The decision

Everything so far has been argument. This chapter is the practical part. I'll put the decision in one sentence and then explain how it works.

When a what-if arrives, I either take a step or give it an appointment.

Those are the only two options, and neither of them is a round. If there's something you can do about the concern, you do it, or you write down when you will. If there's nothing you can do, the thought gets postponed to a fixed time later in the day, and you go back to what you were doing. You don't argue with the thought or try to suppress it, and you don't follow it either.

The method has six parts. Therapists will recognise most of them, since they're adapted from standard cognitive behavioural work on worry. The notes say where each comes from.

Catch it and write it down. When you notice you're in a round, or about to be, put the worry into one line as a prediction. "I'll be criticised in Thursday's chat." A worry that's vague and endless in your head becomes one sentence on paper, and one sentence is much easier to deal with. This is the log from chapter 4, and you've already started it.

Ask whether there's a step. It's a single question, which is whether there's anything you can do about this, today or on a day you can name. Be strict with yourself. "Think about it more" isn't a step. A step is a physical act that you could tick off. It might be to ring the engineer, ask the manager what the chat is about, book the GP appointment, look up the passport renewal time, or move two hundred pounds into the savings account.

If there's a step, take it or date it. Do the smallest version now if you can. If you can't, write down when you will. "Ring engineer, Monday 9am." Then you're finished with that worry until Monday. If it comes back in the meantime, the answer is "that's dealt with on Monday", and you don't need to go over it again. A worry that's been turned into a dated step often loses most of its force on the spot, because the alarm has done its job.

One kind of step needs a special mention. If the worry is about a symptom in your body, the step is an appointment with a doctor. Searching online isn't a substitute for that, and neither is another week of monitoring it. One appointment will settle more than forty evenings of reading. If you're afraid of what you'll be told, that's understandable, but the fear is there whether you go or not.

If there's no step, give it an appointment. Choose a fixed time each day, fifteen to twenty minutes, in the same place. The study in chapter 10 used thirty minutes, and you can too. Don't hold it in bed or in the hour before bed. A kitchen table at six o'clock is good. That's the worry slot. When a what-if turns up outside it and there's no step to take, you say to yourself "not now, at six", add it to the list if it isn't there, and turn back to what's in front of you. This isn't suppression, which the white bear study showed doesn't work. With suppression you tell yourself never to think of the thing. Here you've given it a time, and that turns out to be much easier to live with. Your mind will let go of something it knows it'll come back to.

Keep the appointment. At six, sit down with the list and a pen and go through it. You'll find that a good half of the day's items have gone flat. At eleven that morning they seemed urgent, and by six you can hardly remember why. Cross those off. For the ones that remain, do three things in writing. Write down the worst that could happen, then what's most likely to happen, then what you'd do the next morning if the worst came. That last one is chapter 7's exercise. If a step occurs to you while you're writing, date it. When the time is up, stop, even if you're in the middle of a thought. You'll have another slot tomorrow. Some days you'll sit down and find there's nothing on the list worth the time. In that case, close the notebook and go and do something else.

Check the log weekly. Once a week, mark the predictions that have come due, with a tick or a cross and a word on how bad it was. After a month you'll have your own version of the study in chapter 4, with your own percentage. I don't know what it will be. I'd be surprised if it's far from theirs.

Here's the whole thing applied to the chat on Thursday. On Friday afternoon your manager asks for a chat and the first what-if fires. You write, "She'll criticise my work on Thursday." You ask if there's a step, and there is. You could ask her. "Happy to. Is there anything I should bring or think about beforehand?" It takes fifteen seconds. Either she tells you it's about Leeds, and the weekend is yours, or she says "no, nothing", which tells you nothing, and that's fine. There's no further step. The worry goes on the list. On Saturday at the birthday party it comes back, and you tell it "six o'clock". At six you write that the worst is a formal warning, that the most likely thing is some routine request, and that if it's a warning you'll ask for the specifics in writing and talk to a colleague you trust. That takes four minutes. On Thursday you mark the log with a cross.

A few plans made in advance help, because rounds start at predictable moments. Research on plans in "when, then" form suggests they improve follow-through. These are three I'd suggest.

When I notice a what-if in bed, then I'll write one line on the pad beside me and leave it for the slot.

When I want to check the phone or look up a symptom a second time, then I'll write the worry down and won't check.

When I catch myself halfway through a round, then I'll ask the test question from chapter 2 and get up and do one physical task.

The last one matters more than it looks. After you've postponed a worry, your attention needs somewhere to go, and "nowhere in particular" will lead straight back. Give it something with some grip, such as the washing-up, a walk to the postbox, or the conversation that's going on in front of you.

As for what to expect, the first week feels artificial. The slot seems silly, and some worries won't wait for it. That's normal, and you just go on postponing them. Worries will keep arriving, because the alarm hasn't changed and wasn't supposed to. What usually changes over a few weeks is how long each one stays. I can't promise you a figure or a timetable. The method is a package of my own assembly. Its parts have evidence behind them, but nobody has tested the package. What I can say is that it costs twenty minutes a day at most, and you can check whether it's working from your own log.