Recovering from Depression: Your PHQ-9 Score, How Long to Stay on Treatment and How to Stop Safely
Most people get the first part right and the second part wrong.
They start treatment, they stick with it, and some weeks later things genuinely lift. Sleep settles, the mornings stop being an ordeal, work becomes possible again. And then, reasonably, they stop — because the problem appears to be solved.
That is the single most common sequence that leads back to where they started.
📅 The number that matters most on this page. After symptoms have gone, treatment is normally continued for at least six to twelve more months for a first episode. After two or three episodes it is often two years or longer.
Feeling better is not the finish line. It is the point at which the risk of relapse is highest and the reason to keep going is least obvious, which is precisely why so many people stop there.
What depression is and how it presents is covered in our overview of major depressive disorder. This page is about the part that comes after the diagnosis: measuring it, how long treatment runs, what to do when the first drug does not work, how to come off safely, and how to catch a relapse early.
📊 The PHQ-9: nine questions with real cutoffs
This is the instrument used in clinics worldwide, and knowing your number converts a vague sense of being unwell into something concrete to bring to an appointment and to track over time.
Over the last two weeks, how often have you been bothered by any of the following? Score each: 0 not at all, 1 several days, 2 more than half the days, 3 nearly every day.
| Question | |
|---|---|
| 1 | Little interest or pleasure in doing things |
| 2 | Feeling down, depressed or hopeless |
| 3 | Trouble falling asleep, staying asleep, or sleeping too much |
| 4 | Feeling tired or having little energy |
| 5 | Poor appetite or overeating |
| 6 | Feeling bad about yourself, or that you are a failure, or have let people down |
| 7 | Trouble concentrating on things such as reading or watching television |
| 8 | Moving or speaking noticeably slowly, or the opposite — being restless and fidgety |
| 9 | Thoughts that you would be better off dead, or of hurting yourself |
| Total out of 27 | Band | What it usually means |
|---|---|---|
| 0 to 4 | Minimal | Within ordinary range |
| 5 to 9 | Mild | Watchful waiting, therapy, lifestyle measures |
| 10 to 14 | Moderate | 10 is the usual threshold for probable major depression. Therapy, and medication worth discussing |
| 15 to 19 | Moderately severe | Medication and therapy together usually recommended |
| 20 to 27 | Severe | Active treatment, closer follow-up |
🚨 Question 9 is not part of the total in practice. It is a safety question. Any answer other than zero means contacting a doctor or a crisis line today, not at the next appointment. Those thoughts are a symptom of a treatable illness, they are far more common than people assume, and saying them out loud is what makes them manageable.
Two more numbers make the score useful rather than decorative. A 50 percent reduction is counted as a response. Below 5 is remission — and remission, not improvement, is the target, because leftover symptoms are what relapses grow from.
Write the number down with the date. Repeat it monthly. Depression distorts memory of how bad things were, in both directions, and the written record is more reliable than the recollection.
📋 What the diagnosis requires
For completeness, since a score is not a diagnosis.
- 📏 Five or more of the nine symptoms above, present nearly every day for at least two weeks
- ✔️ At least one of them must be depressed mood or loss of interest and pleasure. Without one of those two it is not this diagnosis
- ⚠️ Causing real distress or impairment
- 🔎 Not attributable to a substance or another medical condition
That last line is the reason blood tests belong in the first appointment. An underactive thyroid, anaemia, low B12 or vitamin D, sleep apnea and several medications all reproduce this picture, and all are straightforward to find. The full list and the tests for each are in our guide to apathy and low drive.
One question also belongs here, before any antidepressant is started: has there ever been a period of unusually elevated mood, needing far less sleep without tiredness, racing thoughts or behaviour out of character? If yes, the treatment differs, because an antidepressant alone can destabilise bipolar depression.
⏳ The real timeline, week by week
Expectations cause more treatment failures here than pharmacology does. This is what actually happens.
| When | What to expect |
|---|---|
| Days 1 to 14 | Often worse before better. Nausea, headache, restlessness, disturbed sleep, sometimes increased anxiety. No mood benefit yet. This is the period when most people stop |
| Weeks 2 to 4 | Side effects fade. The first changes are usually sleep and appetite, and other people often notice before you do |
| Weeks 4 to 6 | Meaningful mood improvement typically appears here |
| Weeks 6 to 8 | The point of assessment. Six to eight weeks at a therapeutic dose is an adequate trial. Judging earlier than this judges nothing |
| Weeks 8 to 12 | Full effect. Energy and motivation lag behind mood, often by several weeks |
Two practical consequences. If nothing at all has changed by week four, that is worth a conversation about the dose rather than a conclusion about the drug — many people are left on a starting dose that was never meant to be the treating dose. And if mood has lifted while drive has not, that is the normal sequence rather than a failure, and it is covered in detail in the guide to residual apathy.
One warning that belongs here: in people under 25, the first weeks of an antidepressant carry a small increased risk of suicidal thoughts. It is the reason for close contact early on, and any new agitation or hopelessness in that window needs reporting the same day rather than waiting.
📅 How long to stay on it
The question people most want answered, and the one most often left vague.
Treatment runs in phases, and the phases exist for a reason: symptoms resolve well before the underlying episode has finished. Stopping at the point of feeling well means stopping in the middle.
| Phase | Duration | Purpose |
|---|---|---|
| Acute | Weeks 0 to 12 | Get to remission — PHQ-9 under 5, not just better |
| Continuation | 6 to 12 months after remission | Prevent return of the same episode. Stopping here roughly doubles that risk |
| Maintenance | 2 years or longer, sometimes indefinitely | Prevent a new episode, where risk is high |
Maintenance is usually recommended where there have been two or more previous episodes, where the episode was severe or involved suicidal thinking, where symptoms never fully cleared, or where there is a strong family history.
The recurrence figures are the argument, and they are worth knowing because they are rarely stated plainly:
- 📈 After one episode — roughly half of people have another at some point
- 📈 After two — around 70 percent
- 📈 After three — around 90 percent
Which reframes the decision. The question is not do I still need this but which episode number am I on. And one more thing worth saying because the worry is so widespread: staying on an antidepressant long term is not dependence. There is no craving, no tolerance requiring higher doses, and no compulsive use — the three things that define addiction. It is closer to staying on a blood pressure tablet.
🔄 When the first one does not work
This section exists because the usual interpretation of this situation is catastrophically wrong.
💡 Roughly a third of people reach remission on the first antidepressant they try. Across several sequential attempts, about two thirds get there.
So the first drug failing is the ordinary case, not the exception, and it predicts very little about the second. The conclusion to avoid is that nothing will work — the evidence says the opposite, provided the sequence continues.
Before changing anything, two things get checked, because both are common and neither needs a new prescription:
- 📏 Was the dose high enough? Starting doses are frequently never increased
- ⏰ Was it long enough? Six to eight weeks at that dose
Then there are three routes, and they are broadly comparable in effectiveness, so the choice depends on what happened the first time.
| Route | When it is chosen |
|---|---|
| Increase the dose | Partial response and tolerating it well |
| Switch | No response at all, or intolerable side effects. Within the same class, or to a different mechanism — an SNRI such as venlafaxine or duloxetine, or bupropion, or mirtazapine |
| Add something | Partial response worth keeping. Bupropion for residual low drive, mirtazapine for insomnia and poor appetite, lithium or a second-generation antipsychotic in specialist care |
Matching the drug to the leftover symptom is where the skill is. Cannot sleep and not eating — mirtazapine. Flat and unable to start — bupropion. Pain alongside the depression — duloxetine. Sexual side effects the main complaint — bupropion or mirtazapine. There is no single best antidepressant, only a better or worse fit to a particular person.
And psychotherapy belongs in every one of these rows rather than competing with them. Cognitive behavioural therapy and behavioural activation perform comparably to medication in mild to moderate depression, the combination beats either alone in moderate to severe, and therapy keeps working after it stops, which medication does not.
💊 Escitalopram, as the common starting point
Lexapro (Escitalopram) is one of the most frequently chosen first-line options, and the reasons are unglamorous: it works as well as anything else in its class, it is among the better tolerated, and it has relatively few interactions with other medicines.
Typical dosing is 10 mg once daily, with 5 mg as a starting step where anxiety is prominent, and 20 mg as the usual maximum. Morning or evening, whichever suits sleep better, and consistency of timing matters more than which. It also treats generalized anxiety at the same doses, which matters because the two conditions travel together — see our guides to the worry loop and anxiety treatment options.
What to expect and what to raise:
- 🤢 Nausea and headache in the first one to two weeks, usually settling. Taking it with food helps
- 🛏️ Sleep can go either way — if it disturbs sleep, move the dose to the morning
- 💔 Sexual side effects are common and under-reported — reduced desire, delayed orgasm. There are specific answers for this, including adding bupropion or switching, so it is worth saying rather than enduring
- 😐 Emotional blunting can appear later and is dose-related. Covered fully in the guide to apathy on antidepressants
- ⚠️ Not with an MAOI, and caution combining with tramadol, triptans or other serotonergic drugs
- 🍺 Alcohol worsens mood, disrupts sleep and makes the whole treatment less likely to work
🔽 Coming off it properly
This is the section that is missing almost everywhere, and the gap causes real distress and a great deal of unnecessary alarm.
When an antidepressant is reduced or stopped, a recognised set of symptoms can appear: flu-like aching, dizziness and unsteadiness, nausea, insomnia and vivid dreams, irritability and anxiety, and unusual sensory sensations — the brief electrical-feeling jolts people describe as brain zaps. They typically begin two to four days after the change.
📌 How to tell discontinuation symptoms from the depression coming back — which is the question that matters, because the two lead to opposite actions.
Discontinuation starts within days of the change, is physical as much as emotional, includes dizziness and the electrical sensations, and improves within days to weeks or immediately on restoring the dose.
Relapse comes on over weeks, is dominated by low mood, loss of interest and hopelessness, and has no physical signature.
Three things govern how hard this is:
- ⏳ How long the drug stays in the body. Short-acting drugs are much worse — paroxetine and venlafaxine are the two most associated with difficult discontinuation. Fluoxetine, which lingers for weeks, is the easiest, and one established tactic is switching to it to cover the final stages of a taper
- 📉 How fast the reduction is. Abrupt stopping is the main cause
- 🕒 How long it has been taken. Longer treatment generally means a slower taper
On the taper itself, current thinking has moved: the reductions should get proportionally smaller as the dose gets lower, rather than coming down in equal steps. The last few milligrams occupy far more of the target than the arithmetic suggests, so the final stretch is where the slowest steps belong — weeks or months rather than days, with the pace set by how each reduction feels and with liquid formulations or pill-splitting used for the small steps.
And the point that most needs saying: these symptoms are not withdrawal in the addiction sense. There is no craving, no escalation, no compulsion. It is a nervous system readjusting to the absence of a drug it had adapted to, which is why it settles. Knowing that in advance is the difference between an uncomfortable fortnight and a frightening one.
Never stop abruptly, never stop because you feel well without discussing it, and if a taper proves difficult, going back up and descending more slowly is a normal adjustment rather than a failure.
🚩 Your own early warning signs
The most useful relapse prevention tool costs nothing and takes ten minutes, and it has to be done while you are well.
Most people have a consistent personal sequence that precedes an episode — and it is usually the same sequence every time, which makes it detectable. Common first signs:
- 🛏️ Sleep drifting later, or waking earlier. Very often the first domino
- 📵 Cancelling things, letting messages pile up
- 🍽️ Skipping meals or living on whatever is nearest
- 😡 Irritability out of proportion, often noticed by others first
- 📺 Losing interest in one specific thing you normally rely on
- 🌀 Rumination increasing, the same thought circling
- 🚬 Drinking more, or starting earlier in the evening
Write down your own three or four, in order, and tell one person what they are. Insight is among the first things an episode takes, so the plan has to exist outside your head and ideally in someone else possession.
Then attach an action to it: at two signs, restore the sleep schedule and contact the doctor. Catching it in week one is a dose adjustment. Catching it in month three is another full episode.
For people with three or more previous episodes, mindfulness-based cognitive therapy has specific evidence for reducing recurrence, comparable in some trials to staying on maintenance medication. Worth asking about by name.
🏃 What holds the ground
These do not replace treatment and they measurably improve its results.
- 🛏️ A fixed wake time, every day including weekends. The strongest single lever, because sleep timing is both symptom and driver. Wake time anchors the rhythm more effectively than bedtime
- 🏃 Around 150 minutes of aerobic exercise a week, which has a genuine antidepressant effect. Where that is unreachable, a ten-minute walk daily counts and beats a plan that never starts
- 🌞 Morning daylight, 20 to 30 minutes, outdoors where possible
- 📅 Scheduled activity before motivation — the core of behavioural activation. Action first, feeling later, which is the reverse of how it seems it should work
- 👥 One social contact a day, however small. Withdrawal deepens the episode and is the easiest thing to let slide
- 🚫 Less alcohol. It worsens mood on a delay, disrupts the second half of the night, and reduces how well treatment works
- 🍽️ Regular meals, because falling blood sugar produces symptoms indistinguishable from anxiety and worsened mood
📈 What recovery actually looks like
The order surprises people and gets misread as failure, so it is worth stating in advance.
| Stage | What changes |
|---|---|
| First | Sleep and appetite. Frequently noticed by someone else before you notice it |
| Then | Function. You do things while still not wanting to — which is real progress and the stage most often dismissed as none |
| Then | Mood. The heaviness lifts, the self-criticism quietens |
| Later | Drive and then pleasure. Usually last, and the reason to keep going when mood alone has improved |
The most sensitive marker to track is not how you feel but how long it takes to recover from a bad day. Hours instead of a week is the shift that matters, and it shows up before anything else is obvious.
Setbacks within recovery are expected and are not relapses. The thing to watch is the direction over a month, which is exactly what the monthly PHQ-9 is for.
📞 When to seek help
🚨 Today, without waiting: thoughts of ending your life, thoughts of harming yourself or anyone else, feeling unable to stay safe, or any answer other than zero to question 9 above. Use emergency services, a crisis line, or tell someone you trust now. This is urgent and it is treatable, and the thoughts pass with treatment.
Also urgent: new agitation, hopelessness or suicidal thinking in the first weeks of an antidepressant, particularly under 25.
Promptly:
- A PHQ-9 of 10 or above, or any score with real impairment at work, at home or in relationships
- No change at all by week four — usually a dose conversation rather than a verdict on the drug
- No remission after six to eight weeks at a proper dose — time to increase, switch or add, not to give up
- Mood recovered but drive, sleep or pleasure still flat — residual symptoms are what relapses grow from and are worth treating
- Thinking about stopping — worth planning a taper rather than simply ceasing, and worth checking which episode number you are on first
- Dizziness, electrical sensations or flu-like symptoms within days of a dose change — discontinuation effects, not relapse, and the pace can be adjusted
- Two or more of your own early warning signs appearing together
- Sexual side effects, weight change or emotional flatness — all have specific solutions and all go unreported
- Any history of elevated mood or reduced need for sleep — before starting or changing an antidepressant
- Drinking to cope, which changes both the diagnosis and the plan
- Pregnancy, planning a pregnancy, or breastfeeding — decisions here are individual and need a clinician, not a guess
❓ Frequently asked questions
What does my PHQ-9 score mean?
0 to 4 minimal, 5 to 9 mild, 10 to 14 moderate, 15 to 19 moderately severe, 20 to 27 severe. A score of 10 is the usual threshold for probable major depression, and below 5 counts as remission.
How long do I need to take an antidepressant?
At least six to twelve months after symptoms have gone for a first episode, and two years or longer after two or three episodes. Stopping when you feel better roughly doubles the risk of the episode returning.
My first antidepressant did not work. Does that mean nothing will?
No. Only about a third reach remission on the first drug, while roughly two thirds get there across several sequential attempts. First check the dose was high enough and the trial lasted six to eight weeks.
How long before an antidepressant works?
Sleep and appetite often shift in weeks two to four, mood in weeks four to six, with full effect by eight to twelve. Six to eight weeks at a therapeutic dose is an adequate trial, so judging earlier judges nothing.
Are antidepressants addictive?
No. There is no craving, no need for rising doses and no compulsive use. Stopping can cause discontinuation symptoms, which is a nervous system readjusting rather than withdrawal from an addictive drug.
What are brain zaps and how do I avoid them?
Brief electrical-feeling sensations that can appear two to four days after reducing or stopping an antidepressant. They are avoided by tapering slowly, with proportionally smaller steps at lower doses, rather than stopping abruptly.
How do I tell discontinuation symptoms from relapse?
Discontinuation starts within days of a dose change and is physical, with dizziness and electrical sensations. Relapse builds over weeks and is dominated by low mood and loss of interest, with no physical signature.
Which antidepressant is hardest to come off?
Paroxetine and venlafaxine, because they leave the body quickly. Fluoxetine is the easiest as it lingers for weeks, and switching to it is one recognised way to cover the final stages of a difficult taper.
What is the escitalopram dose for depression?
Usually 10 mg once daily, starting at 5 mg where anxiety is prominent, with 20 mg as the usual maximum. Timing can be morning or evening depending on sleep, and consistency matters more than which.
How do I catch a relapse early?
Write down your own three or four earliest warning signs while well, in order, and tell one person. Sleep drift is commonly first. Insight is an early casualty, so the plan needs to exist outside your own head.
📑 Sources and editorial
- Diagnostic criteria for major depressive disorder, including the five-symptom threshold, the two-week duration and the requirement for depressed mood or loss of interest
- Validation studies of the PHQ-9 as a severity and outcome measure, including the 5, 10, 15 and 20 cutoffs, response and remission definitions, and the handling of item 9 as a safety item
- Clinical practice guidelines on the acute, continuation and maintenance phases of antidepressant treatment, including recommended durations and indications for long-term maintenance
- Meta-analyses of relapse prevention with continued antidepressant treatment compared with discontinuation
- Longitudinal studies of recurrence risk by number of previous depressive episodes
- Sequenced treatment research in major depression, reporting remission rates at the first treatment step and cumulatively across sequential steps, and comparing switch against augmentation strategies
- Evidence on lithium, second-generation antipsychotics, bupropion and mirtazapine as augmentation strategies in incomplete response
- Literature on antidepressant discontinuation syndrome, including symptom profile, onset within days, relationship to elimination half-life, and differentiation from relapse
- Work on hyperbolic dose-occupancy relationships and proportionally tapered dose reductions when discontinuing antidepressants
- Prescribing information for escitalopram, including dosing range, maximum dose, interactions and the age-related warning on suicidal thinking early in treatment
- Comparative evidence on cognitive behavioural therapy, behavioural activation and combined treatment in depression of varying severity
- Trials of mindfulness-based cognitive therapy for prevention of recurrence in people with three or more previous episodes
- Evidence on residual symptoms following remission as predictors of relapse, and on aerobic exercise, sleep timing and alcohol as modifiers of outcome
- Related reading: major depressive disorder overview, apathy and low drive in depression, the worry loop, obstructive sleep apnea
- Related products: Lexapro (Escitalopram), Sertafine (Sertraline), Effexor (Venlafaxine), Wellbutrin (Bupropion), Remeron (Mirtazapine), antidepressants category
- RXshop Editorial Team — reviewed by Sarah Mitchell, PharmD — Clinical Pharmacist & Chief Editorial
Medical Disclaimer: The information in this article is for educational and informational purposes only and is not a substitute for professional medical advice, diagnosis, or treatment. Always seek guidance from a qualified healthcare provider with any questions you may have regarding a medical condition, and before starting, stopping or changing any medication.