Duty to warn and third-party risk — structured clinical viva

Do not invent statute numbers. If asked for a section, say you would verify the current local Mental Health Act criteria and organisational policy, and answer with principles: mental disorder, risk to others, least restrictive alternative, capacity interfaces, rights to review.[3][4]

Warn vs protect. Tarasoff I teaching: duty to warn identifiable victims. Tarasoff II: duty to protect via reasonable steps — warning is one tool among hospitalisation, police notification, means control, intensified treatment.[1][2][3]

Appelbaum model. (1) Assess dangerousness — named boss, bat in car, tomorrow, manic psychosis. (2) Select protective course — likely emergency containment plus disclosure. (3) Implement and document.[1]

Tonight's actions. Safety of clinic; prevent free departure if imminence high and lawful options exist; seek admission (voluntary if safe/possible; compulsory if criteria met); remove bat access via lawful help; treat mania/psychosis; notify police and warn boss with minimum necessary information; clear crisis plan. Police response varies — communicate actionably.[3][5]

Jurisdiction framing. ABPN: expect US state variation discussion (mandatory/permissive/limited duty). FRANZCP: ethics codes, privacy exceptions, local MHA — not California statute export. Multi-board answers score humility plus concrete clinical action.[4][2]

Alliance. Explain you cannot keep a secret that enables serious harm; care continues; warning may strain trust — repair explicitly.[6]

Diagnosis ≠ duty. Mania raises risk pathway here because of specific threat, means, and target — not because bipolar disorder automatically voids confidentiality.[7][1]