Remote Assessment

Assessing Adult ADHD Over Telemedicine: What the Camera Keeps and What It Loses

Adult ADHD can be diagnosed over telemedicine, and a careful video evaluation is not a lesser diagnosis with an asterisk. What separates a real remote assessment from a prescription mill is rigor, not resolution.

Shafi Lodhi, MD · September 2026

Yes. Adult ADHD can be diagnosed over telemedicine, and a video evaluation done carefully is not a lesser diagnosis carrying an asterisk. The camera keeps most of what matters: the history, the concrete examples, the way a person organizes a narrative when nothing is forcing them to, speech rate and prosody, facial expression, the restlessness that lives above the desk. What it loses is real but bounded — lower-body movement, the whole-room behavioral sample, the waiting room, some of the paralinguistic texture of ordinary conversation. Those losses have workarounds, and the workarounds are not exotic. What makes a virtual ADHD evaluation worthless is almost never the medium. It is the fifteen minutes.

The reputation problem

Virtual ADHD care earned its bad name honestly. A patient fills out a checklist on a website, speaks with a prescriber for a quarter of an hour, and leaves with a stimulant prescription and a diagnostic label they will carry for decades. I have read enough of those records to say where the failure occurred, and it had nothing to do with frame rate. Nobody asked for a single concrete example. There was no developmental history beyond a yes-or-no question about childhood. No collateral informant was considered. The differential was never constructed, so the anxiety, the insomnia, the two decades of undertreated depression, and the seventy-hour weeks of unstructured knowledge work never got weighed against each other. That evaluation would have been just as thin in an office with good lighting and a comfortable chair.

So the question splits in two. Can the medium carry a rigorous adult ADHD evaluation? Yes, with adjustments I will describe. Was this particular evaluation rigorous? A separate question, and clinically the only one that matters. Video did not cause the fifteen-minute assessment. Video made it cheap to scale.

What survives the camera

The most diagnostic material in an adult ADHD evaluation is not an observation at all. It is a story, told in detail, in the patient's own words, about specific occasions. Whether it arrives through a fiber optic cable or across four feet of carpet changes nothing about its evidentiary weight. When a patient tells me she read the same lease clause eleven times last Tuesday because her eyes kept sliding to her open email, then signed it without understanding it, the diagnostic content is fully intact on video. History, functional impairment, developmental trajectory, compensation strategies — all medium-independent.

Some things are easier on screen. A video call is a close-up; I see more of a face on a fourteen-inch display than across a desk at six feet, and expressions that would be marginal in the room are plain — the flicker of recognition when I name a symptom the patient has never had words for, the look of someone assembling a plausible narrative rather than retrieving a memory.

Speech carries well. Rate, volume modulation, self-interruption, the mid-sentence abandonment of one clause for a more interesting one. Tangentiality is fully visible: the patient who answers my question about college with a four-minute account of a roommate's car, then looks at me expectantly, having lost track of what was asked. Narrative organization transmits perfectly, and it is among the richest sources of data in the evaluation — not what the patient says about being disorganized, but whether the account of their life arrives in any usable order.

Upper-body restlessness survives, and there is more of it than clinicians expect. Hands. Repositioning. The pen, the water glass, the hair, the sleeve. The patient who adjusts the laptop nine times.

What the camera loses

The frame ends at the sternum, and a great deal of adult hyperactivity lives below it. In the room I can see that a leg has been going for forty minutes, that a heel is tapping, that the body has slid sideways in the chair by degrees. On video I get none of that unless I ask.

A composite, drawn from a pattern rather than any one person: a man in his late thirties, articulate, self-employed, presenting on video as notably still. Around minute seventy I noticed a faint periodic tremor in the image and attributed it to the laptop fan. It was his heel, striking the floor hard enough to vibrate the desk the computer sat on. He had been doing it since we started, he said, and in every meeting of his adult life. In an office I would have logged it in the first two minutes.

The whole-room behavioral sample disappears. How a person handles paper. Whether they can find the form they were sent. Whether they get up, whether they pace when a question is hard. The waiting room goes with it, which is a larger loss than it sounds. Some of my most useful data used to arrive before I entered the room: arrival time, whether the forms were done or half-done, what they brought, what they forgot.

Paralinguistic cues degrade in a specific and irritating way. Latency becomes ambiguous — a two-second pause before an answer means something in person, and on video it might be the network. Overlapping speech, one of the better natural measures of impulsivity, is distorted by transmission delay, so the patient who talks over me may be impulsive or half a second behind. I hold those observations loosely rather than throw them away. And one loss rarely discussed: on video a patient can have notes beside the keyboard. Preparation is not a sin, but spontaneity is part of what I am measuring, and a prepared list changes what a first answer means.

The camera does not make a diagnosis weaker. Brevity does.

The room is both signal and noise

The patient's environment is genuinely informative. I have learned things from a visible desk that an hour of self-report would not have produced — the stacks, the sticky notes ringing a monitor, the whiteboard with three competing systems on it. Where a person takes the call from tells me something too: the car in a parking lot because home was impossible, the closet, the bed.

It is also noise. A staged or blurred background removes the environmental data entirely, and a virtual background removes it while telling me only about the person who chose it. More subtly, home turf cuts both ways. Some patients disclose more from their own living room than they ever would in an office, and the reduction in situational anxiety is clinically valuable: less masking, a more representative baseline. But the same environment is scaffolding. A patient at their own desk, with their own systems within reach and their phone timer already set, will look more organized than they are. I name that in my reasoning rather than pretend the setting is neutral.

Adaptations that restore rigor

The corrections are unglamorous. First, extend the observation window. If a minute of video carries less behavioral information than a minute in the room, add minutes. Restlessness that declares itself in the first twenty minutes of an office visit may take ninety to surface on screen, and fatigue effects — the decay in attention and narrative coherence as the interview wears on — are among the most useful remote signals available. Evaluation length is its own argument, which I make elsewhere; the methodology answers in the FAQ cover session structure and breaks.

Second, give explicit environmental instructions in advance, in writing, and repeat them the day before. Camera far enough back to show the torso and, where the room permits, the lap. A real chair at a real surface, not a bed. No virtual background and no blur. Phone in another room. A door that closes. Ninety uninterrupted minutes, alone, no notes. Then treat every deviation as data rather than an inconvenience. The patient who cannot assemble those conditions after two reminders has told me something, and so has the patient who arranges them precisely.

Third, let collateral informants carry more of the load. When direct observation is thinner, independent corroboration should be heavier, and remote practice makes collateral easier rather than harder — there is no second office visit to schedule. The companion collateral forms that ship with the LODHI-A come in a partner version and a parent version for exactly this reason, and in a video evaluation I weight them accordingly. A spouse who describes what the patient's feet do at dinner has restored some of what my frame cut off.

Fourth, make the observation deliberate instead of incidental. In person, behavioral observation is ambient; you notice things because they are in front of you. On video it has to be structured — decided in advance, sampled at specified points, recorded as it happens. The LODHI-A ships with a telemedicine adaptation guide covering that, along with what can and cannot be credibly assessed through a screen. Part of it is permission to ask directly: What are your feet doing right now? Patients answer that honestly, and often with surprise.

Why defensibility matters more at a distance

Stimulants are controlled substances, and the rules around prescribing them remotely have been unsettled for years. I will not offer legal specifics; those vary and change. The clinical point stands regardless. An adult ADHD evaluation produces a document other people will read — a future prescriber, a credentialing body, an accommodations office, sometimes an insurer, occasionally a court. Remote evaluations attract more skepticism as a class, some of it earned by the prescription mills and none of it applied selectively to yours.

The defense is the record. Duration documented. Concrete patient-generated examples quoted rather than summarized as criteria endorsements. Developmental history with real anchors. The differential constructed and then explicitly resolved. Collateral sought, obtained or documented as unavailable, and weighed. And — the part clinicians resist — the limits of the modality named plainly. I write down what I could not observe. A note that names its own blind spots reads as more credible, not less, because the alternative implies a thoroughness the medium did not permit.

When I insist on in person

I move an evaluation into the office when external incentive is significant: disability claims, forensic matters, litigation, high-stakes accommodations. Not because I assume dishonesty, but because when incentive is present I want control of the environment and the whole behavioral field, and I want to be able to say I had them. I insist when formal cognitive or neuropsychological testing is indicated, which is a different enterprise with its own standardization requirements. I insist when a patient cannot establish a workable setup and the alternative is ninety minutes of frozen video. I insist when self-report is thin and no collateral is available, so that observation must carry weight it cannot carry remotely. And I insist when the differential turns on something I need to see across a room — significant psychiatric instability, active substance use, sensory or cognitive impairment that makes a screen a poor channel.

The rest of the time, which is most of the time, video is adequate, and the honest reason to prefer an office is convention, not diagnostic superiority. The burden falls where it always falls: on whether the evaluation asked enough, for long enough, of enough people. If you assess adults for ADHD and want to see how the structure handles this, telemedicine guide included, the instrument is free to licensed clinicians and you can request access. The background on why I built it explains the rest.

More articles · FAQ · Request the instrument