Therapy termination: when to end and how to document closure

Everything gets written about the first session: how to run the intake interview, what to ask, what to document. About the last one, almost nothing. And what does circulate splits into two halves that never talk to each other: pieces about the emotional goodbye that never mention the record, and “discharge summary” templates borrowed from inpatient paperwork. This guide joins the halves: when and how to end therapy according to ethics and evidence, and what actually needs to land in the chart when a process closes.

When to end: the three triggers of Standard 10.10

The ethical anchor is Standard 10.10 of the APA Ethics Code. Subsection (a) is uncomfortably clear: psychologists terminate therapy when it becomes reasonably clear that the client no longer needs the service, is not likely to benefit, or is being harmed by continued service. Three triggers; only the first is the happy ending.

No longer needs the service. Termination by goals met is not decided in the moment: it was written the day you drafted the treatment plan. With verifiable goals, “when” becomes a reading of the plan; and if you measure progress session by session, the closing conversation rests on a curve, not an impression.

Not likely to benefit. The stalled process, with no movement in goals or measures. The honest answer is not always more sessions: sometimes it is a change of approach, sometimes a referral.

Being harmed by continued service. The least comfortable trigger to admit: continuing to “accompany” can also be a way of holding on.

Two more subsections complete the standard. Subsection (b) expressly permits ending therapy when the client — or someone in the client’s life — threatens or endangers you. And subsection (c) sets the default duty before any ending: except where precluded by the actions of the client or of third-party payors, you provide pretermination counseling and suggest alternative providers as appropriate. Two neighboring standards round out the map: 3.12 asks you to plan from the start for interruptions — your illness, relocation, or retirement, or the client’s relocation or financial limitations — and 6.04(d) asks you to discuss foreseeable limits due to financing as early as feasible. The ending is not improvised; it is part of the frame.

A note on vocabulary: “discharge” is largely inpatient language, and an outpatient record-keeping framework may define no dedicated termination document at all — an absence that, as you will see below, is precisely what tells you how to document the closure.

Termination planning starts at intake

A 2025 systematic review of 67 articles on psychotherapy termination documents that in CBT the conversation about the ending begins at the outset of therapy, tied to goals and anticipated duration. It also reports that endings agreed upon or initiated by the therapist are associated with higher satisfaction and better outcomes than unilateral dropout, and that client involvement in the decision correlates with perceiving the therapy as successful; endings that go badly show the opposite pattern — lack of closure and no emotional discussion. For severe diagnoses, especially depression, the review describes the need for a more structured process: reviewing gains, setting goals, and giving advance notice.

One reassuring finding for anyone postponing the conversation: studies find more positive than negative emotions at termination. Clients who experience the ending as an achievement report pride, independence, and relief; sorrow, self-doubt, and frustration appear mostly when the client did not feel ready. One more argument for agreeing on the moment rather than announcing it.

Working the closing phase

Manualized therapies provide the script. In interpersonal psychotherapy, the final sessions have defined tasks: reminding the client that the end is near, reviewing their accomplishments — often considerable — to reinforce a sense of capability and independence, and framing the ending itself as a role transition with both good and painful aspects. The same model offers an alternative that works in any approach: for recurrent conditions, recontracting for maintenance — spaced sessions, monthly for example — instead of a hard stop.

The other task of the final phase is relapse prevention, and the numbers justify the time. A meta-analysis of 28 studies with 1,880 adults found that after discontinuing acute-phase cognitive therapy for depression, 29 percent of responders relapse or recur within one year and 54 percent within two; a continuation phase reduced relapse-recurrence by 21 percent at its end and 29 percent at follow-up. Consolidation is not ceremonial: the last sessions are for naming warning signs, rehearsing the response to each one, and putting that plan in writing — in the client’s copy and in your record.

“Session 18 of 20. Reviewed the list of gains M. brought as homework: back to exercising, holding boundaries at work, crying spells down from daily to occasional. Rehearsed the relapse plan: three warning signs and the response to each. She reports feeling proud and a little nervous about ending; we normalized it.”

Documenting closure: the final progress note

Here is the part most templates get wrong: they assume a dedicated discharge document is required everywhere. Check what your own framework actually asks for — often the closure simply rides on the note you already write every session. Mexico’s clinical-record norm (NOM-004-SSA3-2012) is a usefully explicit example: its outpatient chapter lists exactly four documents — clinical history, progress note, interconsultation note, and referral note — and no discharge note of any kind; the “nota de egreso” that vendor templates recycle sits in the hospitalization chapter. The closing note, in other words, is legally the final progress note — written in the knowledge that it is the last one.

The record still matters after the goodbye, and US guidance says why. The APA’s Record Keeping Guidelines note that records document treatment plans, services, and progress, help protect both client and psychologist in legal or ethical proceedings, and are generally required for third-party reimbursement. On retention, the same guidelines suggest — explicitly “in the absence of a superseding requirement” — keeping full records until 7 years after the last date of service for adults, or until 3 years after a minor reaches the age of majority, whichever is later. Notice what that makes of your closing note: it marks the last date of service, so it is the entry that starts whatever retention clock applies where you practice (in Mexico, for example, a 5-year minimum counted from the last act of care). Retention rules and required note content vary by jurisdiction and licensing board — verify yours.

Two documentary companions are worth knowing. If the client wants something to take home, the cleaner instrument is a clinical summary issued on written request, not an improvised “certificate of completion.” And if the ending is a transfer, document it as a referral: receiving provider, reason for referral, diagnostic impression, and treatment provided — the documentary mirror of 10.10(c)’s duty to suggest alternatives.

What goes in the closing note

No special format is required, but the last note should not read like all the others. A solid closing entry covers:

Date and reason for closure. Goals met, client decision, or referral. Without a recorded reason, the chart ends in an ellipsis.
Course summary and current status. Where the process started and where it ends; with initial and final scores if you measured with instruments.
Diagnoses or clinical problems at closure — and what remains open. Unfinished business does not disappear by going unwritten.
Prognosis. Most progress-note structures already ask for it; in the final entry it carries more weight.
Post-therapy plan. Agreed warning signs, the relapse plan, and the conditions for returning.
Referral details, if any.
The formal basics your rules require: date, time, full name, signature.

“Process concluded by mutual agreement upon completion of treatment-plan goals. Course: sustained remission of anxiety symptoms for eight weeks; resumed work and social activities. No open clinical problems. Prognosis favorable. Plan: warning signs and relapse plan reviewed; follow-up available at the client’s request. Clinical summary provided upon written request.”

When the ending is not the one you planned

Not every closure is agreed: sometimes the client simply stops coming. The literature defines dropout as termination without fulfillment of the therapeutic goals and without the benefit a normal ending would have allowed. The numbers are less catastrophic than the ones usually repeated: a classic meta-analysis of 125 studies put dropout near 47 percent, but Swift and Greenberg’s 2012 meta-analysis — as reported in a 2022 open-access study — placed it at roughly 19.7 percent on average, ranging from 0 to 74 percent across studies. The same study documents something rarely said aloud: dropout hits the therapist too, with self-doubt as the most frequent emotional consequence.

Ethical precision helps here: the APA Code never uses the word “abandonment.” Its demand is framed in the positive — orderly resolution of responsibility for the client’s care when the relationship ends (10.09), advance planning for interruptions (3.12), and pretermination counseling with alternatives except where the client’s own actions preclude it (10.10c). Licensing boards and local law may add expectations of their own, so check the rules where you practice. If the client does not return, those duties are exactly what you document: your contact attempts, what you offered — a closing session, a referral, an open door — and a final entry that formally closes the process even though the goodbye never happened. It helps to fix a personal criterion in advance (how many attempts, through which channel, over how long) and apply it consistently.

One case stands apart: an interruption with risk signals present — something said in the last session, a worrying message — is not a filing matter but one of immediate clinical assessment under your own protocol and training, documented as such. It is one more reason crisis and emergency contacts belong in the frame from the informed consent onward, so every client knows where to turn locally if a crisis arrives between — or after — sessions.

The open door

A good ending includes an explicit return clause, spoken and written into the last note: which signs would make coming back sensible, and that coming back is not failure. Two record-keeping facts hold that door open. The chart is retained for years after the last date of service, so a return does not start from zero — it starts from a documented history. And your confidentiality duties do not leave with the client: Standard 6.02 requires maintaining confidentiality in storing, transferring, and disposing of records, and in Mexico, for example, the data-protection law in force states expressly that the confidentiality duty survives the end of the relationship.

Closing well is a process problem and a chart problem at once, and the second one becomes easy when the chart arrived at the ending in order. gesell.ai keeps a structured chart for each client — intake, clinical history, treatment plan, and session notes — so the full arc of the process is in view when termination comes, and it drafts every note, including the last one, from your session record: you review, adjust, and sign. The termination decision will always be clinical and yours; making it effortless to document is exactly the point.