Home → Research → AI in Psychiatry → Telehealth in Psychiatry

Telehealth in Psychiatry: Benefits, Risks, and Oversight

By Ryan S. Sultan, MD
Assistant Professor of Clinical Psychiatry, Columbia University Irving Medical Center
March 28, 2026

The JAMA Psychiatry Research

In 2022, I published a viewpoint in JAMA Psychiatry with Manpreet K. Singh, MD, MS (then at Stanford University, now at UC Davis) and Alice W. Zhang examining how the rapid expansion of telepsychiatry during and after the COVID-19 pandemic had created patient safety risks that the existing regulatory framework was not designed to address.

The paper was not anti-telehealth. I use telehealth in my own practice. The paper was pro-standards -- arguing that telehealth psychiatry should meet the same clinical benchmarks as in-person psychiatry, and documenting what happens when it does not.

What Happened During the Pandemic

Before COVID-19, telehealth accounted for less than 1% of psychiatric visits in the United States. When the pandemic forced clinics to close, telehealth became the primary mode of psychiatric care almost overnight. CMS waived geographic restrictions, the DEA relaxed in-person requirements for prescribing controlled substances, and state medical boards issued emergency orders allowing cross-state practice.

These emergency measures were necessary and, for the most part, beneficial. Patients who would have lost access to psychiatric care during lockdowns were able to continue treatment. The speed of the transition was remarkable and, in many cases, life-saving.

But the emergency measures also created an opportunity for a different kind of operator to enter the psychiatric care space.

The Rise of Digital Psychiatry Platforms

A wave of venture-capital-backed digital psychiatry companies launched during and after the pandemic. Companies like Cerebral, Done, and others positioned themselves as technology disruptors who would democratize access to mental health care. They raised hundreds of millions of dollars from investors and scaled rapidly.

The business model was straightforward: high volume, fast evaluations, streamlined prescribing. Some of these platforms explicitly marketed themselves as places where patients could get ADHD medication quickly and easily -- language designed to attract people seeking stimulant prescriptions.

What my co-authors and I documented in JAMA Psychiatry was the clinical reality behind the marketing:

The Consequences

The consequences of this approach were predictable to anyone in clinical practice:

• Stimulant overprescribing: The DEA investigated several platforms for prescribing patterns that deviated significantly from clinical norms. Reports emerged of patients receiving stimulant prescriptions within minutes of signing up, without the comprehensive evaluation that stimulant prescribing requires.
• Adverse events: When controlled substances are prescribed without adequate evaluation and monitoring, adverse events follow. Cardiovascular complications, psychiatric destabilization, worsening anxiety, and substance misuse were reported.
• Clinical harms: Beyond medication issues, some patients with serious psychiatric conditions -- bipolar disorder, psychotic disorders, active substance use disorders -- received stimulant prescriptions that worsened their underlying conditions. These are precisely the diagnostic distinctions that a 10-minute evaluation cannot make.
• Erosion of telehealth credibility: Perhaps most damaging for the field, these platform failures gave ammunition to critics of telehealth itself. The problem was never telehealth as a modality -- it was telehealth without standards.

The Framework: What Responsible Telehealth Looks Like

Our JAMA Psychiatry paper proposed a framework for integrating telehealth into the mental health ecosystem responsibly. The core principle is simple: telehealth psychiatry should meet the same clinical standards as in-person psychiatry. Technology changes the delivery mechanism, not the standard of care.

Minimum Evaluation Standards

Before prescribing controlled substances via telehealth, the evaluation should include:

• Comprehensive psychiatric assessment: Chief complaint, history of present illness, psychiatric history, substance use history, medical history, family psychiatric history, developmental history (for ADHD), social history, and mental status examination. This takes time. There is no shortcut.
• Differential diagnosis: ADHD symptoms overlap with anxiety, depression, bipolar disorder, trauma-related disorders, substance use, sleep disorders, and thyroid dysfunction. Ruling these out -- or identifying comorbidities -- requires careful clinical reasoning, not a symptom checklist.
• Collateral information: For ADHD in particular, information from someone who knew the patient in childhood is essential for establishing childhood onset. Self-report alone is insufficient, especially in the context of stimulant diversion and misuse.
• Standardized rating scales: Validated instruments (ASRS, Conners, Vanderbilt) should supplement, not replace, clinical assessment.
• Risk assessment: Cardiac history, blood pressure, substance use screening, and psychiatric risk factors must be assessed before prescribing stimulants or other controlled substances.

Mandatory Outcome Monitoring

Prescribing without monitoring is not treatment -- it is dispensing. Our framework requires:

• Scheduled follow-up: Regular appointments to assess medication efficacy, side effects, and clinical status
• Measurement-based care: Standardized outcome measures at each visit to track symptom improvement objectively
• Vital sign monitoring: For stimulants, blood pressure and heart rate monitoring at regular intervals
• Side effect assessment: Systematic screening for appetite changes, sleep disturbance, mood changes, cardiovascular symptoms, and other known medication effects
• Functional assessment: Tracking real-world functioning (work, school, relationships), not just symptom scores

Regulatory Requirements

Our policy recommendations include:

The Benefits of Telehealth Done Right

My research critiques irresponsible telehealth, but I want to be clear: telehealth done right is genuinely beneficial for psychiatric patients. The problem was never the technology -- it was the business model built on top of it.

Access

Telehealth eliminates geographic barriers to psychiatric care. This matters enormously in a country where most counties have zero psychiatrists. Patients in rural areas, patients with mobility limitations, patients who cannot take time off work -- all of these populations benefit from the ability to see a psychiatrist without traveling to a distant office.

Continuity

Patients who move, travel, or face temporary barriers to in-person visits can maintain their treatment relationship through telehealth rather than starting over with a new provider. This continuity is clinically valuable -- a psychiatrist who knows a patient's history over years provides better care than a new provider seeing them for the first time.

Convenience

Eliminating commute time, waiting room time, and the logistics of in-person visits reduces the burden of psychiatric treatment. For patients with ADHD -- who already struggle with time management and organization -- this reduction in friction can be the difference between attending appointments and dropping out of care.

Comfort

Some patients are more comfortable discussing sensitive topics -- substance use, trauma, sexual behavior, suicidal thoughts -- from the privacy of their own home rather than in an unfamiliar office. The therapeutic environment includes the patient's sense of safety, and for some patients, home is where they feel safest.

Clinical Observation

Telehealth actually provides clinical information that in-person visits do not. Seeing a patient's living environment, observing how they interact with family members in the background, noticing environmental factors that contribute to their clinical presentation -- these are observations that occur naturally in telehealth and can meaningfully inform treatment.

Good Telehealth vs. Bad Telehealth

The distinction is not complicated. It comes down to clinical standards:

How Integrative Psych Implements Telehealth

At Integrative Psych, my clinical practice in New York City, we offer telehealth because it genuinely serves our patients. But we implement it with the same clinical standards I advocate for in my research:

• Initial evaluations are comprehensive. Whether in-person or telehealth, the first appointment involves a full psychiatric assessment. There is no abbreviated version for telehealth patients.
• Follow-up visits include adequate time. We do not offer 10-minute medication checks. Follow-up appointments are long enough for clinical assessment, medication adjustment, and therapeutic engagement.
• Controlled substances require proper evaluation. Stimulants and other controlled substances are prescribed only after thorough assessment that includes developmental history, collateral information when appropriate, standardized rating scales, and risk screening. This is true for telehealth and in-person visits alike.
• Monitoring is built in. We track outcomes systematically. We assess side effects at every visit. We monitor vital signs. We require follow-up appointments before refilling controlled substances.
• Continuity is prioritized. Patients see the same provider at each visit. The therapeutic relationship is central to our clinical model, and telehealth does not change that.
• Clinical judgment drives decisions. Every prescribing decision is made by a licensed psychiatrist with adequate information and time. We do not optimize for throughput.

This is not a heroic approach. It is simply standard psychiatric care delivered through a screen. The fact that this feels remarkable in the current landscape says more about the industry than about our practice.

The Policy Landscape

Since our JAMA Psychiatry paper was published, the regulatory landscape has begun to shift, though slowly:

• DEA enforcement actions: The DEA has investigated and taken action against several digital platforms for prescribing practices that violated controlled substance regulations. These actions validate the concerns we raised in our paper.
• State medical board responses: Some states have tightened telehealth prescribing requirements, particularly for controlled substances. Others remain permissive, creating a patchwork of standards that platforms can navigate toward the path of least resistance.
• Congressional attention: Federal legislators have held hearings on digital health oversight, and several bills have been introduced to address gaps in telehealth regulation. Progress has been slow, but the issue is on the legislative radar.
• Professional society guidelines: The American Psychiatric Association and other professional organizations have issued guidelines for telehealth practice that align with many of our recommendations. These guidelines lack enforcement power but establish the standard of care against which platform practices can be measured.

The fundamental challenge remains: regulation moves at the speed of government, and technology moves at the speed of venture capital. Bridging this gap requires sustained advocacy, rigorous research documenting outcomes, and professional organizations willing to define and enforce standards.

Where Telehealth in Psychiatry Is Heading

Telehealth is not going away, and it should not. The clinical benefits are real. What needs to change is the accountability framework around it.

I expect the next several years will bring:

• Outcome-based regulation: Rather than regulating process (evaluation length, visit frequency), regulators will increasingly focus on outcomes. Platforms that produce poor outcomes -- high adverse event rates, low treatment completion, excessive prescribing -- will face regulatory consequences.
• AI-augmented telehealth: AI tools will increasingly support telehealth practice -- clinical decision support, automated screening, measurement-based care tracking. My AI in psychiatry research addresses how to implement these tools responsibly.
• Quality transparency: Patients will eventually have access to quality data about telehealth platforms, similar to hospital quality ratings. This will enable informed choice and create market incentives for quality.
• Hybrid care models: The most effective psychiatric care will combine in-person evaluation with telehealth follow-up and digital monitoring -- using each modality where it is strongest.

Frequently Asked Questions

What did Dr. Sultan's JAMA Psychiatry research find about telehealth?

My JAMA Psychiatry viewpoint, co-authored with Manpreet K. Singh at Stanford (now UC Davis), documented how the rapid expansion of telepsychiatry during the pandemic outpaced regulatory oversight. We found that some digital platforms prescribe controlled substances after evaluations under 10 minutes, regulatory frameworks have not kept pace with telehealth expansion, volume-based business models incentivize speed over quality, and many platforms lack adequate monitoring for side effects and treatment outcomes. We proposed a framework for responsible telehealth integration with minimum evaluation standards and mandatory outcome monitoring.

Is telehealth good or bad for psychiatry?

It depends entirely on implementation. Telehealth done well -- comprehensive evaluations, adequate follow-up, monitoring, continuity of care -- improves access and serves patients who face barriers to in-person visits. Telehealth done poorly -- rushed evaluations, controlled substance prescribing without safeguards, no follow-up -- compromises patient safety. The technology is neutral; the clinical standards built around it determine the outcome.

What are the risks of online psychiatry platforms?

The main risks I identified include controlled substance prescribing without adequate evaluation (evaluations under 10 minutes for conditions requiring comprehensive assessment), lack of follow-up monitoring, no therapeutic relationship between provider and patient, regulatory gaps enabling cross-state prescribing without oversight, and business models that prioritize prescription volume over clinical outcomes. These risks are not inherent to telehealth -- they are the result of specific business decisions that prioritize revenue over patient safety.

How should telehealth prescribing be regulated?

My framework recommends minimum evaluation standards before prescribing controlled substances via telehealth, mandatory outcome monitoring tied to platform prescribing privileges, regulatory requirements specific to digital health companies, quality metrics measuring patient outcomes rather than visit volume, and transparent public reporting of prescribing practices by platform. The goal is not to restrict telehealth but to ensure it meets the same clinical standards as in-person care.

Does Dr. Sultan use telehealth in his practice?

Yes. At Integrative Psych, I offer telehealth with the same clinical standards as in-person visits. Initial evaluations are comprehensive regardless of modality. Controlled substances are prescribed only after thorough evaluation with ongoing monitoring. Follow-up visits include adequate assessment time. Every clinical decision is made by a licensed psychiatrist with sufficient information. The difference between responsible telehealth and problematic telehealth is not the screen -- it is the clinical standards behind it.

Further Reading

• AI in Psychiatry: Research Overview
• Integrative Psych: Our Practice
• PAWS Digital Therapeutic
• The Sultan Lab at Columbia
• Publications