Every technique, with its evidence

Each card states how good the evidence actually is, and anything below the top grade carries its own caveat. If we cannot say what is weak about a recommendation, we should not be making it.

Strong evidenceModerate evidencePreliminary evidenceWell-supported model

Worry triage

Strong evidence3 min

Sort each worry into one you can act on today and one you cannot. They need opposite treatment.

Most worry episodes are a solvable problem and an unsolvable hypothetical welded together. Pulling them apart is the single highest-yield move in the whole programme, because the correct response to each is the exact opposite of the correct response to the other. A current, actionable problem deserves a concrete plan and a next step. A hypothetical β€” "what if the company restructures", "what if the climate news is right" β€” cannot be solved by thinking, only endured or accepted, and every minute spent trying teaches your mind that the worry was worth it.

Evidence

The problem-solving-versus-worry distinction is a core component of IU-based CBT for GAD (Dugas, Robichaud), which outperformed waitlist across IU, worry, anxiety and depression in Dugas et al. (2022).

Open the tool

More on worry triage β†’

Worry postponement

Strong evidence15 min

Capture worries as they arrive, and deal with them only in a fixed daily slot.

When a worry shows up, write one line and return to what you were doing. You are not suppressing it β€” suppression backfires β€” you are deferring it to a scheduled 15-minute period. Two things then happen. Most worries have gone flat by the time the slot arrives, which is direct evidence against "I have to deal with this now". And the ones that are still live get your full attention rather than a leaked hour of it.

Evidence

Stimulus control for worry comes from Borkovec’s model and is a well-established GAD technique that appears in both CBT and MCT protocols.

Caveat: Postponement is a tool for changing your relationship to worry, not a way to never think about anything. If you find yourself using the slot to worry harder, you have turned it into a ritual β€” shorten it.

Open the tool

More on worry postponement β†’

Detached mindfulness

Moderate evidence5 min

Notice a thought arrive, and take no action on it whatsoever β€” including not arguing with it.

A thought appears: "what if I lose my job". The usual moves are to engage it (plan, reassure, check) or to push it away. Detached mindfulness is a third option: register that a thought has occurred, and do nothing. Do not analyse it, do not test whether it is true, do not decide it is unhelpful. Let it sit there and go about your day. The point is not to feel calm; it is to demonstrate to yourself that worry does not have to be responded to, which is what dismantles the belief that it is uncontrollable.

Evidence

A central MCT technique. MCT beat CBT head-to-head in Nordahl et al. and held the advantage at 9-year follow-up (Solem et al., 2021): 57% vs. 38% recovery, GAD re-diagnosis 9.5% vs. 23.1%.

Caveat: The MCT evidence base is much smaller than CBT’s and much of it comes from Wells’ own group. People often convert this into subtle suppression, which is the opposite of the instruction.

More on detached mindfulness β†’

Decatastrophising

Strong evidence10 min

Realistic worst case, most likely case, and β€” the part people skip β€” how you would actually cope.

Catastrophising overestimates both how likely a bad outcome is and how devastating it would be, while systematically ignoring your own capacity to respond. The classic sequence walks all three: what is the genuine worst case (not the cinematic one), what is most likely, and what would you actually do if the worst happened. The coping step matters most and gets dropped most often β€” the fear is usually less about the event than about being unable to handle it.

Evidence

A core Beckian cognitive technique. CBT for GAD shows a large pooled effect vs. control (Cuijpers et al., 2014: g = 0.84, 95% CI 0.71–0.97 across 38 comparisons).

Caveat: Against placebo the effects are more modest (Carpenter et al., 2018: g β‰ˆ 0.56), and roughly 40–50% of GAD patients do not reach recovery thresholds on worry after standard CBT. For some people, engaging worry content this closely feeds the habit β€” if that is you, the metacognitive track is the better fit.

Open the tool

More on decatastrophising β†’

Concreteness training

Moderate evidence8 min

Move from "why is this happening to me" to "what exactly happened, and what is the next step".

Watkins’ finding is that rumination is not harmful because you think about problems β€” it is harmful when the thinking is abstract and evaluative ("why am I like this", "what does this mean about my life"). The same amount of thinking in a concrete, specific, process-focused mode is not harmful and is often useful. This exercise takes one abstract loop and forces it down to specifics: what happened, where, when, what did you do, what is the very next physical action.

Evidence

Watkins’ Phase II RCT found rumination-focused CBT added to treatment-as-usual reduced residual depression and rumination, and reduced comorbid GAD from 11 patients at baseline to 1 post-intervention. Internet-delivered RFCBT (Tulbure et al., 2024) reduced worry, rumination, anxiety and depression at 6-month follow-up.

Caveat: The trials are mostly in depression rather than primary GAD, and the samples are small. Treat it as a strong complement to your primary track rather than a replacement.

Open the tool

More on concreteness training β†’

Uncertainty experiments

Moderate evidence10 min

Deliberately leave small things unchecked, and find out that not-knowing is survivable.

You cannot argue yourself into tolerating uncertainty; you build the tolerance behaviourally, the way you would build any other capacity. Start small and specific: send the email without re-reading it a fourth time, do not check the balance today, leave the message on read for an hour, pick the restaurant without reading the reviews. Predict beforehand how bad it will be and how long the discomfort will last, then record what actually happened. The gap between prediction and outcome is the treatment.

Evidence

Reductions in intolerance of uncertainty during treatment mediate reductions in worry (Bomyea et al., 2015), and experimental manipulation of IU changes worry β€” so targeting it directly is mechanistically supported.

Caveat: Build a ladder. Starting with something high-stakes tends to produce a bad experience that confirms the fear rather than disconfirming it.

Open the tool

More on uncertainty experiments β†’

Values and committed action

Moderate evidence12 min

For the worries that will never resolve: stop trying to settle them, and act on what you care about.

Some uncertainties are not going to close. The climate will not be settled this year; the economy will not issue you a guarantee; nobody can tell you your job is safe forever. ACT’s move is to stop making a good life conditional on resolving them β€” hold the anxious forecast lightly, name what you actually care about, and take one small action in that direction today. Not because it fixes the uncertainty, but because it is how you want to spend the time either way.

Evidence

A-Tjak et al. (2015) found ACT beat controls at g β‰ˆ 0.57; a 2022 group-ACT meta-analysis found g β‰ˆ 0.52 for anxiety. ACT works through its proposed mechanism, psychological flexibility.

Caveat: ACT is roughly equivalent to CBT, not superior to it. Choose it because the frame fits you β€” particularly for uncertainties that genuinely cannot be resolved β€” not because it is stronger.

Open the tool

More on values and committed action β†’

Cyclic sighing

Preliminary evidence5 min

Five minutes a day of double-inhale, long-exhale breathing. Cheap, low-risk, quick.

Two inhales through the nose β€” a full one, then a short top-up β€” followed by a long, slow, complete exhale through the mouth. Repeat for five minutes. The emphasis on the extended exhale is the active ingredient in the tested protocol. This is a base-layer habit, not a worry technique: use it daily on a schedule rather than reaching for it mid-panic, which turns it into another safety behaviour.

Evidence

Balban et al. (Cell Reports Medicine, 2023;4(1):100895, n=111) found exhale-focused cyclic sighing produced greater mood improvement and greater respiratory-rate reduction than mindfulness meditation over one month of 5 min/day.

Caveat: Small, one month long, self-report primary outcomes, no significant HRV change, and the senior author disclosed becoming an advisor to a wearables company during the study period. A cheap low-risk tool with promising-but-early evidence β€” not a cure.

Open the tool

More on cyclic sighing β†’

Aerobic exercise

Strong evidence45 min

The single best-evidenced non-therapy intervention in this whole programme.

Build toward 45–60 minutes, three or more times a week, at least 180 minutes weekly, at an intensity that makes conversation hard. This is not a wellness garnish β€” it has its own effect size, comparable to some psychological interventions, and it makes the cognitive work land better.

Evidence

Wipfli et al. (2008) meta-analysed 49 RCTs and found an effect size of βˆ’0.48 vs. no-treatment controls (Level 1, Grade A). A recent three-level meta-analysis in clinical anxiety samples (27 RCTs, 1,447 participants) found d β‰ˆ βˆ’0.42, with optimal dosing around 45–60 min, 3+ times/week.

More on aerobic exercise β†’

Sleep, alcohol and caffeine

Strong evidence

Fix the timing of sleep, cap caffeine after midday, and cut the evening drink.

Caffeine is a documented anxiogenic and alcohol produces rebound anxiety as it clears β€” often at 4am, which you then interpret as evidence that something is wrong. Poor sleep amplifies everything. None of this is glamorous and all of it moves the numbers. Consistent sleep and wake times matter more than total hours.

Evidence

Sleep, alcohol and caffeine all have well-established relationships with anxiety symptoms; addressing them is treated as foundational in the stepped-care literature.

More on sleep, alcohol and caffeine β†’

Mindfulness practice

Moderate evidence20 min

The best-evidenced option for holding your gains once the acute work is done.

An eight-week structured programme β€” MBSR or MBCT β€” rather than an app streak. Its strongest role in this plan is relapse prevention in Phase 4, once you have already changed how you handle worry.

Evidence

Hoge et al. (JAMA Psychiatry, 2023; TAME trial, n=276) found 8 weeks of MBSR statistically non-inferior to escitalopram (CGI-S improvement 1.35 vs. 1.43; difference βˆ’0.07, 95% CI βˆ’0.38 to 0.23), with far fewer adverse events (15.4% vs. 78.6%). Kuyken et al. (2016) found MBCT reduced depression relapse, HR 0.69 (95% CI 0.58–0.82).

Caveat: Against active comparators the effects shrink substantially β€” in one student population g fell from ~0.42 to ~0.13. Adverse effects occur in roughly 8.3% of meditators and fewer than one trial in five monitors them. Approach intensive practice cautiously, especially with a trauma history.

More on mindfulness practice β†’

Behavioural activation

Strong evidence

Schedule the things you have quietly stopped doing. Withdrawal feeds rumination.

Anxiety shrinks your week without announcing it: the plans not made, the hobby dropped, the friend not replied to. The empty space fills with rumination. Put two or three specific, scheduled activities back in the week β€” by time and place, not by intention β€” regardless of whether you feel like it on the day.

Evidence

Behavioural activation is a well-established, independently effective component of CBT, and counters the withdrawal that maintains rumination.

More on behavioural activation β†’