Can telehealth diagnose ADHD?

The scale of the change

Between March 2020 and mid-2024, telehealth became the modal diagnostic experience for adult ADHD in the US. Commercial-claims data shows the share of new adult ADHD diagnoses delivered via telehealth rose from under 5% pre-pandemic to a peak above 50% during 2021–2022, settling into the 35–45% range across 2024–2025 per Cunningham et al. 2024 in JAMA Psychiatry⁵ and corroborating sources. Half the readers of this article were diagnosed without sitting in a room with the evaluator. The article treats that as the default case.

The DEA telehealth keystone

The regulatory pillar holding up the entire telehealth-ADHD prescribing model is the Drug Enforcement Administration (DEA) extension of COVID-era telemedicine flexibility for controlled-substance prescribing. The underlying law — the Ryan Haight Act 2008 (21 USC §829(e))⁴ — requires at least one in-person medical evaluation before a prescriber can issue a Schedule II controlled substance via telehealth. The US Department of Health and Human Services (HHS) waived this requirement on March 16, 2020, under the COVID public health emergency. The DEA has extended the waiver four times since. The most recent extension (November 2024)³ runs telehealth Schedule II prescribing through December 31, 2025. The DEA’s proposed Special Registration Final Rule remains unfinalized as of Q1 2026.

The practical implication: without this waiver, telehealth ADHD platforms cannot legally prescribe stimulants fully remotely. The waiver is the load-bearing piece of the business model. What happens when (or if) it expires shapes whether anyone’s current telehealth ADHD prescription continues without an in-person visit.

Cerebral, Done, and the platform record

Cerebral. The Federal Trade Commission (FTC) April 2024 proposed consent order¹ requires Cerebral to pay $7 million in consumer refunds and prohibits specific marketing claims about clinical efficacy and prescriber availability. The FTC alleged Cerebral made it intentionally difficult to cancel subscriptions and that its marketing overstated the quality of its evaluations. Separately, Cerebral was the subject of DEA, Department of Justice (DOJ), and HHS Office of Inspector General (HHS-OIG) investigations beginning May 2022 into its stimulant prescribing patterns; in October 2023 it entered a $3.7 million civil settlement with the DOJ over Controlled Substances Act violations.

Done Global (“Done.com”). Founders Ruthia He and David Brody were indicted by the DOJ on June 13, 2024² for an alleged scheme to distribute Adderall and other stimulants outside the usual course of professional practice. The indictment alleges they structured the platform to maximize prescription volume — capping initial visits at 30 minutes, paying prescribers per script rather than per visit, and instructing prescribers against scheduling follow-ups. This is the first federal prosecution of telehealth-platform leadership for ADHD-prescribing-related controlled-substance offenses. The indictment text itself is the public record of what a non-defensible telehealth ADHD evaluation looks like.

What a defensible evaluation contains

Six components, identical whether the evaluation is in-person or video. A 15-to-30-minute intake cannot complete the list in the modal case — the diagnostic question is which components got dropped. Grounded in the American Psychiatric Association’s DSM-5-TR, the American Professional Society of ADHD and Related Disorders’ APSARD 2024 US adult ADHD guideline¹¹, the National Institute for Health and Care Excellence (NICE) guideline NG87, and the Canadian ADHD Resource Alliance (CADDRA) 4th edition.

• Structured DSM-5-TR criteria assessment across both childhood (before age 12) and current presentations, with evidence from at least two settings, and documented functional impairment.
• Collateral history where possible — parent, partner, sibling, old report cards. The pre-12 evidence is the part most adult evaluations lean on collateral for.
• Comorbidity screening — depression, anxiety, bipolar disorder, substance use, sleep disorders, trauma. The differential matters because the treatment for ADHD vs bipolar depression vs anxiety with poor concentration is different.
• Rule-out of medical mimics — thyroid, sleep apnea, iron deficiency, B12 deficiency, perimenopausal cognitive symptoms. Often skipped in short telehealth intakes.
• Medication and family history, including family ADHD history given ~74% heritability.
• Validated rating scale as adjunct— Adult ADHD Self-Report Scale (ASRS) v1.1, Conners’ Adult ADHD Rating Scale (CAARS), Diagnostic Interview for Adult ADHD (DIVA-5), or Barkley Adult ADHD Rating Scale (BAARS-IV). As adjunct, not as the diagnostic itself.

ASRS and DIVA-5: what they actually do

Kessler et al. 2005⁷ developed and validated the Adult ADHD Self-Report Scale for the WHO World Mental Health Survey. The 6-item screener catches most cases that turn out to have ADHD and rarely flags people who don't. The 18-item full version is the symptom checklist. The ASRS v1.1 update (Ustun et al. 2017)⁸ refined the cut-points but didn’t change the role: it’s a screener.

The DIVA-5 is a semi-structured clinical interview based on DSM-5 criteria, designed to be administered by a trained clinician. The patient-facing version exists as preparation for the appointment, not as the diagnosis itself.

Telehealth platforms that present an ASRS or self-administered DIVA result asthe basis for diagnosis are inverting the tools’ intended use. Both instruments were validated as components of a clinical evaluation, not as substitutes for one.

What a 28-minute intake can and can’t do

The data on synchronous-video adult ADHD evaluation is that the modality works when the evaluator and protocol are equivalent. Yang et al. 2022⁶ and Cunningham et al. 2024⁵ both support concordance between in-person and telehealth evaluations conducted by trained psychiatrists or psychologists using comparable protocols. The question is never “does video work” — it does. The question is “what minimum protocol was followed.”

A 28-minute intake by a nurse practitioner under productivity quotas, paid per script rather than per visit, who never sees the patient again, is not the modality that those concordance studies validated. It can confirm self-reported ADHD symptoms; it cannot meaningfully differentially diagnose, cannot screen comorbidity in depth, cannot integrate collateral childhood history, cannot rule out medical mimics, cannot calibrate a treatment plan to the person’s history. The output may be a correct diagnosis. It may also be a misattribution of anxiety, trauma response, long-COVID cognitive symptoms, perimenopausal symptoms, or bipolar II depression — each of which has overlapping presentation and a different treatment pathway.

What to do if a platform under investigation diagnosed you

A platform’s regulatory trouble does not retroactively invalidate the symptom picture in your life. Three moves cover the actual decisions: keep the diagnosis if treatment is working, get a second evaluation if anything is off, and preserve the records either way.

One — keep the diagnosis if the treatment is working and you trust the symptom picture.If you’ve responded to stimulants in the expected way, the functional impairment was real before the diagnosis and improved after, and you don’t have a treatable mimic, the diagnosis is probably right. Confirm it the next time you see a clinician who can do a proper evaluation; don’t panic-reverse it.

Two — get a second evaluation if any of these apply. Treatment response wasn’t what was expected. You have meaningful comorbidity (mood disorder, trauma history, sleep problems) that wasn’t addressed. You have signals of a differential (cyclical mood, severe early-life trauma, sleep apnea risk factors, perimenopausal timing). The second evaluation can be with a psychiatrist, a psychologist with adult-ADHD specialisation, or a developmental-behavioural physician. Insurance coverage for second opinions is usually less hostile than initial assessment.

Three — keep your records.The evaluation notes, prescription history, treatment response. If the DEA waiver expires or the platform shuts down (Done.com’s indicted founders are not running it anymore), continuity of care depends on the next prescriber having a paper trail.

What’s contested

The pre-12 onset criterion (DSM-5-TR) is the most contested piece of adult diagnosis. Sibley et al. 2018 in AJP⁹ documented “late-onset” presentations that meet adult criteria but lack clear pre-12 evidence; APSARD 2024 emphasises functional impairment over rigid age cutoff. A telehealth evaluator who rejects a diagnosis on strict pre-12 grounds may be following older guidance; one who confirms without any developmental history may be skipping the most contested criterion entirely.

Telehealth-prescribing volume itself remains contested in the clinical-quality literature. The good telehealth platforms exist and do thorough evaluations. The indicted ones existed and didn’t. Discriminating between them on a public website is beyond the article’s scope; the regulatory record is the clearest signal currently available, and that’s the signal the article points readers at. The reasonable position: the modality is fine, the platform-level protocol design is what determined evaluation quality, and the FTC/DOJ/DEA record is the public document for which platforms got the protocol wrong.