Assessment Method

How Long Should an Adult ADHD Evaluation Take?

A comprehensive adult ADHD evaluation that can withstand scrutiny typically requires two to three hours of direct clinical time, and it is usually better delivered across more than one session than compressed into a single sitting.

Shafi Lodhi, MD · September 2026

Two to three hours. That is the honest answer for an adult presenting with a question of ADHD, and it is longer than most of the market currently offers. A full administration of the LODHI-A runs 120 to 180 minutes, ordinarily split across two to four sessions. The range is wide on purpose. A thirty-four-year-old with report cards from third grade, a mother available by phone, and no psychiatric history other than the present complaint may finish at the short end. A forty-one-year-old with a decade of treated depression, a heavy cannabis history, two head injuries, and no surviving informant will not, and should not, be finished in ninety minutes. The time is not ceremony. It is the only way to gather the specific material this diagnosis requires.

What the market offers, and what each end gets wrong

At one extreme sits the fifteen-minute diagnosis. A patient completes a symptom inventory online, meets a prescriber briefly, confirms that yes, they lose their keys and interrupt people and cannot finish tax paperwork, and leaves with a diagnosis and often a prescription. The encounter is not fraudulent. The symptoms are usually real. But nothing in that quarter hour distinguishes ADHD from the eleven other things that produce the same self-report, and nothing in it establishes childhood onset beyond the patient's retrospective impression, which is the single least reliable instrument in the room. A fifteen-minute encounter can confirm a hypothesis. It cannot test one.

At the other extreme sits the several-thousand-dollar neuropsychological battery — six or eight hours of testing, continuous performance measures, a report with standard score tables and a percentile for everything. This is a serious, skilled undertaking, and it is genuinely the right call when there is a question of learning disorder, acquired brain injury, or a cognitive decline that needs baselining. For uncomplicated adult ADHD it overshoots in one direction while underperforming in another. ADHD has no confirmatory test. There is no cutoff on a computerized attention task that separates the condition from anxiety, poor sleep, or a bored twenty-six-year-old sitting in a quiet room. The battery generates a large volume of data about current cognitive performance and comparatively little about the developmental trajectory on which the diagnosis actually rests. Patients emerge with a sixteen-page document and a diagnosis that still hinges on the forty-five-minute clinical interview buried in the middle of it.

Both poles make the same error from opposite directions. They treat ADHD as something to be detected — at speed in one case, with instrumentation in the other — when it is something that must be reconstructed.

Where the time actually goes

Reconstruction is slow work, and it is worth naming what consumes the clock, because clinicians who have only done the short version often assume the long version is padded.

The largest single block is developmental history. DSM-5 requires several symptoms present before age twelve, which sounds like a yes-or-no question and is not. Adults do not remember their childhoods in symptom language. They remember rooms, teachers, report card comments, the seat they were moved to, the summer they were made to redo a math workbook. Getting to age twelve honestly means walking through school years in sequence, asking what changed at each transition, asking about the specific accommodations that were made informally before anyone had a word for them — the mother who sat beside them every night through homework until high school, the teacher who let them stand at the back of the room. This takes forty minutes to an hour when done properly, and it cannot be accelerated, because the useful material surfaces late. The first answer to "how were you in school?" is almost always "fine, I think." The fourth or fifth answer is where the history lives.

The second block is differential screening, which in adult ADHD is not a formality. Inattention, restlessness, and disorganization are the common final pathway for an enormous amount of psychiatric and medical illness. Sleep apnea produces it. Untreated hypothyroidism produces it. So do bipolar II, generalized anxiety, complex trauma, early cognitive change, alcohol used nightly to fall asleep, and the ordinary consequences of parenting two children under five while working a demanding job. Screening these seriously — not with a checkbox but with enough questions to know whether the phenomenon is episodic or lifelong, situational or pervasive — takes real time. Most of that time produces negative findings, which is precisely why rushed evaluations skip it.

The third block is the elicitation of concrete examples, and this is where the format matters most. Every symptom claim has to be grounded in something the patient generated themselves, unprompted. That means resisting the temptation to supply the example, which is the most natural thing in the world to do when a patient is struggling to articulate. "Do you lose track in meetings?" is a leading question with a predictable answer. "Tell me about the last meeting you sat through" takes ninety seconds longer and yields something you can actually reason about. Multiply that across eighteen criteria and you have accounted for another substantial piece of the interview. The difference between this and a criterion-by-criterion review is largely a difference in how long you are willing to sit in silence.

The fourth is collateral. A ten-minute call to a parent, sibling, or long-term partner frequently reorganizes the entire picture, and it belongs in the time budget even though it happens outside the room.

A fifteen-minute encounter can confirm a hypothesis. It cannot test one.

The interview is itself a behavioral sample

Here is the point I find most underappreciated, and it took me a few hundred evaluations to state it plainly. A two-hour semi-structured interview is not only a data-gathering exercise. It is a two-hour observation of sustained attention under demand.

You are asking someone to hold a narrative thread, retrieve remote autobiographical detail, tolerate open-ended questions without a form to lean on, and stay with a cognitively effortful task well past the point where it stops being interesting. That is a closer analogue to the demands that actually impair adults with ADHD than anything a fifteen-minute encounter or a timed computer task can produce. I have watched patients who answered every screening item affirmatively settle in and track beautifully across two hours, which is diagnostically meaningful. I have watched others fragment visibly in the second hour — losing the question, circling back to a story they already told, asking me to repeat what I said, apologizing for it — which is also diagnostically meaningful. Neither observation is a criterion. Both inform the formulation.

A short evaluation forfeits this entirely. You cannot observe the decay of effortful attention in a window shorter than the time it takes to decay.

Splitting the sessions, and paying for them

Two to three hours in one sitting is a poor design for this population specifically. Fatigue degrades exactly the retrieval you depend on, and the developmental history done in minute 140 is measurably thinner than the one done in minute 40. I generally run the LODHI-A across two to four sessions: onset and developmental history in the first, symptom elicitation and differential in the second, collateral and feedback after. The gap between sessions is useful in its own right. Patients go home and remember things. They call a parent. They find the third-grade report card in a box. The material that arrives in session two unprompted is often the best material in the file.

Now the uncomfortable part. Time is expensive. Two to three hours of clinician time is a real cost to the practice and frequently to the patient, and I am not going to pretend that reimbursement for extended diagnostic interviewing is generous or that every clinician can absorb it. So the honest position is a conditional one. The full-length workup earns its cost when the presentation is genuinely uncertain — when there is significant psychiatric comorbidity, a substance history, an absent or unreliable developmental account, a prior diagnosis you cannot verify, secondary gain of any kind, or a stimulant already prescribed by someone else on grounds you cannot reconstruct. In those cases the two hours are not a luxury. They are the difference between an opinion and a guess.

And there are presentations where a shorter workup is defensible. A twenty-eight-year-old with intact school records documenting longstanding difficulty, a parent who confirms it on the phone, no psychiatric or substance history, no litigation or academic accommodation at stake, and a symptom picture that has been stable and pervasive since childhood does not require three hours. Perhaps seventy-five to ninety minutes, done carefully, with the developmental section intact. The section you never shorten is onset, because it is the only one that cannot be reconstructed later.

What the rushed version costs downstream

Consider a composite I see some version of several times a year. A woman in her late thirties is diagnosed in a brief encounter and started on a stimulant. She improves for three weeks, then destabilizes — irritable, sleepless, worse than baseline. She is titrated upward, then switched, then augmented. Eighteen months later a longer history reveals two discrete periods in her twenties of four or five days each with reduced sleep, uncharacteristic spending, and pressured speech, neither of which was ever asked about, and a mother who describes her as a quiet, organized, high-achieving child. The inattention was real. It was also not ADHD.

That is the first cost: a controlled substance treating the wrong condition, sometimes for years, with the diagnostic error hardening each time it is copied forward into a new chart. The second cost is subtler and I think worse. A rushed evaluation returns the patient's own hypothesis to them with a clinician's signature on it. Their story goes unexamined. Whatever was actually driving the difficulty — the apnea, the grief, the trauma, the marriage — stays undiscovered, and now there is a diagnosis in the way of finding it. I would rather tell a patient that I am not yet sure and need another hour than hand back a certainty I have not earned.

The instrument exists because I could not do this reliably from memory, and a structure that reliably produces the two hours is more useful than a resolution to be thorough. The LODHI-A is free to licensed clinicians; the FAQ covers administration and session structure in more detail, and you can request a copy if you want to see how the time is allocated before deciding whether your practice can carry it.

More articles · FAQ · Request the instrument