ideation enters the room and the patient is still in the building. It does not comprehensively cover the neurobiology of suicide, the full epidemiology of suicidal behavior, every validated assessment tool, or the extensive literature on specific populations including adolescents, older adults, veterans, and LGBTQ+ individuals. Each of those topics deserves its own careful treatment. The companion clinical reference post covers safety planning, means restriction, chronic suicidality, welfare checks, and professional accountability. The goal here is a tight, actionable account of the warm handoff encounter itself.

---

Suicide Risk Is the Whole Team’s Business

Before turning to the mechanics of assessment, one framing point deserves to come first. Suicide risk assessment does not belong to the BHC any more than depression management belongs to the BHC. The BHC is often the right person to take the lead on assessment and safety planning—with specialized training, time that the PCP cannot spare, and a skill set built for exactly this kind of encounter. But the responsibility for the patient’s safety before they leave the building is shared, and that distinction matters enormously in practice.

The most important consequence: a patient’s suicidal ideation will not wait until the BHC is next in office. If the BHC is unavailable—with another patient, out of the clinic, on leave—the PCP retains the responsibility to conduct an appropriate risk assessment and initiate safety planning. The warm handoff creates the conditions for collaboration; it does not dissolve the team’s collective obligation to the patient in the building (GATHER: Team-based; 4 Cs: Coordination—every member of the care team shares responsibility for the patient in the building, not just the clinician who happens to be available ).

The other consequence is less clinical and more human. When a patient dies by suicide, the emotional toll on everyone who had recent contact with them is significant and lasting—PCPs, BHCs, medical assistants, front desk staff. A team that has built clear workflows and shared ownership of risk encounters is a team that can support each other when the worst happens. A team where everyone assumed the responsibility belonged to someone else is not. Building that shared culture is part of what the BHC’s consulting role makes possible, and it is worth doing before the crisis rather than after.

Clinic workflows matter here in ways that extend well beyond the clinical encounter. A patient who walks up to the front desk and discloses suicidal ideation and is told they can be scheduled six weeks out is at acutely elevated risk. That disclosure warrants an immediate handoff to whoever on the care team is available—not a referral to a future appointment (GATHER: Accessible; 4 Cs: First Contact—crisis disclosure at the front desk is a care contact that demands an immediate response, not a scheduled one.). The same principle applies to call center staff: a patient calling in crisis who is transferred from person to person until they reach a BHC voicemail has experienced a failure of the system, not just the BHC. Solid triage protocols, clear escalation pathways for non-clinical staff, and explicit guidance about when to call 911 versus when to walk a patient to an exam room versus when to pull the PCP out of another appointment—these are not bureaucratic details. They are patient safety infrastructure.

---

The Primary Care Context Is Different

Experienced clinicians coming to PCBH from specialty mental health settings need to recalibrate their expectations—not of their clinical skills, but of the context in which those skills operate.

Outpatient suicide risk assessment typically happens with a patient who has, to some degree, chosen to be there. They are engaged in an ongoing therapeutic relationship, they have had previous contacts that established some clinical baseline, and the session is structured around the clinical concerns at hand. The BHC arriving at a primary care encounter for suicidal ideation has almost none of these advantages. The patient came in for a medication refill, or a sports physical, or a PHQ-9 score that read higher than last quarter. The BHC may be meeting them for the first time. The encounter happens in an exam room between a blood pressure check and the physician’s prescription review, and the patient has fifteen minutes before they need to pick up a child from school.

This is not a degraded version of a specialty encounter. It is a different clinical task entirely, one that plays to different strengths and requires a different orientation. Bryan and Rudd put it plainly: approaches to suicidal risk in primary care must fundamentally embrace the philosophy and clinical reality of that setting and must not mistakenly apply specialty mental health principles where they are inappropriate and cannot be sustained. That reorientation is the prerequisite for everything that follows.

The primary care BHC who approaches suicidal ideation with a specialty posture will be tempted toward two opposite errors. The first is over-triage: treating every item 9 endorsement as a psychiatric emergency, generating unnecessary hospitalizations, and eroding the patient trust the team needs for future encounters. The second is under-triage: treating the setting’s brevity as a justification for incomplete assessment, documenting “patient denied active SI” when what actually happened was that the BHC changed the subject. Neither response constitutes clinical care. Bryan and Rudd describe both error modes as rooted in the clinician’s internal state rather than the patient’s clinical presentation—a point developed further below.

Robinson and Reiter are direct on what competent risk management actually requires: not only identifying high-risk patients, but completing a meaningful staffing with the PCP, following through with patients who decline referral, and documenting clinical reasoning clearly. That is a higher bar than simply noting that the question was asked.

---

What Triggers an Assessment

The previous post in this series covered the most common trigger in detail: an elevated response to PHQ-9 item 9, which asks about thoughts of being better off dead or of hurting oneself in some way. Simon and colleagues found that endorsement of item 9 predicted subsequent suicide attempt and death in a dose-response relationship across more than 84,000 outpatients—a finding that underscores the item’s clinical significance while also demonstrating that ideation frequency alone does not define level of risk. The more important point from that post bears repeating: a positive response to item 9 does not constitute a suicide assessment. It requires one.

Item 9 is not the only trigger. Spontaneous disclosure—a patient who says, unprompted, that they have been thinking about dying—represents the patient actively seeking help rather than being caught by a screen, and the warm handoff in that context carries particular urgency. A PCP flag, a nursing note, a family member who pulls a medical assistant aside in the hallway—any of these can initiate the encounter. And because item 9 has documented limitations in identifying risk in multiracial and multiethnic populations, particularly among older adults , the experienced BHC does not rely on the screen alone. Clinical observation—a patient who seems more withdrawn than usual, a PCP who says simply “something feels off today”—carries weight that a single item cannot capture.

One distinction bears stating plainly: passive ideation (“I wouldn’t mind not waking up”) and active suicidal ideation with intent are not the same clinical picture, and they do not require the same response. The assessment that follows a warm handoff for elevated item 9 needs to distinguish between them. That is the job.

---

The Assessment Itself

The goal of a suicide risk assessment in primary care is not to replicate the depth of a psychiatric evaluation. It is to gather enough clinically meaningful information to make a defensible, documented disposition decision before the patient leaves the building. That requires covering several domains regardless of which structured tool, if any, guides the inquiry. In most cases, a Contextual Interview should be part of this process.

Bryan and Rudd recommend a hierarchical sequencing of questions—starting on familiar ground with the patient’s current symptom picture, moving through hopelessness, and arriving at suicide-specific inquiry—so that the escalation feels contextual rather than interrogative. The sequence matters less than the commitment to cover what must be covered.

A contextual interview approach, even within a suicide risk assessment, gives the care team a fuller and more clinically useful picture: not only the depth and nature of the patient’s pain, but their strengths, their relationships, their reasons for living, and the specific precipitants driving the current crisis . That richer picture directly improves both the risk formulation and the safety plan.

Ideation: What specifically is the patient thinking? How intense are the thoughts, and how long do they last? Bryan and Rudd make an important distinction between frequency and intensity of suicidal ideation: frequency—how often the thoughts occur—is less clinically significant than intensity—how severe and consuming the thoughts are when they occur. The clinician who asks only how often a patient thinks about suicide and receives “not very often” as an answer may be missing the more urgent clinical picture.

Plan and intent: Has the patient thought through how they would act on the ideation? Is there a specific method? Have they taken preparatory steps—counting pills, handling a weapon, writing a note? Preparatory and rehearsal behaviors are among the most significant indicators of suicidal intent , and asking about them directly is non-negotiable.

Behavior history: Previous attempts are the single most robust predictor of future suicidal behavior, outperforming depression severity, hopelessness, and nearly every other risk factor even when those factors are considered simultaneously . The patient who minimizes current ideation but discloses prior attempts warrants careful attention regardless of how they present today.

Means access: Does the patient have access to the means they have contemplated? This question must be asked directly. Clinicians who omit it because it feels intrusive are managing their own discomfort rather than the patient’s risk.

Protective factors: Reasons for living, social support, religious or cultural commitments, responsibility for others, engagement with treatment. These do not eliminate risk, but they contextualize it, inform the safety plan, and provide what Bryan and Rudd describe as a useful clinical turning point—a pivot from cataloguing what is wrong toward identifying what is worth protecting.

Precipitating context: What changed? A loss, a relationship rupture, a medical diagnosis, a job, a discharge from a system of support—understanding the precipitant shapes both the risk formulation and the intervention—a complete Contextual Interview will help with this step.

Several validated brief tools support this structured inquiry in primary care. The Columbia Suicide Severity Rating Scale provides behaviorally anchored ideation and behavior subscales that map naturally to disposition decisions and have been validated across healthcare settings including primary care; its screener version is brief enough for a warm handoff context. The Ask Suicide-Screening Questions is a four-item tool developed specifically for medical settings, validated across pediatric, adult, and psychiatric populations, and brief enough to administer in under two minutes.

No tool replaces clinical judgment, and the choice of instrument matters less than the commitment to use one consistently, document the results, and act on what they reveal.

---

The Problem of Clinician Anxiety

Farber observed, writing about patients who self-harm, that decisions to hospitalize are sometimes made not in response to patient risk but in response to the clinician’s need to assuage their own anxiety. The observation is uncomfortable precisely because it is accurate—and it extends directly to suicidal risk management.

Bryan and Rudd describe the two recognizable manifestations in detail. The first is fear, which produces denial: the provider finds reasons to believe the patient is not really at risk, avoids direct questioning, and delivers interventions at a lower intensity than the clinical picture warrants. The second is anxiety about litigation, which produces overreaction: the “better-safe-than-sorry” approach, overreliance on hospitalization as a default disposition, and clinical decision-making that targets the reduction of the clinician’s distress rather than the patient’s treatment needs. Jobes traces much of this anxiety to the reductionistic clinical stance—the assumption that the clinician’s job is to control the patient’s behavior, so that every bad outcome becomes evidence of clinical failure. The collaborative approach, by contrast, positions the patient as the expert on their own suffering and the clinician’s role as reducing that suffering alongside them—a reframe that distributes both responsibility and anxiety more accurately.

In primary care, these dynamics are amplified by contextual pressures that specialty settings do not impose in the same way. The team is visible—the PCP is in the next room, the medical assistant is nearby, the patient’s family is in the waiting area. Time is compressed. The BHC may not have an established relationship with the patient. The sense that something must be resolved before the appointment slot ends creates pressure toward premature closure, and premature closure in a suicide risk encounter is a clinical error.

The antidote is not to eliminate anxiety. Anxiety about suicidal patients is adaptive and appropriate when it drives careful clinical reasoning rather than replacing it. What distinguishes a defensible clinical decision from an anxiety-driven one is not the outcome but the reasoning behind it. A patient hospitalized because their means access was uncontrollable, their intent was clear, and their social support was absent represents a different clinical picture from a patient hospitalized because the BHC felt uncomfortable and an involuntary hospitalization was the path of least resistance. Both may end up in the same bed. Only one of them needed to be there.

(A forthcoming post in the Training and Education series will address clinician anxiety in the specific context of supervision and trainee development—including how supervisors can support learners through difficult suicide risk encounters without inadvertently signaling that the patient is too complicated to be seen by someone in training.)

---

The Team Communication Loop

The BHC does not close the encounter with a suicidal patient without consulting with another member of the care team. This is not optional and not a matter of preference—it is a structural requirement of the PCBH model and a clinical safeguard that serves the patient, the team, and the BHC.

In most encounters, that consultation is with the PCP: the staffing conversation covers risk level, disposition, safety plan status, means restriction steps taken or recommended, and what follow-up is needed. When the PCP is unavailable, another BHC colleague or a clinical supervisor is the appropriate alternative—and that consultation should be documented in the clinical note regardless of who conducts it. A solo BHC making a high-stakes risk management decision without any collegial input is in a clinically and professionally vulnerable position that the PCBH team structure exists to prevent.

The note format matters here. In settings that use the APSO format, the assessment and plan are visible at the top before the PCP has to scroll—one of many reasons APSO serves urgent encounters particularly well. A face-to-face verbal handback before the patient leaves the building is the gold standard; a message sent through the EHR while the patient is still in the exam room is acceptable when face-to-face is not possible.

The BHC who completes an assessment, builds a safety plan, and then sends an In Basket message that the PCP reads the following day has not completed the clinical loop. The patient’s risk does not pause while the provider reads the message.

This is also a moment for the BHC to lean into what team-based primary care is actually for. Step away from lunch. Pause an existing appointment. Go out of your way when a team member asks for help with a crisis. Suicidal risk encounters are frightening for everyone, and a well-functioning team helps each other through them—not because that is protocol, but because that is what it means to work as a team.

---

When to Escalate

The decision to escalate—to a higher level of outpatient care, an emergency evaluation, or inpatient admission—follows from the assessment, not from the BHC’s discomfort. The relevant clinical variables are: imminent intent with a specific plan, access to means that cannot be restricted before the patient leaves the building, absence of social support sufficient to maintain safety, failure of previous safety planning at this level of care, and active intoxication that renders voluntary safety planning incoherent.

When escalation is indicated, the BHC’s role does not end at the decision. Staying with the patient, involving the PCP directly, contacting family or supports if the patient consents, and documenting the clinical reasoning for escalation are all part of the encounter. Patients referred to emergency settings without a direct clinical handoff face meaningful barriers to engagement—a phone call to the receiving facility, or at minimum a summary the patient can carry, reduces the probability that the referral ends there.

Robinson and Reiter’s competency standard bears restating: the BHC should follow up with high-risk patients who decline referral until they are either engaged in specialty care or their risk level has diminished. That follow-through commitment does not end when the patient walks out of the building.

---

Closing

Melville’s lawyer was not an indifferent man. He is one of the more sympathetically drawn characters in the American canon—genuinely troubled by Bartleby’s suffering, genuinely motivated to help. What he could not do was stay. He could not remain in contact with a problem that did not resolve cleanly, could not resist the organizational pressure to make the discomfort disappear, could not act from reasoning when relief was so much more available. So he moved his office, and Bartleby died in the Tombs.

Primary care gives the BHC something the lawyer never had: a framework, a team, a set of structured tools, and a literature that tells us what actually reduces the probability of a bad outcome. What it cannot provide is the willingness to use those tools without flinching. That willingness does not come from the tools. It comes from the clinician.

---

If you or someone you know is in crisis: The 988 Suicide and Crisis Lifeline is available 24 hours a day, 7 days a week.

• Call or text: 988
• Chat: 988lifeline.org
• For veterans: Press 1 after dialing 988, or text 838255
• For Spanish speakers: Press 2 after dialing 988

If there is immediate danger, call 911 or go to your nearest emergency room.

---

This post was written by the author. Claude (Anthropic) assisted with verifying current‑event details, refining the structure of the series, supporting revision and editing, and generating ZotPress and WordPress shortcode syntax for citations and formatting. All positions, interpretations, and personal statements are the author’s own.

---

References