Substance Use and Addiction

Home > Addiction Psychiatrist NYC

This is my funded research area, and it is most of what I do outside the clinic.

I am a Research Scientist in Columbia's Division on Substance Use Disorders and an affiliate of CHOSEN — Columbia's Center for Healing of Opioid and Other Substance Use Disorders — where my work includes using large-scale health data to model and predict opioid overdose risk. I direct the Sultan Lab for Mental Health Informatics, which uses electronic health record data and natural language processing to find the gaps in addiction treatment.

Published work includes adolescent cannabis use and psychosis risk in Pediatrics, non-disordered cannabis use outcomes, e-cigarette use patterns, and elevated substance use in untreated ADHD — cannabis aOR 2.2, cocaine aOR 2.6.

How I approach this

No moralizing. I ask what you use, how much, and when, because those answers change diagnosis and treatment. Not to judge them.

Your stated goal, not mine. Abstinence works for some people. Reduction works for others. Someone drinking daily who moves to twice a week has improved, and treating that as failure drives people out of care. I'll tell you honestly what I think the evidence supports for your situation and then work toward what you actually want.

Dual diagnosis is the norm. Most people with a substance use disorder have a co-occurring psychiatric condition, and treating one without the other reliably fails. This is the single most common reason addiction treatment doesn't hold.

What responds to what

Alcohol. Naltrexone and acamprosate have the strongest evidence and are substantially underprescribed — most people with alcohol use disorder are never offered medication at all. Disulfiram works in specific circumstances with the right patient. Withdrawal from heavy daily drinking is medically dangerous and sometimes requires supervised detox before anything else.

Opioids. This is where my research sits — Columbia's CHOSEN center exists because the overdose crisis is a data and implementation problem as much as a clinical one. Buprenorphine and methadone are the most effective treatments available and reduce mortality substantially. Naltrexone is an option after full detoxification. The federal X-waiver requirement for buprenorphine was eliminated in 2023, removing a barrier that kept prescribing artificially rare for two decades. Treatment duration is not a fixed course — the evidence supports maintenance for as long as it's helping.

Cannabis. Two findings from my own work, because they cut against the common assumption that the line to worry about is addiction.

Nondisordered Cannabis Use Among US Adolescents — JAMA Network Open, 2023 (N=68,263, nationally representative). Adolescents using cannabis below the DSM-5 threshold for a use disorder had roughly two to four times greater odds of every adverse psychosocial outcome measured: major depression (aOR 1.86), suicidal ideation (aOR 2.08), difficulty concentrating (aOR 1.81), low GPA (aOR 1.80), arrest (aOR 4.15). Non-disordered use was about four times as prevalent as cannabis use disorder.

Cannabis Use Among US Adolescents — Pediatrics, 2025 (N=162,532, Monitoring the Future). A consistent dose-response relationship across academic, self-regulation, and emotional outcomes — present even among monthly users. Adolescents under 16 were more susceptible.

"Not addicted" is not "not affected." No FDA-approved medication exists for cannabis use disorder. Behavioral treatment, with attention to what the cannabis is doing — most of the daily users I see are managing anxiety, insomnia, or PTSD hyperarousal, and treating those directly is usually the more effective route. My research also concerns cannabis and psychosis risk: daily high-potency use, particularly with adolescent onset, carries substantially elevated risk.

Stimulants. No FDA-approved medication. Contingency management has by far the strongest evidence and is chronically underused. Important distinction: prescribed stimulant treatment for ADHD is associated with lower substance use risk than untreated ADHD — untreated ADHD is the risk factor, not the treatment.

Benzodiazepines. Withdrawal is medically dangerous and requires a supervised taper, never abrupt discontinuation. Tapering long-standing benzodiazepine prescriptions is a meaningful part of my practice, and it is slow work done properly.

Adolescents and young adults

My research population. See adolescent psychiatry.

For parents: I can give you the actual data rather than alarmism or dismissal. Occasional use in a nineteen-year-old and daily high-potency use in a fifteen-year-old are different clinical situations requiring different responses, and conflating them costs credibility with the teenager you're trying to help.

Scope

Outpatient practice. Right for evaluation, medication treatment, dual diagnosis, and ongoing care.

Not right for medically supervised withdrawal from alcohol or benzodiazepines, or for anyone needing residential or intensive outpatient care. If that's the level required I'll say so and help find it.

In crisis: call or text 988.

Related

Adolescent psychiatry · ADHD · PTSD · Anxiety · Depression · Cannabis research · DBT

When Substance Use Becomes an Emergency

Medically supervised withdrawal from alcohol or benzodiazepines is not an outpatient office procedure. If that's what's needed, an emergency department or a supervised detox program is the right starting point, and I'll help find it.

Frequently Asked Questions

Is there medication for alcohol use disorder?

Yes. Naltrexone and acamprosate have the strongest evidence and are substantially underprescribed — most people with alcohol use disorder are never offered medication. Disulfiram is an option in specific circumstances.

Do I have to stop completely to get treatment?

No. Abstinence is the right goal for some people and reduction for others. I'll tell you what I think the evidence supports for your situation, then work toward the goal you actually hold.

Does ADHD medication cause addiction?

The evidence points the other way. Treated ADHD is associated with lower substance use risk than untreated ADHD. Untreated ADHD is the risk factor.

Can you prescribe buprenorphine?

The federal X-waiver requirement was eliminated in 2023, so any DEA-registered prescriber can now prescribe buprenorphine for opioid use disorder. Whether this practice is the right setting depends on your situation, and I'll say so directly.

How much cannabis is too much?

Frequency, potency, and age of onset all matter more than quantity alone. Daily high-potency use, particularly beginning in adolescence, carries substantially elevated psychosis risk and interferes with treatment for anxiety, depression, and ADHD.

What if I have both a substance problem and a psychiatric condition?

That's the common case rather than the exception, and treating one without the other reliably fails. Both need addressing, usually concurrently.