# Worse after a treatment change: withdrawal, a returning condition, or another cause?

> “All events seem entirely loose and separate.”

David Hume, An Enquiry Concerning Human Understanding, VII, Part II, §26. [Text of §§26–28](https://davidhume.org/texts/e/7). Hume investigates how the idea of causal connection arises. This sentence neither denies medical mechanisms nor means that causes cannot be investigated.

## What this article is about

Someone may feel worse after a change in treatment. They might think, “I can no longer cope without the medicine” or “The illness has returned.” But feeling similar does not establish the cause. Assessment needs the prescribing history, what actually changed and what is happening now. Help can begin before there is a precise explanation.

## I. First consider how urgently help is needed

An immediate threat to life or risk of self-harm calls for urgent local help. Do not wait until it becomes clear whether the condition relates to withdrawal, the previous illness or another cause. Discussing medication and preparing notes can wait. [NIMH, Depression](https://www.nimh.nih.gov/health/publications/depression).

If suicidal thoughts appear, tell a professional immediately; danger or an inability to keep yourself safe requires a local emergency route. The NHS also warns that some serious adverse reactions require urgent medical help. This article does not list every such reaction or replace individual assessment. [NHS, antidepressants](https://www.nhs.uk/medicines/antidepressants/).

When there is no immediate danger, you can contact the prescriber and say, “I have felt worse since the changes; I need to discuss possible causes and further help.” That contact does not first require proving withdrawal, accepting an addiction diagnosis or compiling a perfect diary.

## II. Similar feelings can have different explanations

Withdrawal symptoms arise in connection with reducing or stopping a particular medicine, sometimes with missed doses. A returning illness means a recurrence of the condition being treated. A side effect or another factor is a further possibility that cannot be excluded simply because treatment recently changed.

For example, the NHS describes sleep problems, anxiety, mood changes and physical sensations during antidepressant withdrawal. Some complaints may resemble the previous condition. Symptoms can be severe, begin later than expected and persist for a long time. This does not permit every deterioration to be labelled withdrawal from a distance, but the absence of an immediate reaction does not exclude it either.

“This feeling is new to me” is important information, rather than ready-made proof of a new illness. “It happened before” does not establish recurrence either. A person's observations deserve attention even when there is no final label yet. Describing a concrete change to a professional is more useful than arguing about labels.

## III. How the history helps the conversation

The Royal College of Psychiatrists' information on stopping antidepressants discusses when complaints begin and how sensations differ from those present before treatment. This information supports a conversation with the prescriber. It does not form an independent test: a single time interval or sensation cannot establish the cause. [RCPsych, stopping antidepressants](https://www.rcpsych.ac.uk/mental-health/treatments-and-wellbeing/stopping-antidepressants).

You can recall why treatment was prescribed, which changes were agreed and what actually happened. If information is incomplete, say so. There is no need to create a new missed dose, stop or restart in order to produce a more convincing history. The body's response to a home experiment does not replace medical assessment.

“Before” and “after” help reconstruct the sequence. Establishing cause takes more: a professional may need to assess other prescriptions and the current condition. Someone does not have to work out every interaction on their own before seeking help. It matters to preserve information about both helpful treatment effects and difficulties.

## IV. Different medicines need different assessment

A shared question about possible withdrawal does not mean that every medicine carries the same risk. The FDA's benzodiazepine warning states that physical dependence can occur even with prescribed use, and abrupt stopping or overly rapid reduction can cause dangerous reactions, including seizures. This is an example concerning a particular class, rather than a prediction for every psychiatric medicine. [FDA, benzodiazepine warning](https://www.fda.gov/drugs/drug-safety-and-availability/fda-requiring-boxed-warning-updated-improve-safe-use-benzodiazepine-drug-class).

Seizures or another life-threatening condition require immediate local emergency help. A scheduled routine appointment does not make waiting safe. Do not look for another substance at home to “compensate” for what is happening.

This article provides no withdrawal regimen, list of substitutes or calendar for several medicines. Changes to a prescribed medicine should be discussed with the responsible professional. If an unplanned change has already occurred, report it; shame and fear of judgement should not become an additional barrier to help.

## V. Preserve treatment benefit and the opportunity for review

Real relief, better sleep or the ability to work matter. They do not establish that a prescription is suitable forever, but difficulty stopping does not erase earlier benefits. RCPsych's information distinguishes physical adaptation from addiction; an individual discussion considers why treatment was prescribed, its benefits and emerging difficulties.

There is no need to prove independence by giving up medication. Both wanting to continue useful treatment and wanting to discuss stopping deserve a conversation. Duration alone does not establish a substance use disorder, weakness or an intention by the system to retain a patient. Assessment needs specific circumstances.

Do not stop, restart or change treatment independently to test the cause of symptoms. This does not mean enduring every reaction or automatically continuing a dangerous prescription: serious deterioration calls for timely medical assessment, with decisions about changes based on the person's condition. Love, meaning and a new environment may support them without replacing necessary help.

## VI. Five optional fields for contacting help

These are the author's preparation prompts for a conversation, rather than a diagnostic scale or treatment plan. Use only what is already known, without carrying out a new experiment. Any field can be left blank; notes are unnecessary during danger.

1. **Prescription and purpose.** What was prescribed, and for which complaint or condition, if known.
2. **Actual change.** What has already changed in treatment, what was agreed and where information is uncertain. Do not create a change to observe it.
3. **Previous and new feelings.** What resembles the condition before treatment, what differs and when it became noticeable, as far as you remember.
4. **Benefits, difficulties and other help.** What helped, how daily life is changing now and which other prescriptions or substances should be discussed.
5. **An accessible contact.** Whom you can contact, which access conditions need clarification and what voluntary help from a relative is possible.

One sentence is acceptable in place of a detailed report: “Difficulties I do not understand appeared after the changes.” Seeking help does not depend on complete notes, perfect motivation or every relative agreeing.

## VII. Agree on a clear next step

For a routine contact, clarify who can assess the current condition and prescribing history. It is important to distinguish medical assessment from a supportive conversation: booking a counsellor alone does not resolve a dangerous reaction or withdrawal.

The author suggests clarifying appointment arrangements and barriers beforehand. You can ask about language, cost, contact format and where to turn if that service is unsuitable. Confidentiality and the sharing of medical information should be clarified with the particular service; conditions differ between countries.

At the appointment, it helps to ask what has been established, what remains unknown, who continues care and how to make contact if things worsen. An individual plan may need review. Asking for clarity does not mean that the patient should calculate the dose, pace or stopping point independently.

## VIII. Relatives help within agreed boundaries

With an adult's consent, a relative can help recall information already known, arrange interpretation, transport or company. They are not obliged to identify the cause of symptoms, choose medication or become solely responsible for treatment. Manageable limits and their own support matter.

Secret checks, withholding medication and demanding proof of “independence” do not replace an agreement about help. A child must not be tasked with monitoring an adult's medicine use, mood or safety. Where there are threats or violence, protecting the person harmed and children matters in its own right; involvement does not require enduring harm.

After making contact, you can discuss whether an actual assessment was received and the next route is clear. A new booking or a new label for the condition does not yet establish that help was provided. Someone is entitled to ask for explanations, report benefits and difficulties together, and keep seeking suitable support.

## IX. Telling a history without turning it into an accusation

Consider a fictional situation. Dasha remembers sleeping worse after a prescription change. She cannot recall the exact onset date and has already heard both withdrawal and relapse from different people. She thinks seeking help requires choosing and defending one explanation. A useful history works differently: known events first, present difficulty next, and hypotheses separately. She can say that sleep changed after a particular event, that the date is approximate, and that the cause is unknown to her. This is valid information even without a diagnosis at the end.

Distinguish what was planned from what happened. An agreed prescription change and actual use sometimes differ: instructions were unclear, a package ran out, part of the conversation was forgotten, or the medicine could not be obtained. Those facts matter for safety, not for assigning shame. State the difference directly: we discussed one thing, but something else happened in practice. The professional needs actual use; a flawless account of compliance may conceal precisely the detail assessment requires.

Accuracy does not require the exact minute. If only the week is known, say so. Distinguish remembering, thinking something probably happened, and having a record. Packaging, dispensing information, or an earlier prescription may help reconstruct details, but missing documents do not erase an experience. Do not add a confident date to sound convincing. Clearly marked uncertainty is more useful than a tidy history whose gaps have been filled with guesses. Nor should medical contact wait until every document has been found.

The next layer describes what is happening now. Feeling anxious can involve different difficulties: not sleeping, finding it hard to leave home, or an unfamiliar bodily sensation. One observable example of the impact on life can be enough. Medical vocabulary and repeated checking are not required. If monitoring becomes distressing, limit the note and seek help with what is already known. The history is meant to communicate information, not create continuous surveillance of every bodily movement.

Useful details are especially easy to lose when several circumstances changed together. Another illness, a new product, or a bereavement may have occurred that week. Reporting this does not label every symptom as merely stress. Mentioning a medication change likewise does not establish it as the only cause. Several facts can coexist in a history. Their significance belongs to assessment, not a contest between psychological and bodily explanations.

If a ready-made conclusion interrupts her, Dasha can return to a specific question: I understand your hypothesis; please also consider that this sensation is new to me. This guarantees neither agreement nor the correctness of her own explanation. It separates being heard from demanding a predetermined answer. If an explanation is unclear, ask what information supports it and what cannot yet be ruled out. Another professional view can be discussed where accessible, without portraying access as easy for everyone.

The result is modest but important: the professional receives a less distorted sequence, and the person no longer needs to present a hypothesis as an entry requirement for care. If several causes coexist, the account was not wasted. If the explanation remains open, it can still guide further assessment. A history should not become proof of a patient's moral innocence: people who became confused or had an unplanned change need care too.

## X. What different traditions allow us not to know

Uncertainty can feel like a personal failing: if I live in this body, should I not understand its response? Philosophical texts can examine the expectation itself. Knowing that something hurts or feels frightening differs from knowing why. A person has access to their experience without becoming an instrument that directly displays its mechanism. Conversely, no mechanism in their account does not make the experience unreal. Preserve both sides without giving all authority either to a confident guess or to a dismissive response.

In Section VII of the Enquiry, Hume distinguishes observed succession from necessary connection, then considers repetition and the habit of expecting one event after another. Here this highlights a leap from after to because of. His theory does not cancel clinical evidence or prove a symptom's cause unknowable. A personal episode, a research generalisation, and medical assessment offer different grounds. His text gives no reason to repeat a dangerous episode for personal proof. [Hume, VII.26–28](https://davidhume.org/texts/e/7).

In chapter I of How We Think, Dewey connects doubt with inquiry into grounds. The emphasis differs: the unknown can become a question rather than a verdict. For contact with care, distinguish what was noticed, what is suspected, and what needs clarification. This need not require endless internal debate; inquiry can mean taking a question to someone qualified to assess it. Dewey's educational work is not a medication-testing protocol, and patients do not have to collect experimental proof before receiving help. [Dewey, I, §§2–4](https://www.gutenberg.org/files/37423/37423-h/37423-h.htm).

At the beginning of the Nicomachean Ethics, Aristotle cautions against expecting the same precision in every subject. This helps challenge false precision in a history: an approximate date should not masquerade as measurement. Respect for limits does not license arbitrary explanations. Where verifiable medical information exists, that information is needed. His ancient ethical discussion does not determine modern diagnostic standards; it raises the question of fitting a claim to its grounds. [Aristotle, I.3](https://classics.mit.edu/Aristotle/nicomachaen.1.i.html).

In the Kalama Sutta, the Buddha's listeners are confused by conflicting teachers. The response does not recommend relying solely on authority, reports, or a pleasing thought; it discusses the consequences of qualities and the judgment of the wise. The limited comparison here is permission to ask what grounds a claim. This is a religious and ethical conversation about skillful and unskillful qualities, not a clinical trial. It cannot become a requirement to stop medication and test the result. Symptoms are not evidence of a person's greed, hatred, or delusion. [AN 3.65, Thanissaro Bhikkhu translation](https://accesstoinsight.org/tipitaka/an/an03/an03.065.than.html).

In Job 2:11–13, friends come to support a suffering person and initially sit beside him without an explanation. The short episode suggests that presence is possible before a theory. Seven days of silence is a detail of the biblical narrative, not a waiting period before medical help. Support can include both listening and urgent contact when needed. This specific passage does not make everything the friends later do an example, or explain illness as a test of faith. [Job 2:11–13, KJV](https://www.biblegateway.com/passage/?search=Job%202%3A11-13&version=KJV).

These approaches differ: Hume considers causal connection, Dewey inquiry, Aristotle appropriate precision, the sutta ethical discernment, and Job an encounter with suffering. Together they provide no medical answer. They help avoid filling uncertainty with blame or demanding unattainable certainty. Medical decisions still need appropriate evidence. Respectful listening does not make every hypothesis correct; accurate assessment does not require taking away a person's voice.

## XI. An unknown cause does not remove the next step

Two questions differ: what explains the symptoms, and what help is needed now? Sometimes the first remains unanswered while the second already calls for action. The opening states this for immediate danger. A less acute situation can also have an understandable intermediate outcome: the professional clarifies the history, arranges further assessment within their competence, and explains contact arrangements. No final label does not mean leaving someone alone until proof appears.

A fictional counterexample to Dasha is Pavel, who is sure the new difficulty is withdrawal because it followed a change. Certainty does not answer whether another condition needs separate assessment. This article cannot determine that from the account. His task is to report the coincidence and current severity, not rule out everything else. The reverse error is equally possible: calling a complaint recurrence simply because a diagnosis already appears in the record. A new observation must remain visible in either case.

A transfer between services presents another difficulty. One professional discussed the change; another sees the person afterward, before documents arrive. Missing information should not require patients to decide which explanation wins. They can report the known plan, actual use, and current symptoms while identifying what is missing. Ask who will request further information and who is responsible for current assessment. Sharing and consent arrangements depend on the service and country; this article promises no universal immediate transfer of every record.

A complete history may be impossible when energy is low. Begin with the present difficulty and a known change. With consent, a trusted helper can read a short note or assist with language. They should not silently replace the person's account with their own causal theory. Having no safe helper does not reduce the human validity of seeking care: need does not depend on family availability. Particular access conditions still need clarification, and their absence cannot be remedied by demanding that someone try harder.

A relative may not know the cause either. A useful role can be describing an observation and offering support: I noticed things have become harder; would help contacting the service be useful? This differs from arguing that everything is due to pills or all in the mind. Agreement on involvement matters; immediate danger needs appropriate urgent help. A child does not become the principal monitor of an adult's condition. A helper's limits do not mean indifference, and care does not authorize changing another person's prescription.

After contact, check concrete outcomes: was urgency assessed, was the history communicated, were remaining questions identified, and is further contact accessible? This is no promise that every appointment resolves uncertainty. If only a new booking was obtained, keep that distinction visible. Ask what help is planned and who will arrange it. Someone need not declare themselves fully reassured to continue a conversation. Respecting uncertainty supports more careful action, not permanent inaction.

## Practice: a short sequence without a conclusion

If there is no urgent danger and your condition allows, say or write three statements. First, what already changed in treatment, as far as known. Second, what happened afterward and which timing details are approximate. Third, what concerns you now and how it affects ordinary life. Keep a hypothesis separate: I think it may be related, but I do not know the cause.

Choose one question for medical contact: what needs assessment to distinguish possible explanations? Add who continues care and where to turn if things worsen. Do not create a missed dose, discontinuation, or restart to complete the note. An app, complete list, and helper are unnecessary; begin with one sentence about the present difficulty.

The outcome is a communicated observation with honest limits of knowledge. It is not withdrawal, recurrence, or a new illness diagnosed alone. Dangerous deterioration takes priority over preparation.

## Key quotation

> “and none spake a word unto him”

Job 2:13, KJV. [Context 2:11–13](https://www.biblegateway.com/passage/?search=Job%202%3A11-13&version=KJV): the friends recognise great suffering. Here it suggests listening before imposing a theory, not delaying medical care or requiring silence about risk.

## Map of connections between traditions

| Text | Relationship | Application and limit |
|---|---|---|
| Hume, VII.26–28 | Succession is not an established necessary connection | Communicate a sequence; do not deny clinical evidence or repeat dangerous events |
| Dewey, I, §§2–4 | Doubt can become a question about grounds | Separate observation and hypothesis; no medication experiment |
| Aristotle, I.3 | Precision should fit the subject and grounds | Mark an approximate date as approximate; arbitrary explanations remain unjustified |
| Kalama Sutta, AN 3.65 | Authority alone does not settle ethical discernment | Ask about grounds; religious practice is neither self-treatment nor blame of illness |
| Job 2:11–13 | Presence can precede explanation | Listen and help seek care; the biblical seven days is not a waiting period |

## Sources and limits

1. [NHS, Antidepressants](https://www.nhs.uk/medicines/antidepressants/), page reviewed 12 June 2025. Reuses main sections and stopping material actually read in this task, including possible late onset and prolonged symptoms; not each product's instructions.
2. [Royal College of Psychiatrists, Stopping antidepressants](https://www.rcpsych.ac.uk/mental-health/treatments-and-wellbeing/stopping-antidepressants). Read sections on timing, symptom quality, adaptation, and review; not the complete guidance or dose schedules.
3. [FDA benzodiazepine warning](https://www.fda.gov/drugs/drug-safety-and-availability/fda-requiring-boxed-warning-updated-improve-safe-use-benzodiazepine-drug-class), 23 September 2020. Read statements about physical dependence with prescribed use and dangerous abrupt withdrawal; one class, not all medicines.
4. [NIMH, Depression](https://www.nimh.nih.gov/health/publications/depression). Fresh displayed information on urgent help and collecting a history read; not the whole booklet or diagnosis of the reader.
5. [David Hume, An Enquiry Concerning Human Understanding, VII.26–28](https://davidhume.org/texts/e/7). Epigraph and continuation concerning repetition checked; not the entire work.
6. [John Dewey, How We Think, I, §§2–4](https://www.gutenberg.org/files/37423/37423-h/37423-h.htm), 1910. Reuses actual task reading concerning doubt and grounds; educational philosophy.
7. [Aristotle, Nicomachean Ethics, I.3](https://classics.mit.edu/Aristotle/nicomachaen.1.i.html), W. D. Ross translation. Read Book I context on precision; not a modern diagnostic standard.
8. [Kalama Sutta, AN 3.65](https://accesstoinsight.org/tipitaka/an/an03/an03.065.than.html), Thanissaro Bhikkhu translation. Translator's note and displayed conversation on doubt, skillful/unskillful qualities, and the wise read; not the full collection or medical evidence.
9. [Job 2:11–13, KJV](https://www.biblegateway.com/passage/?search=Job%202%3A11-13&version=KJV). Complete specified passage read, not the entire book.

Scenes, note organisation, and ethical applications are original exposition. Medical information and philosophical or religious comparisons do not replace one another.

## Tags

medicines; withdrawal; antidepressants; benzodiazepines; physical dependence; recurrence of illness; side effects; treatment history; shared decisions; access to help; relatives' boundaries; patient autonomy

## About the authors

Andrey Zarinsh is the author and founder of United Wisdom. He explores inner transformation, the unity of traditions, intercultural dialogue and contemporary spiritual discipline. Instagram: @andrey.zarinsh.

Maria Evilkut is a co-author of United Wisdom. Alongside her research, she seeks to test these ideas in practice and observe how principles of unity and mutual understanding work in real life between people and cultures. Instagram: @ilkutki.

*MHR-090 r01, 25 October 2026.*