Context Switching with ADHD: Deep Work in a Reactive Job

You are forty minutes into the report, finally holding the whole thing in your head, when someone leans over the partition with a quick question. It takes ninety seconds. Then you turn back to the screen and the whole structure is gone. You read your own last paragraph three times and it belongs to a stranger.

ADHD context switching at work is a specific problem that general focus advice misses. The issue is not starting or even sustaining. It is re-entry: the cost of getting back into a task after you have been pulled out, paid again and again across a day that was never going to be quiet.

Some jobs are reactive by design. This post is about how to do real work inside one anyway, and what to ask for when the switching itself is the problem.

The cost of each switch and why ADHD pays more

Every task you hold has a setup in working memory: what you are doing, why, what comes next, where the pieces are. An interruption clears that setup, and rebuilding it takes time. Research on interrupted work in general populations has measured both a time cost and a stress cost for each switch, and task-switching experiments show a measurable delay every time people change what they are working on. Those findings are about everyone.

ADHD raises the bill in two ways. Working memory holds less to begin with, so the setup that gets cleared was already fragile. And getting attention back onto a half-finished, no-longer-novel task is exactly the kind of initiation that ADHD makes hardest. Many people with ADHD describe a day of ten small interruptions as more tiring than a day of one long crisis, and the mechanism above explains why. Our post on ADHD and working memory challenges goes deeper on the first half of that.

You cannot make switches free. You can make them fewer, make them cheaper, and make the return trip shorter. Those are the three levers in the rest of this post.

Batching the interrupt-driven parts of the job

Most reactive jobs are not uniformly reactive. There are usually two or three categories of interruption, and each has a rhythm. The first step is to find the rhythm so you can gather interruptions into windows instead of taking them as they come.

1. For three days, keep a tally. Every interruption gets a tick in one of a few columns: questions from colleagues, tickets or requests, messages, manager asks, genuine emergencies.
2. Look at the columns. Almost always, one is large and nearly all of it could have waited an hour. That column becomes a batch.
3. Give the batch a window. Requests get processed at 10, 1 and 4. Say this out loud to the people who generate them, and put it in your status.
4. Protect the gaps between windows for the work that needs setup. Two blocks of an hour are worth more than six blocks of twenty minutes.
5. Keep one column, the genuine emergencies, unbatched. Knowing that real urgency still gets through is what lets everyone, including you, relax about the rest.

Batching is a change to when, not whether. Nobody gets a slower answer than they would have on a normal busy day. They get a predictable one.

A re-entry note so you can find your way back

This is the cheapest technique in the post and the one people skip. Before you turn to the interruption, spend ten seconds writing a line for your future self: what you were doing, the next concrete step, and where the thing is.

Section 3 of the budget doc, next: add the Q3 numbers from the finance email, tab is open. That is the whole note. It goes on a sticky on the monitor, in a pinned note on the desktop, or as the first line of the document itself.

When you come back, you read the line instead of re-reading the work. The setup rebuilds from the note rather than from scratch, and re-entry drops from ten minutes to one. If the interruption is so fast you cannot write the note, say one second and write it anyway. People wait ten seconds without noticing.

The same note works at the end of the day. Leaving a next-step line on every open task means tomorrow starts with instructions rather than archaeology.

Office hours and ask me at 2 as a boundary

A batch window for requests only works if people know about it, and the way to make them know is to name it. Office hours is a phrase people understand from school and will accept from a colleague.

• Block two half-hours a day as office hours in your calendar and make them visible. Anyone can book or drop in. Outside them, questions go to a message you will answer in the next window.
• Use a single reply for everything that arrives outside a window: Ask me at 2. It is short, not rude, and gives a time. Most questions will have solved themselves by 2.
• If someone insists it cannot wait, ask one question: What breaks if this waits an hour? Sometimes the honest answer is nothing, and they will say so.
• Tell your manager the plan before you start, framed as a way to give better answers rather than fewer. Managers rarely object to office hours; they object to surprises.

Expect a week of testing. People will come by anyway, you will hold the line gently, and by the second week the traffic moves to the windows on its own.

Roles that are all switching: support, nursing, dispatch

Some jobs cannot be batched because switching is the job. A support desk, a nursing shift, a dispatch console, a front-of-house role: the interruptions are the work, and there is no report waiting underneath. The advice changes.

• Stop trying to protect deep work inside the shift. Put the deep work, charting, reports, follow-ups, in a protected block at the start or end, and treat the middle as pure switching.
• Externalise every open loop immediately. A running list on paper of what is pending, who you are waiting on and what you promised. In an all-switching role, nothing lives in your head for more than a minute.
• Use closed-loop handoffs. When a task moves to someone else, say it, write it, and confirm it. The handoff is where switching roles lose things.
• Build micro-resets. Thirty seconds between calls or patients to look at the list, breathe, and clear the last context. The reset is what stops the switching from compounding into a fog by hour six.
• Take the breaks you are entitled to, fully away from the console. Many people with ADHD in these roles skip breaks because stopping feels harder than continuing, and pay for it later in the shift.

Asking for fewer interruptions as an accommodation, and this week

If the switching load is high enough that your work suffers no matter how well you batch, it is a legitimate thing to raise formally. Accommodations for ADHD can include a quieter workspace, agreed interruption-free periods, written rather than verbal task assignment, and additional short breaks to reset. Our answer on asking for extra breaks at work because of ADHD covers the breaks piece specifically, and the request pattern is the same for the others.

Make the request concrete. Two protected ninety-minute blocks a day, and requests in writing is something an employer can say yes to. Fewer interruptions is not.

For this week: run the three-day tally, pick one batch window and announce it, put a sticky note by the monitor for re-entry lines, and leave a next-step line on every open task before you log off on Friday.