Depression in adults according to NICE NG222: the treatment decision tree, step by step
What the NICE guideline on depression in adults recommends, drawn as two successive pathways from one document: treating the episode, then what follows remission. Each node carries the sentence of the guideline that decides, what the guideline does not say at that step, and what the recommendation rests on. Every card has its step number; hovering lights the path already travelled in amber and the steps that can follow in pale blue. A button switches from what is recommended to what it rests on, and the tree then turns entirely grey: NICE publishes no grades.
This tree reads from the bottom up. Choose the pathway first. Its entry points are at the very bottom, the outcomes and what to do at the top, and the step numbers climb with you.
Colour, the verb used
- Recommended. The text gives a direct instruction: discuss, review, monitor, explain, offer. NG222 writes “offer” only 7 times in its 137 recommendations, and “discuss” 24 times. 24 nodes.
- To consider. The text says “consider”, 39 times in the recommendations. 6 nodes.
- Refusals. Advised against when a risk motivates the refusal, lack of evidence when missing evaluation does, not routinely when the text writes its own exception. This tree uses only the third, once, for 1.5.3.
- Situation, fork. A moment of the pathway, or a point where it divides. No node asks an explicit question, so none carries Yes or No.
Shape, the nature of the act
- Discussion, review or assessment.
- Psychological treatment, referral or specialist care.
- Medicine or dose.
- Monitoring over time.
- Diamond. Clinical situation or fork.
In the tree, the icon takes the colour of its node. The nature of the act never sets the colour.
Line, what the source does not say
- Dotted lines. A link no sentence states. There are none on this tree, and that is a finding: every link rests on a sentence of the text, an “if”, a “for all people having treatment”, a “see the recommendations on”.
- Trunk. Time, from the entry upwards. Pathway 1 carries a main trunk, the psychotic depression column on the right, and a short column on the left where active monitoring continues table 1.
- Twigs. They climb to the canopy, which applies at every step of the pathway shown.
- Step number, an amber tab on each card: from 1 to 10 in pathway 1, from 1 to 8 in pathway 2. The trunk carries plain numbers; a card of the same rank on another branch takes a letter: 1a, 3b.
- Pale blue halo. On hover or click, the steps that can follow light up in pale blue, and the path back to the entry in amber.
- No line between the two roots, nor between the two pathways. The text does not join them; the nodes concerned say so in words.
Your pathway
- Click a card in the tree: the path leading to it from the entry point is listed here.
Nothing leaves your browser. What you type here is neither sent nor stored, on this site or anywhere else, and it is cleared when you close or reload the page. The printed sheet fits on one A4 page and keeps only the essentials of each step. Identify the patient by initials, never by name. Keeping the sheet in the medical record is your responsibility and falls under medical confidentiality.
The tree in full text
What follows is not a copy of the drawing, it is its source. Each node is written once, here, with its coordinates, parent, colour and badge carried as attributes. The script reads this text to draw the tree; it does not produce it. There is therefore no second copy that could drift apart. The two pathways are written one after the other.
Pathway 1, steps 1 and 2: the two entry points, and severity
A new episode of depression in an adult
Where the guideline starts: it sorts new episodes by severity
Entry point.
Next: step 2, Two levels of severity, with 16 on the PHQ-9 as the dividing line.
What the recommendation says
This guideline has therefore defined new episodes of depression as less severe or more severe depression.NICE, Depression in adults: treatment and management (NG222), Recommendations chapter, Definitions of depression and severity, no version label, page 7 of the PDF. Text consulted in its version after the December 2025 amendment. Open on nice.org.ukTwo entry points
This pathway has a second root, step 1a: depression with psychotic symptoms, which section 1.12 treats on its own. No sentence of section 1.12 joins this trunk, so no line is drawn between the two roots.
What depression covers here
The guideline defines depression by the absence of positive affect, low mood and associated emotional, cognitive, physical and behavioural symptoms, and refers to ICD-11 or DSM-5 for detail. Severity is a continuum made of three elements: symptoms, duration, and impact on personal and social functioning. PEB summary, not a quotation.
What it does not say
The guideline does not say how to make the diagnosis: it refers to ICD-11 and to DSM-5 without choosing between them. It does not define a new episode, and the number of previous episodes plays no part at this stage; recurrence appears later, as a risk factor for relapse (1.8.2). Bipolar disorder is set aside by a single bullet of the assessment section (1.2.7), which refers to the NICE guideline on bipolar disorder.
What it rests on
Not graded. This is a definition, not a recommendation, and it carries no version label. NICE publishes no level of evidence for its recommendations: the guideline refers to a separate page, Making decisions using NICE guidelines, for how its wording reflects strength. That page is not in the PEB corpus.
Depression with psychotic symptoms: offer referral to specialist mental health services
Second entry point, with its own short trunk
Entry point.
Next: step 2a, Consider an antidepressant combined with an antipsychotic; step 2b, Does not want an antipsychotic: treat with an antidepressant alone.
What the recommendation says
Offer referral to specialist mental health services for people with depression with psychotic symptomsNICE, Depression in adults: treatment and management (NG222), recommendation 1.12.1, [2022], page 73 of the PDF. Text consulted in its version after the December 2025 amendment. The sentence continues: “where the treatment should include:”, followed by a list. Open on nice.org.ukNo junction with the main trunk
Section 1.12 refers neither to further-line treatment (1.9) nor to relapse prevention (1.8), and neither of those sections mentions psychotic symptoms: no occurrence in 1.8 and 1.9 (search of 15 September 2026). The tree draws no line between the two roots. Whether those sections apply here is not written.
What specialist care should include (1.12.1)
- a risk assessment;
- an assessment of needs;
- a programme of coordinated multidisciplinary care;
- access to psychological treatments, after improvement of acute psychotic symptoms.
Discuss the options and, for people who have capacity, reach a shared decision. PEB summary of a list.
What it does not say
Psychotic symptoms are not defined, and the guideline makes no distinction between mood-congruent and mood-incongruent features. Section 1.13 on ECT does not mention psychotic depression. Bipolar disorder is referred to another NICE guideline (1.2.7).
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 105: combination treatment would not usually be started in primary care, hence the referral. Evidence review G, not in the corpus.
Two levels of severity, with 16 on the PHQ-9 as the dividing line
Subthreshold and mild are less severe; moderate and severe are more severe
Comes after step 1, A new episode of depression in an adult.
Next: step 3, Discuss the person’s preferences, including declining treatment; step 3b, More severe depression: choose together from table 2; step 3c, Less severe depression: choose together from table 1, guided self-help first.
What the recommendation says
More severe depression encompasses moderate and severe depression, and in this guideline was defined as depression scoring 16 or more on the PHQ-9 scale.NICE, Depression in adults: treatment and management (NG222), Terms used in this guideline, More severe depression, no version label, page 86 of the PDF. Text consulted in its version after the December 2025 amendment. Open on nice.org.ukThe same threshold, written twice
Page 7 presents it as an example: thresholds on validated scales were used as an indicator of severity, and a PHQ-9 score of 16 is given “for example”. Page 86, in Terms used in this guideline, the same threshold becomes a definition, “was defined as”. The page 7 sentence also lacks a comparison sign before 16, in the PDF and on the NICE web page alike. The tree quotes the definition and shows the gap.
Not the usual PHQ-9 bands
The PEB page on the PHQ-9 gives the authors’ conventional bands at 5, 10, 15 and 20, and a screening threshold of 10. The 16 used by NG222 is neither: it is this guideline’s own line between its two levels. The PHQ-9 on PEB (in French).
What it does not say
No other scale is named and no threshold is given on any other instrument, although the guideline speaks of thresholds on validated scales in the plural. It does not say what to do when the score and the clinical picture disagree, nor whether severity is reassessed during treatment. Moderate depression no longer exists as a category of its own: a moderate episode is more severe depression.
What it rests on
Not graded. Definitions and terms carry no version label, and neither passage cites a source for the threshold.
Consider an antidepressant combined with an antipsychotic
Olanzapine or quetiapine, as examples
Comes after step 1a, Depression with psychotic symptoms: offer referral to specialist mental health services.
Next: step 3a, Monitor the response, including unusual thought content and hallucinations.
What the recommendation says
Consider combination treatment for people with depression with psychotic symptoms with antidepressant medication and antipsychotic medication (for example, olanzapine or quetiapine).NICE, Depression in adults: treatment and management (NG222), recommendation 1.12.2, [2022], page 73 of the PDF. Text consulted in its version after the December 2025 amendment. Open on nice.org.ukWhat it does not say
No antidepressant is named, no dose is given, and olanzapine and quetiapine are examples with no order between them. No time before judging the combination, and no alternative if these antipsychotics are not tolerated.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 105: “some limited evidence” that the combination may provide some benefit, and “some evidence for olanzapine and quetiapine”; the committee knew that quetiapine has antidepressant as well as antipsychotic actions. Evidence review G, not in the corpus.
Does not want an antipsychotic: treat with an antidepressant alone
Preference decides
Comes after step 1a, Depression with psychotic symptoms: offer referral to specialist mental health services.
What the recommendation says
If a person with depression with psychotic symptoms does not wish to take antipsychotic medication in addition to an antidepressant, then treat with an antidepressant alone.NICE, Depression in adults: treatment and management (NG222), recommendation 1.12.3, [2022], page 73 of the PDF. Text consulted in its version after the December 2025 amendment. Open on nice.org.ukWhat it does not say
Nothing on monitoring specific to this choice, and nothing on what to do if psychotic symptoms persist on an antidepressant alone.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
The rationale, page 105, gives no evidence statement on an antidepressant alone.
Pathway 1, steps 3 and 4: choosing and starting treatment
Discuss the person’s preferences, including declining treatment
The two tables are the basis of the conversation
Comes after step 2, Two levels of severity, with 16 on the PHQ-9 as the dividing line.
Next: step 4, Review how well the treatment is working, 2 to 4 weeks after starting.
What the recommendation says
Discuss with people with depression their preferences for treatments (including declining an offer of treatment, or changing their mind once a treatment has started)NICE, Depression in adults: treatment and management (NG222), recommendation 1.3.3, [2022], page 13 of the PDF. Text consulted in its version after the December 2025 amendment. The sentence continues: “by providing:”, followed by a list. Open on nice.org.ukWhere the two tables lead
Steps 3b and 3c, the two tables of options, both send back here in their own words: “see also the recommendations on choice of treatments” (1.5.2 and 1.6.1). That is why they branch off beside this step.
What the discussion provides (1.3.3)
- information on the treatments NICE recommends, their benefits and harms, waiting times and expected outcomes;
- a choice of treatment, of format (individual or group, in person or remotely) and of place;
- the option of attending with a family member or friend;
- the option of choosing the gender of the professional, keeping a professional already known, or changing if the relationship does not work.
Section 1.3.1 also asks what the person thinks contributed to the depression, their previous episodes and treatments, and what they hope to gain. PEB summary of lists.
What it does not say
No length or format for the discussion. The recommendations name no decision aid: they refer to visual summaries that are not in the PEB corpus. Nothing says what to do when the preferred treatment is not available locally, beyond asking commissioners to make treatments available in a timely manner (1.3.4).
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 93: the evidence showed that people with depression and professionals want time for meaningful discussions and trusting relationships. Evidence review I, patient choice, not in the corpus.
Monitor the response, including unusual thought content and hallucinations
The only monitoring written for this situation
Comes after step 2a, Consider an antidepressant combined with an antipsychotic.
Next: step 4a, Consider continuing the antipsychotic for a number of months after remission.
What the recommendation says
Monitor people with depression with psychotic symptoms for treatment response (in particular for unusual thought content and hallucinations).NICE, Depression in adults: treatment and management (NG222), recommendation 1.12.4, [2022], page 73 of the PDF. Text consulted in its version after the December 2025 amendment. Open on nice.org.ukWhat it does not say
No interval, no scale and no criterion for response.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 105: on the basis of its experience, the committee agreed that the effectiveness of the combination should be monitored.
More severe depression: choose together from table 2
Ten options, all usable as first-line treatment
Comes after step 2, Two levels of severity, with 16 on the PHQ-9 as the dividing line.
What the recommendation says
Discuss treatment options with people who have a new episode of more severe depression, and match their choice of treatment to their clinical needs and preferencesNICE, Depression in adults: treatment and management (NG222), recommendation 1.6.1, [2022], page 47 of the PDF. Text consulted in its version after the December 2025 amendment. Opening of the recommendation, followed by a list. Open on nice.org.ukTable 2, in the order NICE gives
Combination of individual cognitive behavioural therapy (CBT) and an antidepressant ; Individual CBT ; Individual behavioural activation (BA) ; Antidepressant medication ; Individual problem-solving ; Counselling ; Short-term psychodynamic psychotherapy (STPP) ; Interpersonal psychotherapy (IPT) ; Guided self-help ; Group exercise.
1.6.1 states that all these treatments can be used first line. The table title gives the order as the committee’s interpretation of clinical and cost effectiveness and of implementation factors: it is not a ranking of evidence.
What the table says about antidepressants
SSRIs are described as generally well tolerated, with a good safety profile, to be considered as the first choice for most people; TCAs are dangerous in overdose, lofepramine having the best safety profile. A first review usually within 2 weeks, at 1 week for a new prescription between 18 and 25 years or when there is a particular concern about suicide. Typically taken for at least 6 months, with regular reviews. For guided self-help and group exercise, the table asks to weigh first the advantages of options with more therapist contact. PEB summary of table 2.
What it does not say
No antidepressant is named beyond the classes, and no dose is given. The guideline does not say which option to choose when the person has no preference. The rationale places SSRIs and SNRIs first line, while the table names SSRIs alone as the first choice for most people.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 98: “good evidence” for the effectiveness of CBT combined with antidepressants, individual CBT and individual behavioural therapies, the three together; good evidence for antidepressants, with SSRIs and SNRIs first line “because of their tolerability”, a TCA possible for someone who responded well to one, mirtazapine kept for further-line treatment; “some evidence” for counselling and problem-solving. Evidence review B, not in the corpus.
Less severe depression: choose together from table 1, guided self-help first
Eleven options, all usable as first-line treatment
Comes after step 2, Two levels of severity, with 16 on the PHQ-9 as the dividing line.
Next: step 4b, No antidepressant as routine first line, unless it is the person’s informed preference; step 4c, Does not want treatment, or feels better: active monitoring.
What the recommendation says
Discuss treatment options with people with a new episode of less severe depression, and match their choice of treatment to their clinical needs and preferencesNICE, Depression in adults: treatment and management (NG222), recommendation 1.5.2, [2022], page 29 of the PDF. Text consulted in its version after the December 2025 amendment. Opening of the recommendation, followed by a list. Open on nice.org.ukTable 1, in the order NICE gives
Guided self-help ; Group cognitive behavioural therapy (CBT) ; Group behavioural activation (BA) ; Individual CBT ; Individual BA ; Group exercise ; Group mindfulness and meditation ; Interpersonal psychotherapy (IPT) ; Selective serotonin reuptake inhibitors (SSRIs) ; Counselling ; Short-term psychodynamic psychotherapy (STPP).
1.5.2 asks to take into account that all these treatments can be used first line, but to consider first the least intrusive and least resource intensive, guided self-help, and to recognise the right to decline treatment. The rationale places guided self-help first “for pragmatic reasons”; the rest of the order follows the committee’s consensus on average effectiveness and cost effectiveness (page 97).
Guided self-help, as table 1 describes it
Printed or digital materials built on structured CBT, structured behavioural activation, problem-solving or psychoeducation, with support from a trained practitioner, usually 6 to 8 structured regular sessions. PEB summary.
What it does not say
No criterion for moving from guided self-help to another option, apart from preference and clinical needs, and no time limit before judging it has failed other than the general rule of step 5. No product, programme or digital tool is named.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 96: “good evidence” for group CBT and group BA, likely the most cost effective; good evidence for individual BA and individual CBT; “some evidence” for guided self-help, considered first because it is the least intrusive. Evidence review B, not in the corpus.
Review how well the treatment is working, 2 to 4 weeks after starting
Adherence, side effects, suicidal ideation, an outcome measure
Comes after step 3, Discuss the person’s preferences, including declining treatment.
Next: step 5, No response at all after 4 weeks of medication, or 4 to 6 weeks of therapy: look for a cause.
What the recommendation says
review how well the treatment is working with the person between 2 and 4 weeks after starting treatmentNICE, Depression in adults: treatment and management (NG222), recommendation 1.4.3, [2009, amended 2022], page 15 of the PDF. Text consulted in its version after the December 2025 amendment. One bullet of 1.4.3, a complete clause, introduced by “For all people with depression having treatment:”. Open on nice.org.ukIf the episode remits
The guideline never writes “if the treatment works, go to relapse prevention”. It opens relapse prevention “after full or partial remission” (1.8.1). That is pathway 2, which starts there.
The rest of 1.4.3, and the antidepressant timetable
1.4.3 also asks to monitor concordance, side effects and harms, and suicidal ideation, particularly in the early weeks, and to consider routine outcome monitoring with validated sessional measures, for example the PHQ-9. For antidepressants, 1.4.11 sets the first review usually within 2 weeks, or 1 week after starting for a new prescription between 18 and 25 years or when there is a particular concern about suicide, and says the effect, if the drug is going to work, usually shows within 4 weeks. PEB summary.
What it does not say
Neither response nor remission is defined: Terms used in this guideline has 13 entries and none covers them (count of 15 September 2026 on page 02 of the corpus). No score on any scale marks a response.
What it rests on
Not graded. Label [2009, amended 2022]: NICE changed the wording in 2022 without reviewing the evidence (Update information, page 110).
Consider continuing the antipsychotic for a number of months after remission
Stopping is decided by, or with, specialist services
Comes after step 3a, Monitor the response, including unusual thought content and hallucinations.
What the recommendation says
Consider continuing antipsychotic medication for people with depression with psychotic symptoms for a number of months after remission, if tolerated.NICE, Depression in adults: treatment and management (NG222), recommendation 1.12.5, [2022], page 73 of the PDF. Text consulted in its version after the December 2025 amendment. Open on nice.org.ukThe rest of 1.12.5, and how to stop
Whether and when to stop the antipsychotic is decided by specialist services or in consultation with them. 1.4.37: antipsychotics are stopped only in or with advice from specialist services, gradually over at least 4 weeks and in proportion to the length of treatment.
What it does not say
A number of months is not defined. Nothing is said in this section about how long the antidepressant continues after remission.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 105: people should be reviewed regularly and not left on the combination longer than necessary.
No antidepressant as routine first line, unless it is the person’s informed preference
A refusal with its exception written into the recommendation
Comes after step 3c, Less severe depression: choose together from table 1, guided self-help first.
What the recommendation says
Do not routinely offer antidepressant medication as first-line treatment for less severe depression. Only offer it if that is the person’s informed preference.NICE, Depression in adults: treatment and management (NG222), recommendation 1.5.3, [2022], page 29 of the PDF. Text consulted in its version after the December 2025 amendment. Open on nice.org.ukWhy Not routinely
The refusal carries its exception in its own text: the second sentence names the only situation in which to offer medication. It targets antidepressant medication as a class, not SSRIs alone. NICE amended it in October 2025 to clarify its advice on first-line treatment for less severe depression (Update information, page 110).
What the committee adds
If an antidepressant is used at this level of severity, the committee advised that SSRIs would be preferred, because of their safety and tolerability (rationale, page 97). Table 1 lists SSRIs ninth of eleven options.
What it does not say
Routinely is defined nowhere: the word occurs twice in the 137 recommendations, here and in 1.4.23 (count of 15 September 2026). The guideline does not say how to establish that a preference is informed, nor what to do for someone already taking an antidepressant when the episode is assessed. The amendment note does not give the previous wording.
What it rests on
Not graded. Label [2022], on a recommendation amended in October 2025. Rationale, page 97: the evidence suggested that some psychological therapies were more effective than antidepressants and, with the potential for side effects, medication “should not be the default treatment”. Evidence review B, not in the corpus.
Does not want treatment, or feels better: active monitoring
A further assessment, normally within 2 to 4 weeks
Comes after step 3c, Less severe depression: choose together from table 1, guided self-help first.
What the recommendation says
For people with less severe depression who do not want treatment, or people who feel that their depressive symptoms are improving:NICE, Depression in adults: treatment and management (NG222), recommendation 1.5.1, [2009, amended 2022], page 28 of the PDF. Text consulted in its version after the December 2025 amendment. Opening clause of the recommendation, before its list. Open on nice.org.ukWhy this card sits above table 1
1.5.2 ends by recognising that people have a right to decline treatment, and 1.5.1 says what to do then. The card continues the less severe column because it has no choice beside it; it is not a treatment that follows the table.
What 1.5.1 asks for
- discuss the presenting problems, underlying vulnerabilities and risk factors, and the person’s concerns;
- make sure the person knows they can change their mind, and how to seek help;
- provide information on the nature and course of depression;
- arrange a further assessment, normally within 2 to 4 weeks;
- make contact, with repeated attempts if necessary, if the person does not attend.
PEB summary of a list.
What it does not say
The guideline does not say how long active monitoring lasts, nor what finding at the reassessment should lead to treatment. The section covers less severe depression only: nothing equivalent is written for a person with more severe depression who declines treatment, beyond the right to decline in 1.6.1.
What it rests on
Not graded. Label [2009, amended 2022]: NICE changed the wording in 2022 without reviewing the evidence (Update information, page 110).
Pathway 1, steps 5 to 7: when there is no response
No response at all after 4 weeks of medication, or 4 to 6 weeks of therapy: look for a cause
Context, other conditions, adherence
Comes after step 4, Review how well the treatment is working, 2 to 4 weeks after starting.
Next: step 6, Still no response once the causes are addressed: review the diagnosis.
What the recommendation says
If a person’s depression has not responded at all after 4 weeks of antidepressant medication at a recognised therapeutic dose, or after 4 to 6 weeks for psychological therapy or combined medication and psychological therapyNICE, Depression in adults: treatment and management (NG222), recommendation 1.9.1, [2022], page 65 of the PDF. Text consulted in its version after the December 2025 amendment. The sentence continues: “discuss with them:”, followed by a list. Open on nice.org.ukWhat to discuss (1.9.1)
- personal, social or environmental factors, or physical or other mental health conditions, that might explain why the treatment is not working;
- problems adhering to the plan, for example stopping or reducing medication because of side effects, or missing sessions.
If any of these apply, reach a shared decision on how to address them, including the help other agencies may give. PEB summary of a list.
What it does not say
Not responded at all has no threshold on any scale. 1.9.1 covers no response at all; a limited response appears only from 1.9.4, once no obvious cause has been found. A recognised therapeutic dose is not defined: the recommendations give no antidepressant dose apart from fluoxetine 20 mg and 40 to 60 mg, in the advice on stopping (1.4.18).
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 101: the committee made these recommendations based on its knowledge and experience. Evidence review D, not in the corpus.
Still no response once the causes are addressed: review the diagnosis
Alternative or comorbid conditions
Comes after step 5, No response at all after 4 weeks of medication, or 4 to 6 weeks of therapy: look for a cause.
Next: step 7, Other treatments can be tried: the options depend on what has already been tried.
What the recommendation says
review the diagnosis and consider the possibility of alternative or comorbid conditions that may limit response to depression treatmentsNICE, Depression in adults: treatment and management (NG222), recommendation 1.9.2, [2022], page 66 of the PDF. Text consulted in its version after the December 2025 amendment. Main clause. The sentence opens: “If a person’s depression has not responded to treatment after addressing any problems raised (see recommendation 1.9.1), and allowing an adequate time for treatment changes to work,”. Open on nice.org.ukWhat it does not say
An adequate time is not defined. No list of alternative or comorbid conditions and no investigation is given. The recommendation does not say who reviews the diagnosis, nor whether specialist services come in at this point.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 101: reasons for non-response should be explored and addressed before further-line treatment, on the committee’s knowledge and experience. Evidence review D, not in the corpus.
Other treatments can be tried: the options depend on what has already been tried
Three situations, and ECT
Comes after step 6, Still no response once the causes are addressed: review the diagnosis.
Next: step 8, After an antidepressant alone: add exercise, switch to therapy, adjust the medicine, or combine; step 8a, After a combination: another therapy, a dose change or switch, or an added medicine; step 8b, After psychological therapy alone: another therapy, add an SSRI, or switch to an SSRI; step 8c, Electroconvulsive therapy for severe depression, under one of three conditions.
What the recommendation says
Reassure the person that although treatment has not worked, other treatments can be tried, and may be effective.NICE, Depression in adults: treatment and management (NG222), recommendation 1.9.3, [2022], page 66 of the PDF. Text consulted in its version after the December 2025 amendment. Open on nice.org.ukThree recommendations, three starting points
1.9.4 after psychological therapy alone, 1.9.5 after an antidepressant alone, 1.9.6 after a combination. Each asks to discuss the options and reach a shared decision. 1.9.7 adds that vortioxetine is recommended in a separate NICE technology appraisal, TA367, after no or limited response to at least 2 antidepressants within the current episode; TA367 is not in the PEB corpus.
What it does not say
No order among the options within each situation, no number of attempts, and no definition of a limited response. The words treatment-resistant depression do not appear in this section: in the recommendations they occur only in section 1.15, in its title and in its two referrals to other NICE guidance.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
The card carries the badge Fork because its role in the tree is to divide; the sentence itself is an instruction to reassure.
Pathway 1, steps 8 to 10: further-line options
After an antidepressant alone: add exercise, switch to therapy, adjust the medicine, or combine
The only situation the guideline takes further
Comes after step 7, Other treatments can be tried: the options depend on what has already been tried.
Next: step 9, Wants to combine medicines rather than try a therapy: explain the added side-effect burden.
What the recommendation says
If a person’s depression has had no or a limited response to treatment with antidepressant medication alone, and no obvious cause can be found and resolvedNICE, Depression in adults: treatment and management (NG222), recommendation 1.9.5, [2022], page 66 of the PDF. Text consulted in its version after the December 2025 amendment. Opening clause; the recommendation then asks to discuss the options and lists them. Open on nice.org.ukThe options of 1.9.5
- adding a group exercise intervention;
- switching to a psychological therapy, among the options for more severe depression;
- increasing the dose within the licensed range if well tolerated, knowing that higher doses may not be more effective and can increase side effects, with frequent follow-up;
- switching to another drug of the same class, or of a different class (an SSRI, an SNRI, or in secondary care a TCA or an MAOI), with cross-tapering if needed, any MAOI switch in or with advice from secondary care, and TCAs dangerous in overdose;
- changing to a combination of psychological therapy (for example CBT, IPT or STPP) and medication.
PEB summary of a list.
Why this card carries the trunk
1.9.8 and 1.9.9 continue this situation only: they concern a person whose depression “has had no response or a limited response to antidepressant medication”.
What it does not say
No order among the options and no time before judging a dose increase or a switch. No drug is named for a switch, and there is no cross-tapering schedule: the guideline refers to a NICE clinical knowledge summary that is not in the PEB corpus.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, pages 101 and 102: “some evidence” that adding group exercise was effective; “some very limited evidence” for switching antidepressant or increasing the dose; “some evidence” for combining psychological therapy and antidepressants. Evidence review D, not in the corpus.
After a combination: another therapy, a dose change or switch, or an added medicine
The added medicine sends to step 10
Comes after step 7, Other treatments can be tried: the options depend on what has already been tried.
What the recommendation says
If a person’s depression has had no or a limited response to treatment with a combination of antidepressant medication and psychological therapyNICE, Depression in adults: treatment and management (NG222), recommendation 1.9.6, [2022], page 67 of the PDF. Text consulted in its version after the December 2025 amendment. Opening clause, followed by the options. Open on nice.org.ukA link written, not drawn
The third option, adding in another medication, refers to 1.9.9, step 10 of this tree. This card is a leaf, so no line is drawn; the link is written here.
The three options of 1.9.6
- switching to another psychological therapy;
- increasing the dose or switching to another antidepressant, as in 1.9.5;
- adding in another medication, as in 1.9.9.
What it does not say
No order among the options. The warning about side-effect burden (1.9.8) is written for people on medication alone, not for this situation.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
The rationale, pages 101 and 102, gives no evidence statement specific to people already on a combination. Evidence review D, not in the corpus.
After psychological therapy alone: another therapy, add an SSRI, or switch to an SSRI
Options written from experience, without evidence
Comes after step 7, Other treatments can be tried: the options depend on what has already been tried.
What the recommendation says
If a person’s depression has had no or a limited response to treatment with psychological therapy alone, and no obvious cause can be found and resolvedNICE, Depression in adults: treatment and management (NG222), recommendation 1.9.4, [2022], page 66 of the PDF. Text consulted in its version after the December 2025 amendment. Opening clause, followed by the options. Open on nice.org.ukThe three options of 1.9.4
- switching to an alternative psychological treatment;
- adding an SSRI to the psychological therapy;
- switching to an SSRI alone.
After a discussion that includes what other treatments the person has found helpful in the past.
What it does not say
No SSRI is named, no sequence is set among the three options, and no length is given for a second therapy.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 101: no evidence was identified for people whose depression had not responded to psychological therapies as first-line treatment; the committee used its experience and made a research recommendation.
Electroconvulsive therapy for severe depression, under one of three conditions
Further-line, but not only
Comes after step 7, Other treatments can be tried: the options depend on what has already been tried.
What the recommendation says
Consider electroconvulsive therapy (ECT) for the treatment of severe depression if:NICE, Depression in adults: treatment and management (NG222), recommendation 1.13.1, [2022], page 74 of the PDF. Text consulted in its version after the December 2025 amendment. Opening of the recommendation; its three conditions are joined by “or”. Open on nice.org.ukWhy ECT hangs from step 7, and what follows it
The third condition refers in the text to further-line treatment: “see the recommendations on further-line treatment”. The first two do not, and the committee says ECT should be considered “not just as further-line treatment” (rationale, page 102). If the depression responds to a course of ECT, 1.13.9 asks to start or continue an antidepressant or a psychological intervention to prevent relapse, which is pathway 2, and to consider lithium augmentation.
The three alternative conditions
- the person chooses ECT based on their past experience of ECT and what has worked for them;
- a rapid response is needed, for example when the depression is life-threatening because the person is not eating or drinking;
- other treatments have been unsuccessful.
Life-threatening depression is the example of the second condition, not a fourth condition. PEB summary of a list.
What it does not say
No number of sessions, no electrode placement, no frequency. Severe is not defined in this section, and neither is the stable remission at which ECT stops (1.13.8). Section 1.13 does not mention psychotic depression: no occurrence of “psychot” in 1.13.1 to 1.13.9 (search of 15 September 2026). 1.13.2 to 1.13.5 set out information, consent and repeat courses.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 102: “some limited evidence” for ECT as further-line treatment; for a rapid response, the committee relied on its knowledge, experience and awareness of the wider evidence base for ECT. Evidence review D, not in the corpus.
Wants to combine medicines rather than try a therapy: explain the added side-effect burden
Information comes before any combination
Comes after step 8, After an antidepressant alone: add exercise, switch to therapy, adjust the medicine, or combine.
Next: step 10, Accepts the burden: consider specialist referral or advice for a combination.
What the recommendation says
explain the possible increase in their side-effect burden.NICE, Depression in adults: treatment and management (NG222), recommendation 1.9.8, [2022], page 68 of the PDF. Text consulted in its version after the December 2025 amendment. Main clause. The sentence opens: “If a person whose depression has had no response or a limited response to antidepressant medication does not want to try a psychological therapy, and instead wants to try a combination of medications,”. Open on nice.org.ukWhat it does not say
The burden is not quantified and no specific side effect is listed in this recommendation.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 102: there was evidence that combinations were effective, but the committee agreed they would need specialist advice. Evidence review D, not in the corpus.
Accepts the burden: consider specialist referral or advice for a combination
An added antidepressant, an antipsychotic, lithium, ECT, lamotrigine or liothyronine
Comes after step 9, Wants to combine medicines rather than try a therapy: explain the added side-effect burden.
Next: if the depression remits, pathway 2, After remission, starts at its step 1. The guideline links the two only through the words “after full or partial remission” (1.8.1).
What the recommendation says
consider referral to a specialist mental health setting or consulting a specialist.NICE, Depression in adults: treatment and management (NG222), recommendation 1.9.9, [2022], page 68 of the PDF. Text consulted in its version after the December 2025 amendment. Main clause, preceded by “If a person with depression wants to try a combination treatment and is willing to accept the possibility of an increased side-effect burden (see recommendation 1.9.8),”. Open on nice.org.ukThe options of 1.9.9
- adding an antidepressant from a different class, for example mirtazapine or trazodone with an SSRI;
- combining an antidepressant with a second-generation antipsychotic (for example aripiprazole, olanzapine, quetiapine or risperidone) or with lithium;
- augmenting with ECT, lamotrigine or liothyronine.
Some combinations are dangerous and to be avoided, for example an SSRI, SNRI or TCA with an MAOI. With an antipsychotic, its effects on depression, including loss of interest and motivation, are to be reviewed carefully. In June 2022 some antipsychotics, lamotrigine and liothyronine were off-label for this use. PEB summary of a list.
What it does not say
No order among the combinations, no dose, no duration, and no criterion to choose between referral and consulting. The guideline does not say what to do when these options fail: esketamine and vagus nerve stimulation are only referred to other NICE guidance (1.15). Monitoring of lithium and antipsychotics is set out in 1.4.25 to 1.4.37.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 102. The recommendations on monitoring lithium and antipsychotics were made “by informal consensus” (rationale, page 95). Evidence review D, not in the corpus.
Pathway 1: what applies at every step, and what PEB checked
Suicide risk
Suicide risk, at every step
PEB summary
- Always ask directly about suicidal ideation and intent (1.2.8).
- Refer urgently to specialist mental health services when there is considerable immediate risk to the person or others (1.2.9).
- Do not withhold treatment for depression because of suicide risk; take toxicity in overdose into account and, if necessary, limit the amount of medicine available (1.2.12).
- Between 18 and 25 years, or at increased risk of suicide: assess before prescribing, and review 1 week after starting or increasing an antidepressant (1.4.22).
- Do not routinely start treatment with TCAs, except lofepramine, in people at significant risk of suicide, because of their risk in overdose (1.4.23): the second Not routinely of the guideline.
PEB summary.
Where it is written
1.2.8 to 1.2.12, pages 10 to 12; 1.4.22 and 1.4.23, pages 22 to 23. Open on nice.org.uk
Recognition and assessment
Before the tree: recognising and assessing depression
PEB summary
- Be alert to possible depression and consider two questions on the last month: feeling down, depressed or hopeless; little interest or pleasure (1.2.1).
- A comprehensive assessment that does not rely simply on a symptom count (1.2.6).
- Ask about any history of mood elevation, to determine whether the depression may be part of bipolar disorder, which has its own NICE guideline (1.2.7).
- When depression comes with anxiety symptoms, treat the depression first as a rule; when a person has an anxiety disorder with comorbid depression, consider treating the anxiety disorder first (1.2.13).
PEB summary.
Where it is written
Section 1.2, pages 8 to 12. Open on nice.org.uk
When prescribing
What to tell people when prescribing an antidepressant
PEB summary
Agree a management plan: why medication, which choices, the dose, benefits and harms, including withdrawal effects (1.4.10). Explain how long an effect takes, usually within 4 weeks if the drug is going to work; when the first review will be; that treatment might need to be taken for at least 6 months after remission, but should be reviewed regularly; and how withdrawal effects can be minimised (1.4.11). PEB summary.
Where it is written
1.4.10 and 1.4.11, pages 17 to 18. Open on nice.org.uk
Situations with their own section
Four situations the guideline treats in their own sections
PEB summary
- Behavioural couples therapy, to consider when relationship problems may contribute to the depression or involving the partner may help (1.7).
- Chronic depressive symptoms, with their own treatment options (1.10).
- Depression with a diagnosis of personality disorder: do not withhold treatment, consider an antidepressant combined with a psychological treatment (1.11).
- Access and organisation of care, including collaborative care and crisis resolution and home treatment (1.16).
The tree does not draw them: each would be a pathway of its own. PEB summary.
Where it is written
Sections 1.7, 1.10, 1.11 and 1.16. Open on nice.org.uk
Referred elsewhere
Four treatments the guideline only refers to other NICE guidance
PEB analysis
- Vortioxetine: recommended in technology appraisal TA367 after no or limited response to at least 2 antidepressants in the current episode (1.9.7).
- Repetitive transcranial magnetic stimulation: see the interventional procedures guidance (1.14.1).
- Implanted vagus nerve stimulation for treatment-resistant depression: see the interventional procedures guidance (1.15.1).
- Esketamine nasal spray: not recommended in technology appraisal TA854 for treating treatment-resistant depression (1.15.2).
None of these documents is in the PEB corpus. The tree attributes nothing to them beyond what NG222 says. 1.14.1, 1.15.1 and 1.15.2 carry no version label.
Where it is written
1.9.7, page 68; 1.14.1 to 1.15.2, pages 76 to 77. Open on nice.org.uk
The order of the tables
The order of the tables is not a level of evidence
PEB analysis
Both tables are titled “in order of the committee’s interpretation of their clinical and cost effectiveness and consideration of implementation factors”. Guided self-help heads table 1 for pragmatic reasons (rationale, page 97). For table 2, the committee gives good evidence of effectiveness jointly to CBT with an antidepressant, individual CBT and individual behavioural therapies (page 98). And 1.5.2 and 1.6.1 state that every option in each table can be used first line. The rank says what NICE would try first in the NHS, not what the evidence proves best.
Where it is written
Tables 1 and 2; rationale, pages 96 to 99. Open on nice.org.uk
No grades, version labels
No grades, and what the version labels say
PEB analysis
NICE publishes no level of evidence: its wording carries the strength, and the view What it rests on stays grey on every card. What the text does give is a version label. In the section of recommendations: 99 [2022], 1 [2022, amended 2023], 19 [2009, amended 2022] and 14 [2009], 133 labels in all, and 4 recommendations with none. For 33 of the 133 labelled recommendations, the evidence was not reviewed again in 2022. In pathway 1 of this tree: 18 nodes [2022], 2 [2009, amended 2022], 2 definitions without a label.
The PEB appraisal of the same guideline scores 3 of the 6 AGREE II domains, 72, 78 and 62.5%, and gives it as recommended, with adaptation. Read the guideline appraisal.
Where it is written
Count of 15 September 2026 on pieces 01 and 02 of the corpus; PEB appraisal of 15 September 2026.
St John’s wort and light
St John’s wort, light therapy, exercise and lifestyle
PEB analysis
- St John’s wort: NICE acknowledges evidence of possible benefit in less severe depression, yet asks professionals not to prescribe or advise it, because of uncertainty about doses, persistence of effect and preparations, and serious interactions (1.4.38, [2009]).
- Light therapy for seasonal winter depression: tell people that the evidence for its efficacy is uncertain (1.4.39, [2009]).
- Regular physical activity and a healthy lifestyle may help wellbeing (1.4.40 and 1.4.41).
The refusal on St John’s wort mixes an uncertainty and a risk: no single refusal badge would render it, which is one reason it stays out of the tree.
Where it is written
1.4.38 to 1.4.41, pages 27 to 28. Open on nice.org.uk
Pathway 2, steps 1 to 3: preventing relapse
Remission: discuss continuing the treatment to prevent relapse
Entry to pathway 2, after full or partial remission
Entry point.
Next: step 2, Higher relapse risk after an antidepressant alone: consider continuing it, a group therapy, or both; step 2a, Chooses not to continue the antidepressant: explain how to stop, and when to seek help; step 2b, Remission with a psychological therapy, alone or combined: discuss continuing it.
What the recommendation says
Discuss with people that continuation of treatment (antidepressants or psychological therapies) after full or partial remission may reduce their risk of relapse and may help them stay well.NICE, Depression in adults: treatment and management (NG222), recommendation 1.8.1, [2022], page 62 of the PDF. Text consulted in its version after the December 2025 amendment. Open on nice.org.ukComing from pathway 1
This pathway begins wherever pathway 1 ends in remission: after the review of step 4, after a further-line option, or after ECT, since 1.13.9 refers here. The guideline draws no passage between the two; it writes “after full or partial remission”.
The second sentence, the risk factors, the long-term risks
Reach a shared decision on whether or not to continue, based on clinical needs and preferences (1.8.1). 1.8.2 lists what may increase the likelihood of relapse: recurrent episodes, particularly if frequent or within the last 2 years; incomplete response to previous treatment, including residual symptoms; unhelpful coping styles such as avoidance and rumination; a history of severe depression; other chronic physical or mental health problems; personal, social and environmental factors that contributed and are still present. 1.8.3 asks to discuss the risks of long-term antidepressants, such as an increased bleeding risk, long-term effects on sexual function and difficulty stopping. PEB summary.
What it does not say
Neither full nor partial remission is defined, and no score marks either. 1.8.2 lists risk factors without saying how many make the risk higher, and names no tool to assess it.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 100: the risk factors come from the committee’s knowledge of the wider literature and its experience. Evidence review C, not in the corpus.
Higher relapse risk after an antidepressant alone: consider continuing it, a group therapy, or both
Keep the dose that led to remission
Comes after step 1, Remission: discuss continuing the treatment to prevent relapse.
Next: step 3, Continuing the antidepressant: review at least every 6 months; step 3a, Group CBT or MBCT for relapse prevention: a course focused on staying well.
What the recommendation says
For people who have remitted from depression when treated with antidepressant medication alone, but who have been assessed as being at higher risk of relapse, consider:NICE, Depression in adults: treatment and management (NG222), recommendation 1.8.5, [2022], page 63 of the PDF. Text consulted in its version after the December 2025 amendment. Opening of the recommendation, followed by three options joined by “or”. Open on nice.org.ukThe three options of 1.8.5
- continuing the antidepressant at the dose that led to full or partial remission, unless there is good reason to reduce it, such as side effects;
- a course of group CBT or mindfulness-based cognitive therapy (MBCT) for people who do not wish to continue antidepressants, following the advice on stopping;
- continuing the antidepressant together with group CBT or MBCT.
PEB summary of a list.
What it does not say
No length of continuation is given, and the guideline sets no maximum anywhere. The information given when prescribing says treatment “might need to be taken for at least 6 months after the remission of symptoms, but should be reviewed regularly” (1.4.11). The “up to 2 years” of the rationale, page 100, describes how long the available data lasted, not a recommended duration.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 100: “good evidence” that SSRIs, SNRIs and TCAs, group CBT and MBCT were effective for relapse prevention and, on average, cost effective for people at high risk of relapse, with data for treatment periods up to 2 years. Evidence review C, not in the corpus.
Chooses not to continue the antidepressant: explain how to stop, and when to seek help
Straight to the stopping advice
Comes after step 1, Remission: discuss continuing the treatment to prevent relapse.
What the recommendation says
If a person chooses not to continue antidepressant medication for relapse prevention, advise them:NICE, Depression in adults: treatment and management (NG222), recommendation 1.8.4, [2022], page 63 of the PDF. Text consulted in its version after the December 2025 amendment. Opening of the recommendation, followed by two points. Open on nice.org.ukThe two points of 1.8.4
How to stop the antidepressant, with a referral to the recommendations on stopping, which are steps 4 to 8 of this pathway; and to seek help as soon as possible if depressive symptoms return or residual symptoms worsen.
What it does not say
No follow-up interval after stopping is given. The guideline does not say here how to tell a return of depression from withdrawal symptoms; that distinction is made at step 7 (1.4.19).
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Remission with a psychological therapy, alone or combined: discuss continuing it
At least 4 more sessions if the therapy continues
Comes after step 1, Remission: discuss continuing the treatment to prevent relapse.
What the recommendation says
whether they wish to continue with their psychological therapy for relapse prevention. Reach a shared decision on further treatment.NICE, Depression in adults: treatment and management (NG222), recommendation 1.8.8, [2022], page 64 of the PDF. Text consulted in its version after the December 2025 amendment. End of the first sentence and second sentence. The recommendation opens: “Discuss with people who have remitted from depression when treated with a psychological therapy alone, but who have been assessed as being at higher risk of relapse,”. Open on nice.org.ukThe recommendations that follow
1.8.9 does the same after a combination of antidepressant and psychological therapy: continue one or both treatments, by shared decision. 1.8.10: a person who stays on a psychological therapy continues the same one, adapted for relapse prevention, with at least 4 more sessions. 1.8.12: reassess the risk of relapse at the end of that treatment, and the need for follow-up. PEB summary.
What it does not say
Higher risk is assessed without any named tool. No maximum number of sessions, and no interval between them.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, pages 100 and 101: the committee used its knowledge and experience to set follow-up arrangements, so that people did not remain on therapy indefinitely.
Continuing the antidepressant: review at least every 6 months
Mood, side effects, risk factors, and the wish to stop
Comes after step 2, Higher relapse risk after an antidepressant alone: consider continuing it, a group therapy, or both.
Next: step 4, Stopping: the dose is usually reduced in stages, and most people succeed.
What the recommendation says
Review treatment for people continuing with antidepressant medication to prevent relapse at least every 6 months.NICE, Depression in adults: treatment and management (NG222), recommendation 1.8.11, [2022], page 64 of the PDF. Text consulted in its version after the December 2025 amendment. Open on nice.org.ukWhy step 4 follows
The last point of 1.8.11 sends to the recommendations on stopping in so many words: if the person wishes to stop, “see the recommendations on stopping antidepressant medication”. The trunk follows it.
At each review (1.8.11)
- monitor mood with a validated rating scale;
- review side effects;
- review medical, personal, social or environmental factors that may affect the risk of relapse, and encourage help from other agencies;
- discuss whether the person wishes to continue.
PEB summary of a list.
What it does not say
The rating scale is not named. No end point is set: the review repeats with no stated limit.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Group CBT or MBCT for relapse prevention: a course focused on staying well
Usually 8 sessions over 2 to 3 months
Comes after step 2, Higher relapse risk after an antidepressant alone: consider continuing it, a group therapy, or both.
What the recommendation says
For people starting group CBT or MBCT for relapse prevention, offer a course of therapy with an explicit focus on the development of relapse prevention skills and what is needed to stay well.NICE, Depression in adults: treatment and management (NG222), recommendation 1.8.6, [2022], page 63 of the PDF. Text consulted in its version after the December 2025 amendment. Open on nice.org.ukThe rest of 1.8.6, and 1.8.7
The course usually consists of 8 sessions over 2 to 3 months, with the option of additional sessions over the next 12 months. 1.8.7 lists what relapse prevention components may include: reviewing what was learnt and helpful, concrete plans to maintain progress, identifying warning signs and triggers with contingency plans, and plans for challenging events over the next 12 months. PEB summary.
What it does not say
No group size and no criterion for the additional sessions.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 100: psychological therapies used for relapse prevention should explicitly focus on relapse prevention skills.
Pathway 2, steps 4 to 6: stopping the antidepressant
Stopping: the dose is usually reduced in stages, and most people succeed
Talk to the prescriber first
Comes after step 3, Continuing the antidepressant: review at least every 6 months.
Next: step 5, Taper in proportional steps, at a pace led and agreed by the person.
What the recommendation says
Explain that it is usually necessary to reduce the dose in stages over time (called ’tapering’) but that most people stop antidepressants successfully.NICE, Depression in adults: treatment and management (NG222), recommendation 1.4.12, [2022], page 18 of the PDF. Text consulted in its version after the December 2025 amendment. Second sentence of 1.4.12. The first advises people who want to stop to talk with the person who prescribed their medication. Open on nice.org.ukBefore stopping
1.4.13 to 1.4.15 describe what to tell people about withdrawal symptoms and the support that may help: they are in the canopy, since they hold at every step of this pathway.
What it does not say
The guideline does not say when stopping is appropriate beyond the person’s wish and the reviews of step 3, and gives no length of treatment after which stopping should be raised.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 94: “some limited evidence” that tapering antidepressants may reduce withdrawal effects. The evidence reviews behind these recommendations are those of the NICE guideline on safe prescribing, not in the corpus.
Taper in proportional steps, at a pace led and agreed by the person
A proportion of the previous dose, then smaller cuts as the dose gets low
Comes after step 4, Stopping: the dose is usually reduced in stages, and most people succeed.
Next: step 6, Monitor during the reduction: withdrawal symptoms, and the return of depression.
What the recommendation says
slowly reduce the dose to zero in a step-wise fashion, at each step prescribing a proportion of the previous dose (for example, 50% of previous dose)NICE, Depression in adults: treatment and management (NG222), recommendation 1.4.16, [2022], page 20 of the PDF. Text consulted in its version after the December 2025 amendment. One bullet of 1.4.16, a complete clause, introduced by “When stopping a person’s antidepressant medication:”. Open on nice.org.ukThe rest of 1.4.16
- take into account the half-life, a short half-life needing a slower taper, and the length of treatment;
- consider smaller reductions, for example 25%, as the dose becomes lower;
- consider liquid preparations, if available, once very small doses cannot be reached with tablets or capsules;
- let the person lead and agree the speed and length of withdrawal, with withdrawal symptoms resolved or tolerable before the next reduction;
- a more rapid withdrawal may be appropriate for serious or intolerable side effects, or when switching antidepressants;
- withdrawal may take weeks or months.
PEB summary of a list.
Drug-specific points (1.4.18)
Withdrawal symptoms occur with TCAs, SSRIs, SNRIs and MAOIs. Paroxetine and venlafaxine are more likely to cause them and need particular care. Fluoxetine can sometimes be stopped this way: at 20 mg a day, a period of alternate-day dosing can provide a suitable reduction; at 40 to 60 mg a day, a gradual schedule; allow 1 to 2 weeks to evaluate each reduction. PEB summary.
What it does not say
No schedule per drug, no total duration, and no interval between steps other than the 1 to 2 weeks written for fluoxetine. Only paroxetine and venlafaxine are named as higher risk; the other antidepressants are not ranked.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Rationale, page 94: the detail on tapering techniques, on drugs associated with more withdrawal symptoms and on fluoxetine was added from the committee’s knowledge.
Monitor during the reduction: withdrawal symptoms, and the return of depression
How often depends on clinical and support needs
Comes after step 5, Taper in proportional steps, at a pace led and agreed by the person.
Next: step 7, Withdrawal symptoms are not a relapse: reassure and explain.
What the recommendation says
Monitor and review people taking antidepressant medication while their dose is being reduced, both for withdrawal symptoms and the return of symptoms of depression.NICE, Depression in adults: treatment and management (NG222), recommendation 1.4.17, [2022], page 20 of the PDF. Text consulted in its version after the December 2025 amendment. Open on nice.org.ukWhat it does not say
No interval, no scale, and no criterion for telling the two apart at this stage.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Pathway 2, steps 7 and 8: withdrawal symptoms
Withdrawal symptoms are not a relapse: reassure and explain
Common, and relapse does not usually come at once
Comes after step 6, Monitor during the reduction: withdrawal symptoms, and the return of depression.
Next: step 8, Mild withdrawal symptoms: monitor, reassure, and ask for a call back if they persist; step 8a, More severe withdrawal symptoms: consider restarting the previous dose, then reduce more slowly.
What the recommendation says
If a person has withdrawal symptoms when they stop taking antidepressant medication or reduce their dose, reassure them that they are not having a relapse of their depression.NICE, Depression in adults: treatment and management (NG222), recommendation 1.4.19, [2022], page 21 of the PDF. Text consulted in its version after the December 2025 amendment. Open on nice.org.ukWhat to explain (1.4.19 and 1.4.14)
- these symptoms are common;
- relapse does not usually happen as soon as the antidepressant is stopped or the dose lowered;
- even if the antidepressant is restarted or the dose increased, withdrawal symptoms may take a few days to disappear.
1.4.14: withdrawal symptoms can be mild, appear within a few days and usually go within 1 to 2 weeks; they can last longer, in some cases several weeks and occasionally several months; they can sometimes be severe, particularly after stopping suddenly. PEB summary.
What it does not say
No clinical criterion and no timeline separate withdrawal from relapse, beyond relapse not usually happening as soon as treatment stops. No scale is named.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Mild withdrawal symptoms: monitor, reassure, and ask for a call back if they persist
Usually time-limited
Comes after step 7, Withdrawal symptoms are not a relapse: reassure and explain.
What the recommendation says
If a person has mild withdrawal symptoms when they stop taking antidepressant medication:NICE, Depression in adults: treatment and management (NG222), recommendation 1.4.20, [2022], page 21 of the PDF. Text consulted in its version after the December 2025 amendment. Opening of the recommendation, followed by three points. Open on nice.org.ukThe three points of 1.4.20
- monitor the symptoms;
- reassure the person that such symptoms are common and usually time-limited;
- advise them to contact their prescriber if the symptoms do not improve or get worse.
What it does not say
Mild is not defined, and nothing says how long to wait before treating the symptoms as persistent.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
More severe withdrawal symptoms: consider restarting the previous dose, then reduce more slowly
Smaller decrements once symptoms have resolved
Comes after step 7, Withdrawal symptoms are not a relapse: reassure and explain.
What the recommendation says
If a person has more severe withdrawal symptoms, consider restarting the original antidepressant medication at the previous dose, and then attempt dose reduction at a slower rate with smaller decrements after symptoms have resolved.NICE, Depression in adults: treatment and management (NG222), recommendation 1.4.21, [2022], page 22 of the PDF. Text consulted in its version after the December 2025 amendment. Open on nice.org.ukWhat it does not say
More severe is not defined. The guideline does not say how long to stay at the previous dose, nor how much smaller the new decrements should be.
What it rests on
Not graded: NICE publishes no level of evidence for its recommendations. Label [2022]: the evidence was reviewed for the 2022 update (Update information, page 110).
Pathway 2: what applies at every step, and what PEB checked
No maximum duration
How long to continue: at least 6 months, reviewed, and no maximum
PEB analysis
The only duration written is in the information given when prescribing: treatment “might need to be taken for at least 6 months after the remission of symptoms, but should be reviewed regularly” (1.4.11). No recommendation sets a maximum. The “up to 2 years” sometimes quoted comes from the rationale, page 100, where it describes the length of the data on relapse prevention, not a duration to follow. Search of 15 September 2026: one occurrence of “up to 2 years” in the PDF, on page 100.
Where it is written
1.4.11, page 18; rationale, page 100. Open on nice.org.uk
Withdrawal symptoms
What to tell people about withdrawal symptoms
PEB summary
Stopping abruptly, missing doses or not taking a full dose may cause withdrawal symptoms, which do not affect everyone and vary in type and severity: unsteadiness or dizziness, altered sensations such as electric shocks, altered feelings (irritability, anxiety, low mood, panic, confusion, very rarely suicidal thoughts), restlessness, sleep problems, sweating, nausea, palpitations, tiredness, headaches and aches (1.4.13). People may fear stopping and may need support, especially after previous difficult attempts: written or online resources, more frequent contact (1.4.15). PEB summary.
Where it is written
1.4.13 to 1.4.15, pages 18 to 19. Open on nice.org.uk
Lithium, antipsychotics
Lithium and antipsychotics are stopped by, or with, specialists
PEB summary
Only stop lithium in specialist mental health services or with their advice, reducing gradually over 1 to 3 months whenever possible (1.4.32). Consider at each review whether to continue an antipsychotic, then stop it only in or with advice from specialist services, gradually over at least 4 weeks and in proportion to the length of treatment (1.4.36 and 1.4.37). PEB summary.
Where it is written
1.4.32, 1.4.36 and 1.4.37, pages 25 to 26. Open on nice.org.uk
All [2022], none graded
Twelve recommendations from 2022, none graded
PEB analysis
All twelve nodes of this pathway carry the label [2022]: the evidence was reviewed for the update. None is graded, since NICE publishes no grades. The committee writes “good evidence” for continuing antidepressants, group CBT and MBCT to prevent relapse (page 100), and “some limited evidence” that tapering may reduce withdrawal effects (page 94). The detail of how to taper, which drugs need care and how fluoxetine can be stopped comes from the committee’s knowledge, as the rationale says.
Where it is written
Rationale, pages 94 and 100.
How to read this tree
Colour carries the verb, never the nature of the act. The label says whether a card is a discussion, a medicine or a referral. What the reader needs is whether to do it, and how firmly the text asks. That is what colour answers.
On this document the line runs through one word, “consider”. Starting ECT, adding an antipsychotic in psychotic depression, keeping an antidepressant after remission: the text writes “consider”, and those cards are blue. Discussing, reviewing and monitoring are green.
Why two pathways
NG222 opens relapse prevention “after full or partial remission” (1.8.1), whatever the route that led there: a response at the first review, a further-line option, or ECT. A single pathway would have needed two links the text does not write. Two successive pathways need none, and nothing is drawn twice.
Depression with psychotic symptoms is the second entry of pathway 1. Section 1.12 does not refer to further-line treatment or to relapse prevention, and those sections do not mention psychotic symptoms. The tree therefore draws no line between the two roots, and says so in the cards.
What this tree adds to the guideline
A gap inside the text. The 16 on the PHQ-9 is an example on page 7 and a definition on page 86; the page 7 sentence is also missing its comparison sign. The severity card shows both.
What the text does not say, at each step. This is the most useful part. Response, remission and treatment-resistant depression are never defined. No antidepressant dose is given outside the stopping advice for fluoxetine. No duration of treatment has a maximum, and the “up to 2 years” often quoted is the length of the data, not a recommendation. “Routinely”, “adequate time” and “a number of months” are not defined.
One refusal, with its exception
Not routinely, for 1.5.3: “Do not routinely offer antidepressant medication as first-line treatment for less severe depression.” The second sentence names the exception, the person’s informed preference. The recommendation targets antidepressants as a class. No node carries Advised against or Lack of evidence.
Why the view What it rests on is grey
NICE publishes no grades. Every card therefore reads Not graded, and What it rests on reports instead the committee’s own words in the rationale, good evidence, some evidence, very limited evidence, knowledge and experience, with the evidence review they come from. None of the evidence reviews is in the PEB corpus, so none of those statements could be checked against the studies.
No dotted step
Each link rests on a sentence of the text: the “if” of 1.9.1, 1.9.2, 1.9.4 to 1.9.6, 1.9.8 and 1.9.9, the reference to choice of treatments in 1.5.2 and 1.6.1, the reference to stopping in 1.8.11. That no step had to be deduced is a result, and it is reported as one.
How this content was checked
Each quotation comes from the guideline text, one sentence per node, 35 words at most, found word for word and only once in the document by an automated check. Lists of the source are rephrased and announced as PEB summaries; two quotations are complete clauses taken from a bullet, and say so. The figures were counted again. The two tables were read as tables on the NICE web page, since their columns interleave in the PDF extraction. How we check what we publish.
What could not be checked
As of 15 September 2026. The evidence reviews A to I, the methods, the scope, the committee membership and the declarations of interests. The seven visual summaries the guideline refers to. The technology appraisals on vortioxetine (TA367) and esketamine (TA854), the interventional procedures guidance on transcranial magnetic stimulation and vagus nerve stimulation, and the NICE page on how its wording expresses strength. The tree attributes nothing to these documents beyond what NG222 says.
Analysis from Psychiatry Evidence Base, evidence-based psychiatry, explained with rigour.
