When depression comes back: why a recurrence is not a failure

“I’m back where I started. All of it was for nothing.”

That sentence lands in my room with the same weight every time. It is spoken by someone who came to therapy two or three years ago, got better, stopped, and is now here – frightened, disappointed in themselves and convinced they have just proved something bad about who they are. The worst part is not the return of the symptoms. The worst part is the interpretation they add: that there was no real progress after all, that it was only a good run of months, that something is fundamentally wrong with them.

In this piece I want to explain why that interpretation is almost never correct – and what a recurrence actually is.

Recurrences are the rule, not the exception

Let’s start with the data, because in this case it is unusually comforting. For a large share of people, depression is a recurring condition. After a first episode, roughly half go on to have another; after a second and a third the likelihood is higher still. This does not apply to everyone – a substantial number of people have a single episode in their lifetime and never another – but the pattern of recurrence in depression is common enough that the field treats it as an ordinary course, not as a complication.

If recurrence is that common, then it cannot simultaneously be proof of personal failure. Something that happens to most people is not a sign that you did something wrong. It is a feature of the condition.

The comparison I often use is imperfect, but it helps. Nobody says that asthma treatment failed because a bad week turned up two years later. We say it is a condition that recurs and that we know how to manage. In this respect depression is more like asthma than like a broken arm – and our whole picture of “being cured” is modelled on the broken arm.

Why a return doesn’t erase what therapy did

Here is the core of it. Therapy does not remove the possibility that a person will one day go under again. No therapy can do that and no serious research promises it. What therapy changes is three other things: how quickly you notice, what you do about it, and how far down you go.

And that is almost always what a recurrence reveals. Someone who has been in therapy comes to me after three weeks, not after eleven months. They know how it begins in them. They no longer believe absolutely everything their head tells them. They tell the people at home instead of disappearing. These are not small differences – these are the differences that decide whether the episode lasts a month or a year.

What usually sets off a recurrence

Rarely a single event. More often several things assemble at once, and that is exactly why it is so hard to see coming.

Among the people I see, the three commonest are these: a period of genuinely heavy load that goes on too long; the collapse of sleep; and the quiet abandonment of everything that had been holding the person up – movement, contact, free time – because “there is no time for that now”. If the last one sounds familiar, that is no accident: the road into a recurrence and the road into burnout often overlap.

A particular case is stopping medication. Many people stop taking an antidepressant as soon as they feel better – which is understandable and at the same time one of the riskiest periods. Guidelines usually speak of continuing for several months after improvement, and longer in recurrent forms. This is not a decision to make alone, or with a therapist; it is made with a doctor. I have written about the relationship between medication and therapy here.

Cognitive reactivity: why the road opens faster

There is one finding I consider among the most useful in this field. In people who have been depressed before, a small dip in mood triggers a considerably larger avalanche of negative thinking than it does in people who never have. This is called cognitive reactivity.

Translated into everyday terms: two people have a bad day. The first, who has never been depressed, thinks the day was bad. The second, who has, arrives within half an hour at the conclusion that the whole thing is pointless and that it is coming back. The pattern has been walked once and is therefore more quickly accessible – much like a path through snow that someone has already trodden.

This sounds like bad news, but it is actually useful. It means that the first wave of a recurrence is often inflated by this mechanism. Part of what you experience then as evidence about your life is only an old path opening quickly. That is not a reason to ignore the warning – but it is a reason not to believe every part of it straight away. I have written about how a thought catches and spins in the piece on rumination.

A map of early signs

The most practical thing I do with someone who has had depression before is a personal map of early signs. We write it while the person is well, not while they are unwell – at that point there is neither the will nor the clarity for it.

Two things matter here more than the list itself. First: the signs must be written by the person, in their own words, because early signs are highly personal – for one it is that they stop cooking, for another that they stop answering messages, for a third that they suddenly start working excessively. Second: someone close needs to know what is on the sheet. Not so that they can police you, but because the very capacity to recognise it is the first thing depression takes.

Where my argument does not hold

Now I have to correct myself, because otherwise this piece would be consolation without a backbone.

First: sometimes a recurrence really does say something about the treatment. If someone relapses for the third time in four years, each time in the same pattern, and each time we do the same thing, that is not a sign that therapy is generally powerless – but it is a sign that this approach is missing something essential in them. It might be an unrecognised bipolar disorder, an unresolved physical condition, or a relationship or set of circumstances that no conversation will change. So a recurrence can also be a useful piece of information, and it is unwise to smooth it over with reassurance.

Second: not every bad stretch is a recurrence. Three hard weeks after a funeral, a redundancy or a house move is neither depression nor relapse. If we declare every dip in mood to be the illness returning, a person ends up living in fear of their own feelings and loses any trust that they are allowed to be sad. I have unpicked the difference in depression is not sadness.

And third, the most uncomfortable: telling someone that a recurrence is not a failure does not lighten what they are living through. I know people who liked that sentence and who could still barely move for the next six weeks. An explanation does not heal. It only stops a person from loading shame on top of the episode – which is not nothing, but is also not everything.

What works in prevention

The best researched route is continuing whatever helped. With medication that means maintenance treatment agreed with a doctor, which measurably lowers the likelihood of recurrence. In psychotherapy, mindfulness-based cognitive therapy and continuation cognitive behavioural therapy have the strongest evidence in people with recurrent episodes; both were developed for exactly this job.

Beyond that there is something that sounds too simple to matter: a steady rhythm. A fixed waking time, meals, movement, at least one weekly contact that is not conditional on mood. In people prone to recurrence, the collapse of rhythm is often the first link in the chain. And finally I have learnt that in a recurrence it matters greatly what voice a person uses with themselves; I have written about why self-compassion here is a tool rather than a softness in the piece on self-compassion.

What we do differently the second time

When someone comes back, we do not start from scratch. For many this is a surprise – they expect to have to go through all of it again.

First we look at what helped the first time and pick that up immediately; there is no reason to hunt for a new method if the old one worked. Then we look at what was different before this episode, and what has changed in their life. And third, the most useful part: we look at what the person already did right when it began – because they almost always did something, they just don’t count it, since it didn’t rescue them. That you arrived three months earlier than last time is a result of the first course of therapy, not evidence of its failure.

If you are wondering how progress is recognised at all, I have written about it in the piece on the signs that therapy is working. None of them is “it will never be bad again”.

What about the people close to you

A recurrence is hard not only for the person living it. A partner who has already survived eighteen months of such a period is often more frightened by the first signs than the patient is. And in relatives that fear usually shows up as one of two reactions: as monitoring – questions, reminders, counting – or as withdrawal, because the first time was too much.

What works better than either is an agreement made in advance. While the person is well, they can agree with someone close what that person may say if they notice the signs, and what the two of them will then do. “If you don’t go out for two weeks, I’ll say so and we’ll book the appointment together.” Such an agreement is an enormous relief for both, because when things worsen there is no longer any need to invent a response on the spot, in the middle of an argument.

And one more thing that relatives are rarely told: you are not obliged to be a therapist. Someone who checks your mood every day is not a partner but a supervisor, and it exhausts them both. It is more useful to stay the person who remains a person – and to keep hold of your own life even when things are hard at home.

What to remember

Recurrence is a common course in depression, not a personal failing. Something that happens to roughly half of people after a first episode cannot be proof that you didn’t try hard enough.

Therapy does not remove the possibility of a return. It shortens the road from the first sign to action, and it is that road which determines how long and how deep the episode will be.

The first wave of a recurrence is often inflated by cognitive reactivity. The conclusions you draw about your life at that point are spoken at the worst possible moment and do not have to be taken as truth.

Write your map of early signs while you are well, in your own words, and tell someone close where it is. That is the most concrete thing you can do today.

And if it has come back right now: you are not at the beginning. You are on the same ground, only this time you have a map.