Depression, anxiety, guilt, loneliness, relief

The honest headline is better than the internet’s version and worse than the greeting card’s. Most people recover most of the way. The best long-run evidence comes from a German panel study that followed more than thirty thousand people for eighteen years: life satisfaction falls in the years leading up to a divorce, rises again afterwards, and on average does not quite return to where it started. That paper is titled “Time Does Not Heal All Wounds”, which is both bleak and, read properly, reassuring — the recovery is real, it is just not total.

A second study of the same panel, covering 1984 to 2015 and 1,220 divorces, found the strain lands differently on each partner rather than equally: men’s strain was larger and more transient, women’s smaller at the time and more chronic, driven largely by money rather than mood.

Lucas, R.E., “Time Does Not Heal All Wounds: A Longitudinal Study of Reaction and Adaptation to Divorce”, Psychological Science (2005). Leopold, T., “Gender Differences in the Consequences of Divorce: A Study of Multiple Outcomes”, Demography (2018).

Grief after a divorce is ordinary and it is not a disorder. What distinguishes it from a depressive episode is shape, not intensity. The American Psychiatric Association’s own account of the difference is useful here: in grief, the pain comes in waves, mixed in with memories that are good; self-esteem is usually left intact. In a depressive episode the mood and the thinking are almost constantly negative, and worthlessness — not sadness, worthlessness — is the symptom that marks the line.

Grief also has no schedule, and there is nobody to send a card. Grief that has no funeral is the page for that part.

The threshold that clinicians actually use is not a feeling, it is a duration. Symptoms present most of the day, nearly every day, for at least two weeks, including either depressed mood or loss of interest and pleasure, plus symptoms from the list below. Two weeks is short. If you have been reading this page wondering whether you are “bad enough” to see someone, that is the answer.

• Hopelessness, or a persistent flat emptiness rather than sharp pain.
• Feelings of guilt or worthlessness — the difference from ordinary regret is that these are about you, not about what happened.
• Fatigue that sleep does not fix, and trouble concentrating or deciding anything.
• Sleeping or eating far more or far less than usual.
• Loss of interest in things that used to matter, including the ones you would expect to still matter.
• Thoughts of death or suicide. This one does not wait two weeks. Call or text 988 today.

Symptom list and the two-week threshold: Depression, National Institute of Mental Health (2024). Grief versus depression: American Psychiatric Association, Major Depressive Disorder and the Bereavement Exclusion (2013).

Divorce anxiety mostly does not feel like fear. It feels like a mind that will not stop running scenarios — the money, the children, the hearing, the thing you said in an email in March. The clinical threshold for generalised anxiety disorder is excessive, hard-to-control worry on most days for at least six months, plus at least three of: restlessness, fatigue, trouble concentrating, irritability, muscle tension, and disturbed sleep.

Six months is a long time, and a divorce takes longer than that, which is why this one is easy to normalize until it is entrenched. The treatment with the strongest evidence is cognitive behavioral therapy — NIMH calls it the gold standard for GAD — with SSRIs and SNRIs the main medication options.

What nobody has measured: we went looking for research quantifying the specific divorce anxieties — the custody one, the money one, the dating-again one — as named constructs with prevalence figures, and there is none. They are universally described and never counted. That is a real hole and it is on our list of them.

Generalized Anxiety Disorder: What You Need to Know, National Institute of Mental Health (2024).

These get used as synonyms and they behave completely differently. The research base is June Tangney’s, and the distinction is the one worth carrying around: guilt is about a thing you did; shame is about the kind of person you are. Guilt tends to push people toward repair — apologizing, fixing, doing better by the children. Shame tends to push people toward hiding, and hiding is what makes a divorce lonely on top of everything else.

The practical use of that distinction: when the 3am thought arrives, notice which one it is. “I handled that conversation with my daughter badly” is guilt, and it has an action attached. “I have ruined my children’s lives” is shame, and it has no action attached, which is how you can tell it is not information.

What nobody has studied: whether the person who initiated carries more guilt than the person who was left. It is assumed everywhere. We could not find a single study that compared them.

Tangney, J.P. & Dearing, R.L., Shame and Guilt, Guilford Press (2002).

This is the one with the hardest evidence behind it, and the evidence is not gentle. A meta-analysis of 148 studies covering more than 300,000 people found that people with stronger social relationships had a 50% greater likelihood of survival over an average seven and a half years of follow-up. The US Surgeon General’s 2023 advisory put the risk of premature death from lacking social connection at 26 to 29% higher, and compared the mortality risk to smoking up to fifteen cigarettes a day.

And it is unevenly distributed after a divorce. In the German panel, more than 40% of men reported frequent loneliness in the year they divorced, against roughly 20% of women. The usual explanation is that the marriage was holding the friendships, and one of you kept them.

Which makes rebuilding contact a health intervention rather than a nice idea. It is also the single most actionable thing on this page: buddy connections, the community boards, traveling with friends.

Holt-Lunstad, J. et al., “Social Relationships and Mortality Risk: A Meta-analytic Review”, PLOS Medicine (2010) — OR 1.50, 95% CI 1.42–1.59. Our Epidemic of Loneliness and Isolation, Office of the US Surgeon General (2023). Loneliness split: Leopold, Demography (2018).

Almost nothing written about divorce takes relief seriously, and it is one of the most reliably reported feelings there is. It also has real evidence behind it, and the evidence explains something that causes an enormous amount of unnecessary cruelty between two people.

A 2022 study separated the two people in a break-up instead of averaging them, and found that the person who initiated improved their wellbeing afterwards, across most areas of life except money, while the person who did not initiate took a significant short-term loss and then recovered over the longer run. Averaging the two together hides both facts. A separate study following 1,462 Australian women for more than twenty years found life satisfaction stable beforehand, a sharp drop around the separation, and then long-term increases that the matched still-partnered women did not show.

Two things follow. If you are the one who left and you feel lighter: that is the expected finding, not evidence that you never cared. And if you are the one who was left and their apparent ease looks obscene: you are watching someone who started grieving privately a year or two before you did. You are not behind. You started later.

Brüning, R. et al., “Separations of romantic relationships are experienced differently by initiators and noninitiators”, PNAS (2022). Arcangeli & Ejova, “Does Time Heal All Wounds? Life Satisfaction Trajectories in Australian Middle-Aged Women”, Journal of Happiness Studies (2025).

Not what should help, and not what a book says helped its author. What has trials behind it.

• Cognitive behavioral therapy. The best-supported psychotherapy for both depression and anxiety, and NIMH’s named gold standard for generalised anxiety. If a therapist cannot tell you what modality they practice, that is information.
• Behavioral activation. A Cochrane review found no meaningful short-term difference between it and CBT for adult depression. It is simpler, more structured, and often easier to get: the mechanism is scheduling activity rather than waiting to feel like it.
• Exercise. A Cochrane review found it may reduce depressive symptoms compared to no treatment, and found probably little to no difference between exercise and psychological therapy, or between exercise and antidepressants. The long-term evidence is uncertain and it is well tolerated. Treat it as a real option rather than a consolation prize.
• Medication. SSRIs and SNRIs are the main options for anxiety, and the antidepressant comparison above is a comparison, not a dismissal. Medication, and the custody question covers the thing people are actually afraid of when they hesitate.
• Rebuilding contact with people. See the loneliness numbers above. This is not soft advice.

And one honest gap. Divorce support groups are recommended constantly, including by us in places. We could not find a trial or a review evaluating them for adults. That does not mean they do not work — the loneliness evidence suggests something in that direction should — it means nobody has checked.

Cochrane reviews CD004366 (exercise for depression) and CD013305 (behavioral activation). Psychotherapies and Generalized Anxiety Disorder, National Institute of Mental Health (2024).

• Therapy, and how to find it — what the modalities are, what they cost, and how to get one
• Grief that has no funeral
• Trauma bonds, and breaking free — if leaving is the part you cannot do
• Medication, and the custody question
• Sleep
• Drinking
• Buddy connections — the loneliness answer, in practice
• What nobody has studied yet — including the two holes named on this page