You already know the name Tarasoff. Every California law and ethics course you’ve sat through has walked the 1976 decision, the shift from a duty to warn to a duty to protect, and the long academic argument about what the court actually meant. That history isn’t background material. The duty comes from the case law, and it still does.
Civil Code §43.92 does something narrower, and the difference matters the first time you have to explain your decision to someone who wasn’t in the room.
Read the structure, not the summary
The statute says there shall be no monetary liability and no cause of action against a psychotherapist for failing to protect from a patient’s threatened violent behavior — except if the patient has communicated to the psychotherapist a serious threat of physical violence against a reasonably identifiable victim or victims.
That’s an immunity provision with one hole in it. The duty came from Tarasoff and stays there. The hole marks where the statute’s immunity gives way.
Subdivision (b) closes the other half: no monetary liability against a psychotherapist who, in those circumstances, discharges the duty to protect by making reasonable efforts to communicate the threat to the victim or victims and to a law enforcement agency. Reasonable efforts toward both are what the statute asks — not a completed conversation with either — and making them takes failure-to-protect liability off the table.
So §43.92 protects you at both ends of the question and nowhere else. You aren’t liable for failing to protect when the statutory trigger wasn’t met. You aren’t liable for failing to protect when a threat was communicated and you made reasonable efforts to reach the victim and law enforcement. Your exposure sits in the gap between them — a threat was communicated, and you didn’t act.
The term “psychotherapist” reaches you either way. Evidence Code §1010(c) covers a person licensed as a clinical social worker engaged in applied psychotherapy of a nonmedical nature, and §1010(g) covers a registered associate clinical social worker practicing under supervision. Associates sometimes assume the statute runs to the supervisor. It runs to you.
The trigger is a communicated threat
This is the part that cuts against instincts many of us develop over years of practice.
You may be sitting across from a client whose affect, history, and escalating agitation convince you they could become violent. Your body is telling you something is wrong. But if no threat has been communicated by the client — or in the circumstance recognized in Ewing v. Goldstein (2004), where the client’s own statement reached the psychotherapist through an immediate family member relaying it for the purpose of advancing the client’s therapy — the immunity holds and no notification duty attaches.
Your clinical inference alone doesn’t open the exception. That doesn’t mean you do nothing. It means the thing you do isn’t notification.
The reverse holds too, and it’s the half practitioners skip. Communication isn’t a switch that flips the duty on. The statute asks for a serious threat, and Ewing itself turned on whether the therapist believed the patient intended to kill or cause grave bodily injury. Your assessment of what you heard stays inside the analysis. It isn’t a step you get to skip because something alarming was said out loud.
And Ewing is narrower than practitioners tend to remember. The relationship, the purpose, and the origin of the information all mattered to the court. What reached the therapist there was the patient’s own threat, carried by his father — not the father’s independent worry that his son was dangerous. A family member’s belief about the client, without a reported threat from the client, isn’t what the case recognized. What it does mean is that you should be paying attention to what immediate family members involved in the client’s treatment tell you, including the things that arrive by voicemail at nine at night rather than inside the clinical hour.
“Reasonably identifiable” does more work than it looks like
A client who says he’s going to kill his ex-wife has named someone. A client who says he’s going to shoot his supervisor at Greenfield Elementary may likewise have identified the intended victim sufficiently, even without giving you a name.
Now take a client who says someone at his job is going to get what’s coming to them.
That’s menacing. It raises clinical alarm. It may keep you up. But it doesn’t identify a victim — it describes a group by location with no way to narrow it to a person. You’d do clinical work in the session to clarify, and if the client names someone or gives you enough to identify someone, the picture changes. If he doesn’t, you’re holding a clinically concerning statement that doesn’t meet the statutory threshold.
That distinction is where California practice actually lives. Most threat-assessment work happens in the gray space, not in the clean scenarios that exam questions favor. A client mutters something about a gun. A client describes a violent fantasy and insists he’s just venting. A client’s tone shifts and your stomach drops but the words themselves are ambiguous.
None of that automatically triggers the duty. All of it demands a clinical response — structured risk assessment, attention to history and access to means and substance use, adjustments to session frequency, coordination with other providers, a real conversation about voluntary hospitalization if it’s warranted.
Treating §43.92 as the ceiling of your responsibility rather than the floor is a mistake that can cost a life. The statute defines the narrow circumstance in which you lose immunity. It was never a description of competent care.
The authority to disclose comes from somewhere else
Here’s the misunderstanding that surprises people most.
Section §43.92 doesn’t immunize you against a client’s breach-of-confidentiality claim for having made the disclosure. Practitioners assume it does. The assumption is reasonable — you were told to notify, so surely you’re covered for notifying — but the text doesn’t reach that far.
Your authority to disclose sits elsewhere. Evidence Code §1024 removes the psychotherapist-patient privilege where disclosure is necessary to prevent a threatened danger. And Civil Code §56.10(c)(19) permits a psychotherapist to disclose medical information, consistent with applicable law and standards of ethical conduct, where the psychotherapist in good faith believes disclosure is necessary to prevent or lessen a serious and imminent threat to the health or safety of a reasonably foreseeable victim — and where the disclosure goes to someone reasonably able to prevent or lessen that threat, the target of the threat included.
Note the standards those two run on. Good faith belief, imminence, a recipient who can actually do something. None of that language appears in §43.92, which is one reason the sections don’t substitute for each other.
Knowing which authority does which job is not academic. If you’re ever asked to justify the disclosure — by a board investigator, by opposing counsel, by the client themselves — “§43.92 let me” is the wrong answer.
What the California duty to protect asks you at 4:45 on a Friday
Tarasoff and Ewing aren’t decoration. They’re the source of the duty and the leading interpretation of the statute’s key phrase, and a court would read §43.92 against both.
But when the session has run long and the client has just said something that changed the temperature of the room, the question that determines your exposure is the statutory one. Did the client — or an immediate family member relaying the client’s own words for the purpose of the client’s therapy — communicate a serious threat of physical violence against a reasonably identifiable victim?
If yes, the immunity for inaction is gone. You make reasonable efforts to communicate the threat to the victim and to a law enforcement agency.
If no, you’re in clinical territory. Risk assessment, safety planning, documentation. Real work, and often harder work. But you’re not in §43.92 territory, and the note you write that night needs to show that you knew the difference.
This is an excerpt from California Law and Ethics for Social Workers, available at SWTP CEUs . Looking for nationwide CEs? Most courses are designed for social workers across the U.S. Get them all for one low fee–priced for social workers.