DSM-5 ADHD Criteria: What They Ask, and What They Can't See
By Valerio Cozzi, founder of Omniprism. Omniprism began after my own ADHD diagnosis at 42.
To meet the DSM-5 criteria for ADHD, a person needs a persistent pattern of inattention, hyperactivity-impulsivity, or both, lasting at least six months, with several symptoms present before age 12, evidence of the pattern in two or more settings, and clear functional impairment that isn’t better explained by something else. Children and adolescents under 17 need at least six symptoms in a domain. From 17 onwards, five is enough.
That’s the whole gate. Six numbers and a clinician’s judgement.
I met that gate at 42, without ever having read it. Nobody hands you the criteria before they’re applied to you. You go in with a life that hasn’t been working the way you expected, and you come out with a category — and only afterwards do you go and find out what the category was actually built to do.
So this article does two things. It sets out what the criteria ask, precisely. And then it says what they can’t see, because that second part is what nobody told me and it turned out to matter just as much.
The five criteria, in plain terms
The DSM-5 organises ADHD into five criteria, labelled A through E. All five must be satisfied.
| Criterion | What it asks |
|---|---|
| A | Enough symptoms in at least one of the two domains — six for under-17s, five from 17 up — sustained for six months or more, and out of step with the person’s developmental level |
| B | Several of those symptoms were already present before age 12 |
| C | The symptoms show up in two or more settings — home and school, work and relationships |
| D | There is clear evidence the symptoms interfere with social, academic or occupational functioning |
| E | The picture isn’t better accounted for by another condition, and doesn’t occur only during a psychotic disorder |
The two symptom domains contain nine items each. Inattention covers things like losing track of details, difficulty sustaining attention, trouble organising multi-step tasks, avoiding sustained mental effort, losing necessary items, and forgetfulness in daily routines. Hyperactivity-impulsivity covers physical restlessness, difficulty remaining seated or staying quiet, excessive talking, answering before questions finish, difficulty waiting, and interrupting.
I’ve described these rather than reproduced them. The DSM-5 wording is the American Psychiatric Association’s copyrighted text, and the full item lists are available free from the CDC and the NCBI Bookshelf, which is where you should read them if you want the exact language a clinician will be working from.
Criterion C is the one people underestimate. Struggling in one place — a bad classroom, a badly designed job — is not ADHD. The requirement for two settings exists to separate a person from a situation.
Where the line sits, and why it moves
Criterion A is a counting exercise, and counting is where the trouble starts.
The DSM-5 lowered the adult threshold from six symptoms to five, on the reasoning that some symptoms genuinely soften with age while the underlying condition does not. It also moved the onset requirement from before age 7 to before age 12, because the evidence showed little clinical difference between people identified at six and people identified at ten — and because almost no adult can reliably reconstruct their own behaviour at age seven.
| DSM-IV | DSM-5 | |
|---|---|---|
| Adult symptom count | 6 in a domain, all ages | 5 from age 17 |
| Age of onset | Several symptoms before 7 | Several symptoms before 12 |
| Two-settings rule | Impairment in two settings | Symptoms present in two settings |
| Adult examples | None | Written into the items |
| Alongside autism | Not permitted | Permitted |
That last row changed a lot of lives. Under DSM-IV, an autism diagnosis blocked an ADHD diagnosis outright. A generation of autistic people were told their attention and impulse difficulties were simply part of autism, and were left without the interventions that would have helped.
The DSM-5-TR, published in 2022, is the current edition. It left the ADHD criteria themselves essentially untouched and updated the surrounding text and prevalence data. If a clinician says DSM-5 or DSM-5-TR, they mean the same gate.
If you’re not American, you may be measured differently
This is the part almost no English-language guide mentions, and it matters to most of the people reading this.
The DSM is the American Psychiatric Association’s manual. Across most of Europe, and much of the rest of the world, clinicians work from the WHO’s ICD instead. The two systems are not identical, and the differences are large enough to change an outcome.
Under ICD-10, ADHD as such did not exist. The category was hyperkinetic disorder, coded F90, and it required inattention and hyperactivity-impulsivity together. That maps roughly onto what the DSM-5 calls the combined presentation — which means the predominantly inattentive presentation had no ICD-10 equivalent at all. The onset requirement was also earlier. If you were the quiet, disorganised, internally chaotic child who never climbed the furniture, ICD-10 had no name for you. That is a substantial part of why inattentive ADHD went unrecognised across Europe for decades, and why so many of us arrive at diagnosis in our forties.
ICD-11, in effect since 2022, closes most of that gap. It uses the name ADHD, code 6A05, moves it into neurodevelopmental disorders, recognises the inattentive and hyperactive-impulsive presentations separately, and permits co-occurrence with autism. It is far closer to the DSM-5 than its predecessor was.
But one difference survives, and it is the interesting one. ICD-11 does not set a fixed symptom count. It describes the condition and asks the clinician to judge whether the presentation fits, rather than requiring a tally of six or five. The same person, assessed carefully by two competent clinicians on two continents, can land on different sides of a line in one system and not encounter a line at all in the other.
Both systems are trying to describe the same reality. Only one of them pretends the reality has an edge.
The same pattern at three ages
The DSM-5 items are written broadly so they apply across a lifespan, which means “fails to follow through” has to cover a seven-year-old and a thirty-five-year-old at once.
In children, it is visible and external. Unfinished worksheets, lost permission slips, getting up during circle time, a bag that functions as a bin. Teachers usually raise it first, because a classroom demands sustained attention in a way that home routines often don’t.
In adolescents, the hyperactivity moves inward. The child who ran around the room becomes the teenager who taps a pen and feels wrong sitting still. Inattention gets harder to spot, because adolescents can burn effort to compensate — and the compensation itself hides the pattern.
In adults, it is almost entirely internal. Not running around a room but a constant mental restlessness. Chronic lateness. Projects begun with real energy that stall in the tedious middle. Bills paid late because the task cannot be started, not because the money isn’t there. A gap between capability and output that never quite closed. Research on lifespan presentation finds this consistently: adult symptoms are subtler, more internalised, and easier for everyone including the person themselves to explain away.
Many adults who seek assessment for the first time were the ones who coped. They compensated through school, through university, sometimes through a whole career — until a promotion, a child, or a bad year removed the slack that had been holding it together.
What the criteria get right
It is fashionable to be dismissive about diagnostic manuals. I’m not going to be.
The criteria are a real achievement, and their strongest feature is not the symptom lists — it’s the guardrails. The six-month duration, the two-setting requirement, and Criterion E exist specifically to stop a hard week, a hostile workplace or a mismatched classroom from being converted into a disorder. Inattention and restlessness are not distinctive. Anxiety produces them. Depression produces them. So does chronic sleep deprivation, thyroid disease, grief, a learning disorder, and a life with too much in it.
Criterion E is the sentence that makes the whole thing medicine rather than description. A clinician who counts symptoms without ruling out alternatives isn’t following the criteria — they’re skipping most of them.
There is also this: a shared, specified definition means a diagnosis in Lisbon and a diagnosis in Turin refer to the same thing. Research can accumulate. Treatments can be compared. Accommodations can be argued for in a school or a workplace with something behind them. That is not bureaucracy. That is what makes help transferable between people who have never met.
The label helped me. Once I had a name for the pattern, twenty years of evidence rearranged itself into something legible. That is not a small thing to be given.
What they can’t see
And yet.
What the criteria describe is a boundary. What they are describing is not shaped like one.
Attention, restlessness, impulse regulation, sensory tolerance, executive load — none of these are conditions you have or don’t have. They are dimensions everybody sits somewhere on. A diagnostic threshold takes a continuum and draws a line across it, then treats everything on one side as a kind of thing and everything on the other side as ordinary life. The line has to go somewhere. But the person at five symptoms and the person at four are not different in kind, and the manual has no vocabulary for saying so.
This is exactly what the shift from six to five for adults reveals. Overnight, a large number of people who had been outside the category were inside it. Nothing about them changed. The line moved.
Three specific things fall through:
The people who compensate. Criterion D asks for clear impairment. Someone who has spent thirty years building elaborate coping structures may show very little visible impairment while paying an enormous private cost to maintain it. The exhaustion of holding it together is not a listed symptom.
The people who are masking. Symptoms suppressed in front of others still count as absent when the assessment relies on what others observed. The two-setting requirement, which does such good work at separating a person from a situation, works against the person who has learned to disappear in every setting.
The people just under. Four of six inattention symptoms, with a life that isn’t working, is a real and common position. The manual’s answer is that you do not have the condition. That answer is correct and it is useless.
None of this means the categories are wrong. I don’t think they are. I think they are an honest attempt to impose usable structure on something that is genuinely prismatic — a continuous spread of traits that shade into one another, where any two people with the same diagnosis can have almost nothing in common. The category helps because it makes the thing discussable. It harms because a finite label, once applied, quietly persuades everyone that the thing it names is finite too.
Hold both. The criteria are a good tool used at a threshold. They were never a description of a person.
What actually happens in an assessment
No checklist diagnoses ADHD, and the CDC and NIMH are both explicit about this. A real evaluation involves a clinical interview covering current symptoms and developmental history; a reconstruction of childhood, usually from school records and family recollection, to satisfy Criterion B; structured rating scales such as the ASRS for adults or the Conners and Vanderbilt scales for children; collateral reports from someone who sees you in a second setting; a medical review to exclude thyroid problems, sleep disorders, vision and hearing issues and medication effects; and finally a formulation against the criteria, with a presentation specifier and a severity rating.
Scales support that judgement. They do not make it.
One thing that should not wait: if ADHD-like symptoms appear alongside significant mood instability, self-harm, or a sudden collapse in functioning, those need assessing first. Bipolar disorder and severe depression can present with distractibility and impulsivity, and stimulants given before mood is stabilised can make things worse.
What to bring
-
School reports or teacher notes from childhood — the single most useful document for Criterion B
-
A written timeline: when each difficulty first appeared, and in which settings
-
Two people who can speak to what they’ve observed, in different contexts
-
Current medications and any previous mental health diagnoses
-
Concrete examples of impairment — the missed deadline, the specific argument, the specific failure
Write the timeline before you go. Nobody reconstructs twenty years of their own history under pressure in a first appointment.
What Omniprism is for
I built Omniprism because I arrived at my own assessment with nothing organised, and wasted most of it trying to remember my own life out loud.
The self-assessment covers 14 domains, including ADHD traits, and gives you a profile in plain language. It is free, it needs no account and no email address, and it runs entirely in your browser — your answers are never sent to a server. There are guided versions for children aged 7–10 and young people aged 11–13, designed for a parent to complete from what they have observed. The optional Specialist Briefing, which summarises your profile in a format built for a clinical conversation, is a paid add-on.
It cannot tell you whether you have ADHD. Nothing that runs in a browser can. What it can do is make sure that when someone qualified asks you the questions in this article, you already know your own answers. You can read the assessment methodology before you start.
Sources
-
Diagnosing ADHD — Centers for Disease Control and Prevention
-
ADHD: What You Need to Know — National Institute of Mental Health
-
DSM-5 Diagnostic Criteria for ADHD — American Academy of Family Physicians clinical table
-
Differences between DSM-5-TR and ICD-11 revisions of ADHD: a commentary
-
ICD-11 for Mortality and Morbidity Statistics, 6A05 Attention deficit hyperactivity disorder — World Health Organization (add the direct WHO ICD-11 browser link before publishing)
This article is educational. It is not medical advice, not a diagnosis, and not a substitute for assessment by a qualified clinician.