One of the most useful recent developments in psychiatric medication conversations is also one of the least dramatic: a 2026 expert consensus statement on deprescribing argued that medication decisions should include clear treatment aims, ongoing reassessment, and prespecified endpoints.
That sounds straightforward. It isn’t.
Most medication decisions are framed as a choice between continuing and stopping, or between staying the course and making a change. But those are actions. They are not goals.
Before a medication is continued, reduced, added, or replaced, the more important question is this: What would count as a better outcome, and how would we know?
Without an answer, almost any result can be interpreted after the fact. Feeling worse may be called relapse. Feeling better may be credited to the change, even if several other things shifted at once. A side effect may be treated as the unavoidable price of stability, or a temporary disruption may be treated as evidence that the medication was essential forever.
None of those conclusions follows automatically.
“Better” needs a job description
Medication is often started for a specific problem: panic, insomnia, depressed mood, agitation, intrusive thoughts, psychosis, mood instability, or something else. Over time, the original reason can become fuzzy.
The medication remains on the list, but its current job description is unclear.
Is it still helping the original symptom? Is it serving a different purpose now? Is it preventing a recurrence? Is it being continued because no one wants to disturb a fragile equilibrium? Is it being used partly to offset the effects of another medication?
These are different situations. They call for different reasoning.
A meaningful definition of success is more specific than “feel better” or “be stable.” It may include symptom intensity, sleep, ability to work, relationships, concentration, sexual functioning, appetite, weight or metabolic concerns, physical comfort, emotional range, and the practical burden of appointments, laboratory monitoring, cost, or managing multiple prescriptions.
Not every variable carries equal weight. That is the point. A person may reasonably place a high value on preventing a particular symptom from returning. Another may place equal weight on reducing a side effect that has been quietly eroding daily life. Neither priority can be inferred from a prescription record.
Improvement and tolerability are separate questions
I keep coming back to this distinction because it gets lost so easily: a medication can be beneficial and burdensome at the same time.
It can reduce one symptom while contributing to another problem. It can make a difficult period more manageable while becoming less clearly useful years later. It can be tolerated in the narrow sense that a person has not stopped it, while still carrying costs that deserve a real accounting.
The reverse is also true. A person can experience difficult effects during a medication change without that difficulty proving the original medication was harmful or unnecessary.
Side effects, adverse effects, withdrawal, rebound, return of prior symptoms, recurrence, and new symptoms can overlap. Similar symptoms do not settle the question of cause.
For example, anxiety or insomnia after a dose reduction might reflect withdrawal, rebound, recurrence of an earlier condition, a life event, another medication, or some combination. The timing matters. The symptom pattern matters. What existed before treatment matters. What changed during prior dose adjustments matters. Even then, certainty may not be available.
That is not a reason to ignore the experience. It is a reason not to force a simple explanation too early.
Why a timeline changes the quality of the decision
Medication decisions are often made from a snapshot: how someone feels this week, what symptom is most frustrating today, what a clinician remembers from the last visit.
A timeline is usually more informative.
It can reconstruct what was happening before the medication began, what changed after starting, what happened after dose increases or other additions, which symptoms improved, which problems emerged later, and what occurred during any previous attempts to change treatment.
Temporal association does not prove causation. But it is still evidence. Dismissing chronology because it cannot provide proof is not rigorous. Neither is treating chronology as proof when competing explanations remain plausible.
The goal is not to win an argument about the medication. The goal is to make the next decision less blind.
Set the observation window before interpreting the result
Another overlooked variable is time.
Some medication effects emerge quickly. Others develop gradually. Some withdrawal symptoms occur soon after a change, while others may persist well beyond the period when a drug has largely left the bloodstream. That does not automatically establish protracted withdrawal, but drug elimination and physiological adaptation are not interchangeable concepts.
Likewise, a short stretch of feeling well does not necessarily establish that a change was successful, and an early rough stretch does not necessarily establish that it failed.
This is where prespecified endpoints matter. What changes would be watched? Which outcomes would matter most? What degree of change would be considered meaningful? What uncertainty would remain even after observing the result?
A 2026 American Society of Clinical Psychopharmacology task force reached consensus that psychotropic medications should be periodically reassessed and that the impact of changes on functioning should be part of the discussion. It is important to read this correctly: consensus is not the same thing as definitive outcome evidence. But it identifies a gap in ordinary practice. Medication review is too often administrative when it needs to be interpretive. ([jamanetwork.com](https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2845497))
The real question is not always “continue or stop?”
Sometimes the decision is whether the current medication still has a clearly defined benefit. Sometimes it is whether the burden has changed. Sometimes it is whether a symptom has been incorrectly assigned to the original diagnosis, to withdrawal, or to a medication effect. Sometimes it is whether several medications are obscuring the picture.
And sometimes the honest answer is that the evidence is incomplete.
That answer is not a failure of judgment. It may be the beginning of better judgment.
A medication decision becomes more credible when it can name its aims, its tradeoffs, its time horizon, and its uncertainties. Before changing the treatment, define the question the change is meant to answer.
Written for Dr. Teralyn Sell, PhD
Psychology · Brain Health · Human Behavior
Sources worth knowing
- Recommendations for the Deprescribing of Psychotropic Medications: A Consensus Statement From the American Society of Clinical Psychopharmacology Task Force
- Comparison of antidepressant deprescribing strategies in individuals with clinically remitted depression: a systematic review and network meta-analysis