How to stop worrying and overthinking: what the trials actually show

You stop worrying by changing what you do with the worry, not by trying harder to stop it. Telling yourself to relax does not work, and the trials show why: worry is a habit that feeds on attention, and the techniques that reduce it all involve doing something specific with that attention. The three with the best trial evidence are postponing worry to a fixed slot each day, cognitive behavioural therapy (CBT: a structured talking therapy that works on the link between what you think, what you feel and what you do), and versions of CBT built specifically around worry and rumination, which is the word psychologists use for going over the same ground again and again.

The honest headline is that self-help gives small gains and structured therapy gives moderate ones. Neither is a cure. If worrying has been happening most days for six months and is getting in the way of your life, the HSE (Health Service Executive, the public health service) says to talk to your GP. From there, medical card holders can get free counselling, and anyone can be referred to an online CBT course with weekly support. Details and numbers are at the end.

Why does worrying keep going when it never solves anything?

The leading explanation is the Contrast Avoidance Model, set out in a 2011 review of experimental studies. It argues that people who worry a lot are especially sensitive to being caught off guard by bad news, and that worry “is employed to prolong and maintain a negative emotional state thereby avoiding an unexpected negative emotional shift” (Clinical Psychology Review). In plain terms, staying braced feels safer than being blindsided, so the mind keeps you braced. That relief from never being surprised is what keeps the habit going, even though the worry itself fixes nothing.

A second account focuses on what you believe about the worry. People with pathological worry tend to see it as uncontrollable (Clinical Psychology in Europe), and that belief is itself a target for treatment. Both of these are theories supported by evidence rather than settled fact, but they point the same way: the loop is maintained by what worrying does for you, not by the problems you are worrying about.

When is it ordinary worry, and when is it an anxiety problem?

Everyone lies awake before a mortgage decision or a hospital appointment. The HSE describes generalised anxiety disorder as feeling “anxious about a wide range of situations and issues, rather than about a specific event” (HSE). The worry floats from money to work to the kids and back without settling.

The HSE’s rule of thumb on duration is “worrying more than you need to - most days for at least 6 months” (HSE). That is a guide, not a diagnosis. The other test is impact: the HSE says “It’s important to talk to your GP if anxiety is affecting your daily life” (HSE). If you are avoiding things, sleeping badly or snapping at people because of it, that counts.

Does ‘worry time’ actually work, or is it just a trick?

Worry postponement, sometimes called worry time, means that when a worry arrives you note it and put it off to a fixed slot later in the day, then deal with your worries together in that slot. The HSE suggests it as a self-help step: “Try 15 to 30 minutes each day to start with” (HSE).

It has real trial evidence, and the evidence is modest. A meta-analysis (a study that pools the results of earlier trials) combined 7 randomised trials with 999 participants, of whom 250 experienced worry as a burden and most of whom were women (Springer). Compared with simply writing worries down, postponement gave small effects on how long people worried (d = 0.313) and on how often (d = 0.189) (Springer).

A word on those numbers, because they come up again below. Trials report results as an effect size: a single number on a shared scale that researchers group into small, moderate and large. A small effect means the treated group did a little better on average than the comparison group, and plenty of individuals would notice no difference. A moderate effect is a clear average improvement that most people would notice. A large effect means the treated group’s average is well clear of the comparison group’s. Whether the letter is d, g or SMD, it is the same kind of number on the same scale, and a minus sign in the figures below simply means fewer worrying thoughts. So for worry time: the average postponer worried a bit less than the average recorder, not dramatically less. The trials ran for a week to a month, and the authors note that long-term follow-up studies are still lacking.

A 2024 randomised trial tested a stronger version in people with a diagnosis. It recruited 47 people with generalised anxiety disorder and 35 people with hypochondriasis, which is severe health anxiety, and randomised them to a two-session intervention or a waiting list (Clinical Psychology in Europe). The sessions aimed at the belief that worry cannot be controlled, and participants postponed their worry to a set time each day over the six days between sessions. In the generalised anxiety group the effect on worry was large and held at a four-week follow-up. The health anxiety group did not clearly benefit, which the authors put down to the small sample. This is one small trial with a waiting list rather than an active comparison, so it is promising rather than proven.

The practical reading: worry time is worth a week’s honest trial because it costs nothing and the direction of the evidence is consistent. Expect a bit less worry, not none.

What has the strongest trial evidence: CBT, rumination-focused therapy, or self-help?

Structured therapy, and specifically therapy built around the worry itself. A 2025 meta-analysis pooled 55 randomised controlled trials with 4,970 participants across diagnoses and measured repetitive negative thinking, the umbrella term for worry and rumination (Psychological Medicine). CBT reduced it with a moderate effect (g = -0.67). Treatments designed specifically to target worry and rumination did better, with a large effect (g = -0.99), than more general approaches, which had a moderate one (g = -0.56) (Psychological Medicine). That comparison is across trials rather than a head-to-head, so it could partly reflect differences in who was studied. Notably, delivery format did not matter: efficacy “was not significantly enhanced by individual or in-person settings”, which supports online and guided self-help versions.

One honest caveat on the question itself. That meta-analysis lumps worry-focused and rumination-focused treatments together as “RNT-specific”, so it cannot say whether a therapy aimed purely at rumination beats one aimed at worry. The only rumination-only pooled evidence in the research behind this piece is a review of 8 studies in 856 people with cancer, mostly using mindfulness-based approaches, which found a large reduction in rumination (effect size -0.894) but with results so varied between studies that the authors say to interpret them with caution (Journal of Advanced Nursing). Only half of those studies were randomised, and the population is specific, so treat it as a hint rather than an answer.

For generalised anxiety disorder specifically, a 2024 network meta-analysis (a type that compares several treatments at once across trials) of 65 randomised trials and 5048 participants found CBT reduced symptoms compared with usual treatment with a moderate-to-large effect (SMD, -0.74; SMD stands for standardised mean difference, the same effect-size scale described above). Third-wave CBTs, meaning newer variants that add mindfulness and acceptance techniques to standard CBT, did about the same (SMD, -0.76) (JAMA Psychiatry). Relaxation therapy looked helpful at first but lost its advantage once weaker trials were removed. When the authors looked at anxiety 3 to 12 months after treatment, “only CBT remained significantly associated with greater effectiveness than treatment as usual”, with a moderate effect (SMD, -0.60) (JAMA Psychiatry).

Two cautions. First, gains shrink over time: a separate meta-analysis of long-term outcomes found that at 12 months or more after treatment the effect for generalised anxiety disorder was small (Hedges g, 0.22), from 10 studies (JAMA Psychiatry). Second, effect sizes depend heavily on what CBT was compared against; trials using a waiting list produce much larger numbers than trials using care as usual (JAMA Psychiatry). None of these trials were run in Ireland.

What do you do at 3am when the spiral starts?

First, a line that matters more than anything else in this section. If the spiral includes thoughts of harming yourself, worry postponement and to-do lists are not the tools for tonight. Freetext HELLO to 50808, or call the Samaritans free on 116 123; both are available around the clock (HSE).

For the ordinary 3am loop about money, work or the kids, the HSE’s sleep advice is not to lie there fighting it: “get up but keep the light levels low”, do something calm, and go back when you feel drowsy (HSE). It also suggests you “make time to plan for the next day, such as writing a to-do list” before bed (HSE).

There is a small experiment behind the to-do list idea. In a sleep laboratory, 57 healthy adults aged 18-30 spent 5 min writing before bed, randomised to either a to-do list for the coming days or a list of things already done (Journal of Experimental Psychology: General). The to-do list group fell asleep faster, and the more specific the list, the faster they dropped off. This was one night, in young people without insomnia or an anxiety diagnosis, so treat it as a cheap thing to try rather than an established treatment. The logic fits the postponement approach: writing the task down is a way of parking it.

If broken sleep is the main problem, CBT for insomnia (CBT-I: a short structured programme that changes sleep habits and beliefs about sleep) is the standard treatment. A meta-analysis of 15 randomised controlled trials found it had moderate-to-large effects on worry (Hedge’s g range: -0.41 to g = -0.71) but small and unreliable effects on rumination (g = -0.13), and it seemed to act more on worry about sleep than on worry in general (Sleep Medicine Reviews). The same review found no clear link between the drop in negative thinking and the drop in low mood. So fixing sleep helps you sleep, and may quieten worry about sleep, but do not expect it to switch off daytime overthinking on its own.

Do sleep, alcohol, exercise and your phone really make it worse?

Exercise has the best evidence of the four. An umbrella review (a review of reviews) covering 97 systematic reviews, 1039 trials and 128 119 participants found physical activity had a medium effect on anxiety symptoms (median effect size=-0.42) compared with usual care, with higher intensity giving larger improvements (British Journal of Sports Medicine). Two caveats: most of the reviews were rated critically low quality, and the outcome was anxiety symptoms broadly, not worry or rumination as such.

Alcohol is guidance rather than trial data. The HSE says “Alcohol affects your mental health and can make anxiety worse in the long-term” and warns that using it to cope with difficult feelings “can create long-term problems, including addiction” (HSE). It also names the next-day dread many people know as “the fear”. For sleep, the HSE advises to “avoid drinking alcohol within 4 hours of bed time” if your sleep is disrupted (HSE).

Caffeine follows the same rule: “avoid caffeine at least 4 hours before your bedtime” (HSE).

Your phone is where the evidence is thinnest. The HSE advises “Do not use any device with a screen in the hour before bed” (HSE). That is sensible sleep hygiene, but the research behind this piece found no trial showing that cutting phone use specifically reduces worry, and the blue-light explanation is debated among sleep researchers. If scrolling at midnight is what starts your spiral, stopping is worth doing for that reason alone. Just do not expect it to be the fix by itself.

Where do you get help in Ireland, and what does it cost?

Start with your GP. The HSE’s threshold is worry on most days for six months or worry that is affecting your daily life. A GP can rule out physical causes, talk through options, and make the referrals below.

• Free counselling if you have a medical card. The HSE’s National Counselling Service “provides free counselling for people age 18 or older who have a medical card”, and “up to 8 free counselling sessions may be available” depending on your needs (HSE). The service is called Counselling in Primary Care (CIPC), and it is the HSE service that runs this referral pathway. You need a referral from a GP or another health professional. One catch worth knowing: “After your GP makes a referral, you must phone your local CIPC office within 2 weeks to confirm you want counselling” (HSE), or the referral lapses. Your GP can give you the local office number. Waiting times are not published on the page. If you do not have a medical card, this route is closed to you.
• Guided online CBT by referral. The HSE’s SilverCloud courses “are based on a talk therapy called cognitive behavioural therapy (CBT)” (HSE). You need a referral, which can come from your GP, Primary Care Psychology, Counselling in Primary Care, a community mental health team or Jigsaw. You work through it in your own time, usually over 6 to 8 weeks, and “a supporter from SilverCloud will check in with you weekly for the first 6 to 8 weeks through the website” (HSE). Given the meta-analysis finding that format did not change how well CBT worked, this is a reasonable first step rather than a second-best one.
• Information line. The HSE runs a freephone line, 1800 111 888, “for information on services and how to access them” (HSE).
• If it becomes a crisis. The HSE lists two round-the-clock supports: freetext HELLO to 50808, and the Samaritans on freephone 116 123 (HSE).

Private counselling is the other route, and fees vary by counsellor and area, so no figure is quoted here. GP visit fees also vary, and whether you pay depends on whether you hold a medical card or GP visit card.

The practical order for most people: try worry postponement and the bedtime list for a week or two, keep moving, and go easy on the drink. If the worry has been most days for six months, or it is costing you sleep, work or relationships, book the GP. The trials say the loop responds to treatment. They also say it responds better to a structured programme than to willpower, and there is no shame in using one.

Related Compound guides

If worry is part of a broader period of sustained pressure, start with Stress symptoms: how to recognise stress and what actually helps.

If the loop is driven by harsh self-judgement or fear of not being good enough, read How to improve self-esteem and confidence.