Shared Decision-Making in Psychotropic Tapering: Clinical Frameworks

Shared decision-making (SDM) in psychotropic deprescribing is no longer optional. The Maudsley Deprescribing Guidelines (Horowitz & Taylor, 2024), NICE NG215 (2022) on safe medicine withdrawal, and the Royal College of Psychiatrists' 2020 position statement all converge on the same point: tapering decisions belong to the patient, supported by a prescriber who frames risks, options, and personal priorities clearly. This article outlines a structured SDM framework adaptable to psychiatric and primary care settings, with suggested patient language the clinician can lift directly into the visit.

Why SDM is load-bearing in deprescribing

Initiation of a psychotropic is often a low-friction decision: symptoms are present, the patient is distressed, and the prescriber has a near-term outcome to optimize. Discontinuation is the inverse. The patient is typically stable on the medication, the trigger for tapering is rarely acute, and the time horizon for harm (relapse) and benefit (resolution of long-term adverse effects, autonomy) extends over months or years. Standard informed consent, designed for a single discrete intervention, fits poorly. SDM — explicitly modeled on the three-talk framework of Elwyn et al. (BMJ, 2017) — handles these features better because it treats the decision as iterative.

Two empirical observations reinforce the point. First, Read & Williams (2018) found in a survey of more than 1,800 antidepressant users that 56% reported withdrawal effects, of which roughly half rated them severe — yet fewer than 2% recalled being warned at initiation. Second, the PANDA trial (Kendrick et al., NEJM, 2019) and the ANTLER trial (Lewis et al., NEJM, 2021) both demonstrated meaningful relapse risk on antidepressant discontinuation, with ANTLER reporting 56% relapse at 52 weeks in the discontinuation arm versus 39% in the maintenance arm. The information asymmetry at initiation and the genuine relapse risk at discontinuation together create an unusual decisional load: the patient is being asked to weigh risks they were not previously informed of, against benefits that are individual and difficult to predict.

A structured SDM framework converts this load into a series of discrete, documentable decisions rather than a single global one.

The three-talk model adapted for tapering

Elwyn's three-talk model — team talk, option talk, decision talk — maps cleanly onto deprescribing if the prescriber resists the temptation to compress it into one visit.

Team talk

The opening move is to establish that a decision exists. Many patients on long-term psychotropics have never been told that discontinuation is a legitimate option, or have absorbed the message that the medication is "for life." Naming the decision is the intervention.

Suggested patient language:

"You've been on sertraline for about six years. The reason we started it has changed, and there are now better data on what long-term use looks like and what coming off looks like. I want to walk you through what we know, hear what matters to you, and decide together whether to continue, taper, or pause that decision for now. There is no urgency today — we are just opening the question."

Three points are worth flagging in team talk: (1) the current dose and duration, (2) the original indication and whether it still applies, and (3) the explicit statement that staying on the medication is a valid choice. Patients who feel that the prescriber has a preferred answer disengage from the rest of the conversation.

Option talk

Option talk requires the prescriber to enumerate plausible courses of action and their associated risks. For most psychotropics, the realistic option set is:

1. Continue at the current dose with periodic review.
2. Continue at the current dose with adjunctive non-pharmacological treatment (CBT, exercise, sleep optimization) and revisit tapering in 6–12 months.
3. Begin a gradual taper with monitoring.
4. Switch to a longer-half-life agent (e.g., for short-half-life SSRIs and benzodiazepines) before tapering.
5. Cross-taper to a different agent if the original is poorly tolerated and discontinuation is not the primary goal.

Each option should be paired with the principal risks the patient cares about: relapse, withdrawal, adverse effects of continued use, and the practical demands of the taper itself (visit frequency, formulation changes, the time horizon). The Maudsley Deprescribing Guidelines emphasize that gradual hyperbolic tapering — slower at lower doses, reflecting the non-linear receptor occupancy curve described by Sørensen et al. (Lancet Psychiatry, 2022) — substantially reduces withdrawal severity compared with linear tapers, and option talk is the right place to introduce this pharmacological reality without yet committing to a schedule.

A common pitfall in option talk is anchoring on a number. Quoting a specific reduction percentage before the patient has expressed values pre-empts the decision. The schedule belongs to decision talk, after the patient has indicated which risks they weight most heavily.

Decision talk

Decision talk is where the prescriber elicits values and translates them into a plan. A useful structure is to ask the patient to rank, in their own words:

• The cost of relapse (lost time, occupational impact, prior episode severity).
• The cost of withdrawal (duration tolerable, life events on the horizon, support available).
• The cost of continuing the medication (sexual dysfunction, weight, emotional blunting, fatigue, drug-drug interactions, autonomy concerns).
• The cost of indecision (cognitive load, ambivalence as itself a stressor).

Patients who weight relapse cost highest typically want a slower taper with longer hold periods and clear stop-rules. Patients who weight ongoing adverse effects highest typically want a more proactive schedule and accept higher withdrawal risk. Patients who weight autonomy highest often want a written plan they can execute with minimal visits. The schedule, monitoring frequency, and rescue plan flow from this ranking — not from a one-size-fits-all guideline.

Eliciting values without leading

The hardest skill in SDM is asking values questions that do not embed an answer. Compare:

• Leading: "You don't want to risk a relapse, do you?"
• Neutral: "When you think about the period after we last changed this medication, what was the part you most wanted to avoid happening again?"

The neutral form invites the patient to specify what they are protecting against, which may not be relapse — it may be the side effect of the index episode's treatment, the loss of work, or the experience of involuntary admission. Each of these maps to a different optimal taper.

The Ottawa Personal Decision Guide and the SURE test (Sure of myself, Understand information, Risk-benefit ratio, Encouragement) are validated, brief instruments that can be completed in 2–3 minutes. The SURE test in particular flags decisional conflict before the patient has committed to a course they will later abandon. A SURE score below 4 of 4 is a signal to extend team talk and option talk before moving to a plan.

Documenting the SDM conversation

Documentation serves three purposes: medico-legal protection, continuity for covering clinicians, and a written record the patient can refer to between visits. A defensible note for a tapering decision should include:

• The indication for the original prescription and its current status.
• The options discussed (including continuation).
• The risks named for each option, in plain terms.
• The patient's stated values and which risk they weighted most heavily.
• The plan agreed, including the first reduction, the monitoring interval, and the explicit criterion for pausing or reversing the taper.
• The patient's contact route for emergent symptoms.

A template paragraph that satisfies most of this in 4–5 sentences:

"Discussed continuation, gradual taper, and switch to a longer-half-life agent. Reviewed relapse risk citing ANTLER trial data, withdrawal profile citing Maudsley Deprescribing Guidelines, and adverse effects of continued use specific to this patient (sexual dysfunction, emotional blunting). Patient prioritized resolution of sexual dysfunction over relapse risk and chose gradual taper with hyperbolic dose reductions. Plan: reduce at the next clinical review with patient-initiated contact for emergent symptoms; criteria to pause are specified in the patient-facing taper letter. Patient verbalized understanding and consented to the plan."

The note is short, names the evidence cited, names the values elicited, and specifies stop-rules. This is the standard the General Medical Council's Decision Making and Consent (2020) effectively requires for high-stakes elective decisions.

Common decisional traps

The "stable on it" anchor

Patients and clinicians both interpret stability on a psychotropic as evidence the drug is responsible. For most antidepressants and benzodiazepines used for more than 12 months, this is a testable rather than settled hypothesis. Naming the anchor explicitly during team talk — "stability while on the medication doesn't tell us whether you'd be stable off it; it tells us the current combination is working" — opens the option set.

The asymmetric framing of risk

Relapse is salient and recent in patient memory; chronic adverse effects are diffuse and have been adapted to. Without explicit prompting, the patient will overweight relapse. Asking the patient to enumerate the costs of staying on the medication, in detail, restores symmetry.

Premature commitment

The patient who arrives saying "I want off everything" is often making an emotionally driven decision that will not survive the first withdrawal symptom. The patient who arrives saying "I will never come off" is often deferring to a previous prescriber's framing rather than expressing a current value. SDM requires the prescriber to slow both patients down and revisit team talk.

The proxy decision-maker problem

Family members, particularly partners and parents, frequently have stronger preferences about the patient's medication than the patient does. SDM is between prescriber and patient. Family input is data; it is not the decision. Documenting whose values drove the plan is part of defensible practice.

Special populations

Patients with prior severe episodes

Patients with a history of suicide attempt, psychotic features, or involuntary admission deserve a more conservative default and a longer team talk. SDM is not consent-by-default to whatever the patient prefers; it is a structured negotiation in which the prescriber retains the duty to flag asymmetric risk. The Maudsley Guidelines recommend that for patients with two or more severe depressive episodes, maintenance treatment is generally favored, and the SDM conversation should make this evidence visible rather than presenting all options as equally weighted.

Polypharmacy

Patients on three or more psychotropics rarely benefit from simultaneous tapering of more than one agent. SDM in this setting includes deciding the order of deprescribing. The general principle, supported by the BNF and Maudsley Guidelines, is to taper the agent with the worst adverse-effect profile or weakest current indication first, and to allow stabilization before the next change. The patient's input is essential because the order changes which adverse effects resolve first, and patients prioritize differently.

Pregnancy and the perinatal period

The decision frame shifts when teratogenic risk, postpartum relapse risk, and breastfeeding considerations enter. SDM remains the appropriate framework, but option talk must include the obstetric and pediatric data the patient may not have seen. Consultation with perinatal psychiatry is generally indicated; the decision is not one a generalist should finalize without specialist input.

Cognitive impairment

When capacity to weigh information is impaired, SDM does not become unilateral prescribing. It becomes shared decision-making with a substitute decision-maker, and the prescriber's documentation should name the substitute, the basis for capacity assessment, and the patient's expressed wishes where elicitable.

When to defer the decision

Not every tapering conversation should end with a plan. Acceptable outcomes of an SDM visit include continuing the medication, beginning a taper, switching agents, and explicitly deferring the decision to a named future date. Deferral is a decision; it should be documented as such, with the trigger for revisiting (a life event, a planned pregnancy, an annual review).

Premature plans generate failed tapers, and failed tapers reinforce the patient's belief that the medication is required indefinitely. A patient who defers, returns in six months, and then begins a confident taper has a better prognosis than a patient who begins ambivalently and abandons the attempt at the first withdrawal symptom.

Clinical pearls

• Open every long-term psychotropic review by naming the decision: continuation, taper, and deferral are all legitimate outcomes, and the patient should know this before option talk begins.
• Quote evidence by name in the visit. "Maudsley Deprescribing Guidelines," "ANTLER trial," and "Horowitz & Taylor" are short enough to say aloud and substantially shift the patient's perception of the conversation's rigor.
• Elicit the patient's values before naming a schedule. Anchoring on a percentage or interval pre-empts the decision and produces compliance, not consent.
• Use a brief validated instrument (SURE test or Ottawa Personal Decision Guide) when decisional conflict is suspected. A score below threshold is a signal to extend team talk, not to push for a plan.
• Document the decision in the structure: indication, options discussed, risks named, values elicited, plan with stop-rules. Four or five sentences suffice and meet GMC consent standards.
• Treat deferral as a legitimate outcome. A confident taper begun six months later outperforms an ambivalent taper begun today.
• Revisit the SDM conversation at every dose reduction. Tapering is not a single decision; it is a sequence of small ones, and the patient's values may shift as withdrawal symptoms or quality-of-life changes emerge.

References

• Elwyn G, Durand MA, Song J, et al. A three-talk model for shared decision making: multistage consultation process. BMJ. 2017;359:j4891.
• Horowitz MA, Taylor D. The Maudsley Deprescribing Guidelines: Antidepressants, Benzodiazepines, Gabapentinoids and Z-drugs. Wiley-Blackwell; 2024.
• Lewis G, Marston L, Duffy L, et al. Maintenance or discontinuation of antidepressants in primary care. N Engl J Med. 2021;385(14):1257–1267 (ANTLER trial).
• Kendrick T, Stuart B, Leydon GM, et al. Patient-reported outcome measures and observed symptoms in patients prescribed long-term antidepressants. NEJM. 2019.
• NICE NG215. Medicines associated with dependence or withdrawal symptoms: safe prescribing and withdrawal management for adults. 2022.
• Read J, Williams J. Adverse effects of antidepressants reported by 1,431 people from 38 countries. Curr Drug Saf. 2018;13(3):176–186.
• Sørensen A, Ruhé HG, Munkholm K. The relationship between dose and serotonin transporter occupancy of antidepressants — a systematic review. Lancet Psychiatry. 2022.
• General Medical Council. Decision Making and Consent. 2020.

---

For more clinician resources on safe deprescribing and tapering, visit tapermeds.com.