Differential Diagnosis

ADHD or Anxiety: Reading the Shape of the History

Anxiety and ADHD produce nearly identical complaints in adults. The differential rarely turns on which symptoms are present; it turns on when they started, whether there were ever good years, and what the mind is doing at the moment attention fails.

Shafi Lodhi, MD · September 2026

Anxiety and ADHD arrive in my office sounding the same. Both patients say they cannot concentrate, cannot sit still, cannot finish what they start. The differential almost never turns on the symptom list — it turns on the shape of the history. ADHD is trait-like: present in some form since childhood, continuous, worst under tedium, and content-free. Attention simply goes. Anxiety is usually episodic: it has an onset, it has better years, it sharpens under evaluation, and its inattention has content. Attention goes somewhere, and that somewhere is a worry. So the questions that matter are when this started, whether there was ever a stretch without it, and where the mind actually travels when it leaves. Then plan for the fact that both conditions are frequently present at once.

What anxiety does to attention

Anxiety is not a disorder of mood that happens to be unpleasant. It is a disorder of allocation. A person who is worried is running a background process — scanning for threat, rehearsing outcomes, drafting contingencies — and that process consumes exactly the resources that reading a contract or following a meeting requires. The result is a genuine attentional deficit. Not a complaint, not a metaphor. The patient really cannot hold the sentence.

Everything else follows. Restlessness, because sustained arousal has to go somewhere: the leg bouncing, the standing up, the inability to sit through a film. Procrastination, because the task is aversive and avoidance relieves that immediately. Forgetfulness, because encoding fails when working memory is already occupied. Irritability, poor sleep, the sense of being scattered all day. Every one of these maps onto a DSM-5 inattention or hyperactivity item. A patient with a generalized anxiety disorder can endorse most of the adult ADHD criteria honestly, and will, if you ask in a way that permits it.

Which is the real problem with brief screeners in this population. They ask whether the symptom is present, not what produces it, and the anxious patient answers yes truthfully. A high ASRS score in an anxious adult tells you that something is interfering with attention. It does not tell you what, and treating the score as though it did is how ADHD gets diagnosed in a room where the actual finding was untreated panic.

The longitudinal key

The single most useful distinction is developmental. ADHD is a neurodevelopmental condition; DSM-5 requires that several symptoms be present before age twelve. But the criterion is usually applied too thinly — clinicians ask whether the patient had trouble in school, accept a yes, and move on. Anxious children also have trouble in school.

What I want instead is continuity. ADHD does not have good years. It has well-scaffolded years. The patient who had a structured household, a small school, a sport that burned off the restlessness, and a mother who checked the backpack every night may have looked fine until the scaffolding came off — at college, at a first job, at the birth of a second child. That is not late onset. That is a fixed trait meeting a rising demand, and the history will show the trait in the margins of the good years if you look: the homework that took four hours instead of one, the room that was never once clean, the teacher comments about potential.

Anxiety, by contrast, usually has edges. There was a year that was bad and a year before it that was not. There was a job, a diagnosis, a breakup, a move, a pandemic. I ask directly: Tell me about the last time you felt like your attention was working. A patient with primary anxiety can often name the period, sometimes the month. A patient with ADHD looks at me as though I have asked about a country they have never visited. They may name an era of better function — the year with the good boss, the semester with the four-hour exam schedule — but not an era of different wiring.

This is what I mean by temporal triangulation, and it is why the structured instruments that front-load symptom checklists leave the hardest work undone. The diagnostic information is in the course, not the cross-section.

Anxiety takes attention somewhere specific. ADHD lets it go nowhere in particular.

Probes that actually separate them

Where does your mind go?

This is my highest-yield question, and I ask it about a specific failure the patient has already described. Not do you lose focus but: you said you reread that email four times — during the rereading, what was in your head? The anxious patient answers immediately and in detail. They were thinking about how the recipient would take the second paragraph. They were thinking about the deadline, the money, their mother. The content is retrievable because worry is vivid and self-aware.

The patient with ADHD often cannot answer. They were not thinking about anything in particular. They noticed a sound, or a tab, or nothing at all, and some number of minutes later they came back and had no idea they had gone. The absence of retrievable content is itself the finding. Anxious inattention is an occupied mind. ADHD inattention is a vacated one.

Procrastination and what it is made of

Both groups procrastinate; the texture differs. Anxious avoidance has an object. There is dread attached, usually to an imagined outcome — the criticism, the mistake, the exposure — and the patient can tell you what they are afraid will happen if they open the file. ADHD initiation failure is flatter and stranger. The patient wants to do the task, has no fear of it, may even find it interesting, and still cannot start. What they describe is not dread but an absence: nothing turns on. When I ask what would happen if they did it badly, the anxious patient has a vivid answer and the ADHD patient shrugs.

Errors

Anxiety generates checking. The work goes out late because it was reviewed six times. ADHD generates uncorrected errors: the wrong date, the missing attachment, the transposed digits, discovered by someone else. A history of meticulous, slow, anxious accuracy is evidence against inattentive ADHD — and a history of careless errors that persist despite genuine effort and real consequences is difficult for anxiety alone to explain.

Sleep

Both delay sleep onset, and again the content differs. Anxious patients lie down and rehearse: the conversation, the bill, the symptom they googled. Patients with ADHD more often describe a mind that will not settle but is not distressed — branching from one interesting thing to another, or a sudden productivity at eleven at night that they cannot summon at two in the afternoon. Ask what the thoughts are about. The answer sorts more cleanly than the complaint does.

Caffeine and performance context

I ask about caffeine in every evaluation because it is a crude natural experiment the patient has already run. Anxious patients frequently report that more coffee makes everything worse — the heart, the hands, the sense of dread — and that they have learned to cut back. Patients later diagnosed with ADHD more often report large intakes with modest jitter and a settling effect on their thinking. I treat this as a hint and nothing more.

Performance context is stronger. Anxious impairment is state-dependent and pressure-sensitive: the presentation, the exam, the meeting with the director, the moment of being watched. The same person alone in a quiet room with low stakes may work perfectly well. ADHD runs the other way. The quiet room with no stakes is where attention dies, and acute deadline pressure is often when the patient finally performs, because urgency supplies the activation their neurology does not. When a patient tells me they only work well the night before something is due, that is not a moral failing. It is a physiological report.

A composite case

A composite from many evaluations: a thirty-four-year-old operations manager, referred after a colleague's diagnosis, describing two years of not being able to think. She cannot follow meetings. She rereads. She has started making errors in a role she has held for six years without incident, and she has begun waking at four in the morning.

The two-year frame is the first thing that matters — the errors began after a reorganization, a new supervisor, and a mother's cancer diagnosis in the same six months. When her attention fails, she can tell me precisely where it went: to a Slack message she interpreted as displeasure. She checks her work obsessively and submits it late. Caffeine, she stopped. Before all this, by her account and her husband's, she was the one who kept everyone's calendar.

Childhood is unremarkable in the relevant way: decent grades, no teacher concerns, homework finished before dinner, a tidy room. There is no trait to find in the margins. This is anxiety with secondary attentional impairment, and the correct output of the evaluation is a diagnosis she did not come in for. Had the same woman described thirty years of lost keys, an unfinished thesis, three abandoned filing systems, and a childhood in which she was the family joke about being late — with the anxiety arriving at twenty-six, on top of it all, as a reasonable response to a life of underperforming her own intelligence — the formulation would invert, and both diagnoses would be correct.

When both are present

Which is the common case. Anxiety and ADHD co-occur at rates high enough that finding one should prompt a careful search for the other, and the relationship is frequently causal in the direction clinicians forget: two decades of missed deadlines, disappointed employers, and private conviction that you are lazy is an excellent way to become an anxious adult. In those patients the anxiety is real, meets criteria, and is downstream.

The organizing question is what the history looks like underneath the anxious period. If you subtract the last three years, is there anything left? When the answer is a childhood and adolescence of trait-level inattention, both diagnoses stand, and the sequence tells you something about treatment order. When the answer is a competent, organized, unremarkable developmental history, the ADHD question is answered, and the temptation to diagnose it anyway — because the patient is suffering and the symptoms are technically present — should be resisted.

I do not think it is necessary to treat anxiety to remission before considering ADHD, and I have seen that sequencing delay care for years. But treatment response is data. An adult whose concentration normalizes on an SSRI, or after twelve weeks of decent CBT, has told you something the interview could not. Say so in the note, and leave the ADHD question genuinely open rather than quietly settled.

None of this requires a new instrument. It requires an interview organized around course and mechanism instead of endorsement — which is what I built the LODHI-A to force, after too many evaluations in which the checklist and I had confirmed each other and learned nothing. It is free to licensed clinicians; the methodology questions are answered in more detail there, and the instrument itself can be requested here. Mostly, though, I would ask one thing of the next anxious patient who tells you they cannot concentrate: ask them where their mind goes. The answer is usually the diagnosis.

More articles · FAQ · Request the instrument