AI Scribe for Substance Use Disorder Treatment

Substance use disorder care lives at the intersection of two realities. Clinically, it is some of the most demanding documentation in behavioral health — multi-dimensional assessments, controlled-substance prescribing, frequent reauthorizations, and group modalities that no SOAP template was designed for. Legally, it sits under 42 CFR Part 2, a federal regulation that is stricter than HIPAA about how SUD records may be created, stored, and re-disclosed. Scriber is built for clinicians who have to satisfy both at once.

1. The SUD documentation problem

SUD documentation is heavier than almost any other behavioral health vertical, and the burden compounds in ways that general therapy notes do not.

The Joint Commission's Behavioral Health Care and Human Services standards expect SUD programs to document a thorough biopsychosocial assessment, an individualized treatment plan tied to ASAM placement, ongoing reassessment at defined intervals, and explicit linkage between treatment goals and the interventions delivered each session. For accredited programs, "the chart did not show how the level of care was justified" is a finding that puts contracts at risk.

ASAM criteria — the field's de facto standard for placement and continued-stay decisions — require the clinician to think and write across six dimensions: acute intoxication and withdrawal potential, biomedical conditions, emotional and cognitive conditions, readiness to change, relapse and continued-use potential, and recovery environment. A complete assessment needs to address all six, justify a level of care from outpatient through medically managed inpatient, and re-justify continued stay at predictable cadences. Most clinicians can carry the clinical thinking; the bottleneck is putting all six dimensions on paper without losing an hour per intake.

Reauthorizations make it worse. Commercial payers and many Medicaid plans demand utilization review for residential, partial hospitalization, and intensive outpatient programs as often as every 7 to 14 days. Each reauth packet typically pulls the most recent ASAM dimensions, observable progress, current treatment plan goals, and a justification for ongoing intensity. A program with 30 IOP clients can spend 10 to 20 clinician hours a week purely on reauthorization paperwork.

Layer on top of that the realities specific to SUD: MAT visits that need controlled-substance documentation; group counseling that is the dominant treatment modality in IOP and PHP; relapse-prevention plans that are themselves required by many payers; mandated drug-screen and PDMP documentation; and case-management notes for housing, employment, probation, and family services. The result is a workforce that, in surveys, reports spending 40 to 55 percent of its workday on documentation — among the highest in any clinical field.

This is the problem Scriber is built to solve for SUD specifically. Not "generate a SOAP note." Generate the assessment, generate the MAT visit, generate the group note, generate the relapse-prevention plan — and do it in formats your reviewer, your payer, and 42 CFR Part 2 all expect.

2. What Scriber captures during SUD visits

Scriber captures audio with patient consent and turns it into a structured draft note. What the draft contains depends on the visit type.

SUD intake. During a 60- to 90-minute intake, Scriber listens for the elements a biopsychosocial SUD assessment is built from: substances used (with onset, frequency, route, last use), withdrawal history and current symptoms, prior treatment episodes and outcomes, co-occurring mental health and medical conditions, family and social history, legal involvement, employment and housing, motivation language consistent with the stages of change, and risk factors including overdose history and suicidality. The draft populates an intake template the clinician selects — including a full ASAM-criteria assessment if that is the program's intake standard.

MAT visit. For induction, stabilization, and maintenance visits with buprenorphine, methadone, or extended-release naltrexone, Scriber captures dose changes, side-effect review, cravings and use since the last visit, urine drug screen results when discussed, PDMP review when mentioned, counseling component, and the plan including the next visit interval. The output is a MAT visit note structured the way a DEA-aligned auditor or a state opioid treatment program reviewer expects to see it.

Group counseling. Scriber's diarized transcription separates speakers in a group session. The clinician identifies each speaker once at the start of the recording; from there, Scriber attributes turns to participants and produces a distinct draft progress note per member, plus a group-level summary describing theme, curriculum module, and clinical observations.

Individual counseling. For 1:1 SUD counseling, Scriber generates the progress note in your preferred format — DAP, BIRP, GIRP, or a custom SUD progress template — capturing presented concerns, interventions, response, homework, and plan, with attention to recovery-specific content like cravings, triggers, recovery capital, and relapse prevention.

In every visit type, the audio is processed for transcription and note generation, then discarded. The note is the artifact you keep; the recording is not stored.

3. SUD note formats Scriber generates

Scriber ships purpose-built SUD templates, not generic SOAP variants pretending to fit. Each is editable, and each can be tuned per organization.

SUD intake

A long-form biopsychosocial intake covering chief complaint, substance use history (per substance, with route, frequency, last use, longest period of abstinence), psychiatric and medical history, family and developmental history, social and legal history, current medications, allergies, mental status exam, diagnostic impression with DSM-5-TR criteria documented, and an initial treatment recommendation. Optional addenda for trauma screening, suicide risk assessment, and overdose-risk assessment.

ASAM-criteria assessment

Sectioned strictly by the six ASAM dimensions, with a placement summary that names the recommended level of care (outpatient, IOP, PHP, residential, medically managed) and the dimensional findings that justify it. Designed to support both initial placement and continued-stay reviews. Reauthorization-ready.

MAT visit note

Structured for buprenorphine, methadone, or naltrexone visits. Captures phase (induction / stabilization / maintenance), current dose and any change with rationale, observed dosing where applicable, withdrawal and craving scales when used (e.g., COWS, CIWA), urine drug screen results discussed in the visit, PDMP review acknowledgment, counseling component, prescription details discussed, and the follow-up interval.

Group counseling progress note

Two-part output: (1) a per-participant progress note attributing observed engagement, contributions, and clinical impressions to the right group member; (2) a group-level summary capturing curriculum content, group dynamics, and clinical themes. Length is handled gracefully — 90-minute IOP groups produce per-member notes that are concise and individualized rather than identical boilerplate.

Relapse-prevention plan

Generated either as a standalone document or as an addendum to a treatment plan. Covers identified triggers, high-risk situations, early warning signs, internal and external coping strategies, support contacts, recovery community involvement, and a written plan for managing a lapse. Structured so the document can be shared with the client and revisited.

SUD discharge summary

Captures admission and discharge dates, levels of care across the episode, ASAM dimensions at discharge, treatment goals and progress, medications at discharge (including MAT continuation plan), aftercare plan with named providers and follow-up dates, and any continuing-care recommendations.

4. 42 CFR Part 2 and SUD privacy specifics

42 CFR Part 2 is stricter than HIPAA in three ways that matter operationally.

Heightened consent. Disclosure of Part 2-protected SUD records generally requires patient consent that names the specific recipient and the specific purpose. HIPAA's broader treatment, payment, and operations exceptions do not flatten this. A note written about a MAT visit in a Part 2 program cannot be re-disclosed to a primary care physician without appropriate consent, even if a HIPAA-covered entity could otherwise share it.

Re-disclosure prohibition. Recipients of Part 2 records cannot re-share them without the patient's renewed consent. This shapes how SUD programs structure release-of-information forms, how they answer subpoenas, and what gets sent in a referral packet.

Court-order standard. Disclosures pursuant to a court order require findings the court is supposed to make on the record — a higher bar than a HIPAA-permitted disclosure to law enforcement.

What changes for an AI scribe in this environment? The practice still owns the consent ledger, the release-of-information form, and the record-of-truth where the note ultimately lives. Scriber's role is narrower and more conservative: capture audio with consent, generate a draft note, and hand the draft to the clinician for entry into the practice's Part 2-compliant record system.

Three Scriber design choices reduce Part 2 exposure surface:

1. Audio non-retention. The recording is processed for transcription and discarded. There is no archive of the spoken Part 2 content sitting in a vendor's cold storage that a future request could compel. The note exists in the practice's record system; the audio does not exist anywhere.
2. AES-256 encryption in transit and at rest for the brief window during which transcription and note generation happen, on infrastructure (Azure) covered by a Business Associate Agreement.
3. BAA at signup, not gated to enterprise contracts. Solo practitioners and small SUD groups get the same compliance footing as a large agency.

Scriber does not replace your Part 2 program policies, your consent forms, or your record-of-truth EHR. It works alongside them — and is designed not to create new disclosure pathways.

5. Telehealth SUD treatment with Doxy.me + Scriber

Telehealth has been a structural change for SUD treatment. The post-2020 telehealth flexibilities allowed buprenorphine induction by video in many circumstances, expanded the geographic reach of MAT, and made IOP and individual counseling viable for clients who could not otherwise sustain attendance.

Scriber has a real Doxy.me integration — not a marketing claim, a shipped feature. The workflow:

1. Patient enters the Doxy.me waiting room. The clinician confirms identity and obtains the verbal consent for recording that the practice's policy and applicable state laws require.
2. Clinician starts the visit and starts Scriber inside Doxy.me. Audio is captured directly from the Doxy.me session. There is no separate recorder, no second device, and no copy-paste of audio files.
3. Visit happens normally. Clinician and client talk; Scriber listens.
4. Visit ends; Scriber generates the draft note. The clinician chooses the template (MAT visit, individual SUD counseling, intake, etc.). The transcription runs on xAI's fast pipeline; in production benchmarks, this materially outpaces alternatives, which means the draft is ready by the time the clinician is finished with closing tasks.
5. Clinician reviews, edits, and exports. The note goes into the record-of-truth EHR. The audio is discarded.

For SUD telehealth specifically, this matters because: MAT visits are short, the documentation is dense, and clinicians often run consecutive 15- to 20-minute follow-ups. The minutes saved per visit by not retyping a structured MAT note compound across a panel.

A few telehealth-specific details that come up in SUD program rollouts. Buprenorphine telehealth induction notes need to capture verbal confirmation of identity, the location of the patient at the time of the visit (state matters for prescribing authority), the COWS score if a withdrawal scale was used, the dose induction pattern, and any observed dosing limitations. Scriber's MAT template prompts for those elements when the visit is flagged as an induction. Cross-state prescribing edge cases are clinician-owned — the documentation captures what is said, but the licensure decision and DEA registration footprint stay with the prescriber. Recording consent is verbal and documented in the note; Scriber does not record without that affirmative step inside the workflow.

The result is a telehealth SUD workflow that does not require a clinician to switch tools. Doxy.me handles the visit, Scriber handles the documentation, and neither tool retains the audio after the note is signed.

6. Group counseling note generation

Group counseling is where AI scribes typically fail SUD programs. A 90-minute IOP group with eight participants is not a SOAP note — it is eight progress notes plus a group summary, each of which has to reflect what that specific person did or said.

Scriber handles it in three steps:

Diarization. The transcription pipeline separates speaker turns. The clinician identifies each speaker once at the start of the recording — typically by saying the participant's first name or initials as they speak — and Scriber maintains the attribution through the rest of the session.

Per-member attribution. When the note is generated, observations and contributions are routed to the right participant's record. If a participant is largely silent, the note reflects that honestly rather than fabricating engagement. If a participant disclosed a recent lapse, that goes in their note and not in everyone else's.

Length handling. Long group recordings are processed without forcing every participant note to look identical. The output for a heavily engaged member is longer and more specific; the output for a quiet member is shorter and accurate. The clinician then reviews each note before signing.

The result: a 90-minute group session produces eight clinician-reviewable progress notes plus a group-level summary, in the time it currently takes most programs to write two.

Two practical considerations worth surfacing. First, diarization quality scales with audio quality. A group seated around a single table with one room microphone produces good attribution; a group on a video platform where each participant has their own audio channel is even better. For in-person groups, programs should standardize on a multi-channel or omnidirectional microphone. Second, clinical attribution is the clinician's job, not the model's. Scriber's diarization gets the speaker turns right; the clinician's responsibility is to confirm that the named participant matches the speaker and that the clinical observations are accurate before signing. This review step is part of the workflow, not an afterthought.

For programs running curriculum-driven groups — Matrix Model, Seeking Safety, relapse-prevention modules — the group summary captures which curriculum module was delivered and how participants engaged with it, which is useful for both clinical supervision and payer documentation that wants to see evidence-based content delivered.

7. Common SUD-clinician questions

How does Scriber handle controlled-substance MAT documentation? MAT visit templates capture the elements DEA-aligned reviewers and payer auditors look for: phase, dose, formulation, observed dosing where applicable, side effects, cravings, urine drug screen results discussed in the visit, PDMP review acknowledgment, counseling component, and follow-up interval. Scriber generates the documentation; it does not e-prescribe, does not query the PDMP for you, and does not replace clinician judgment. The clinician signs.

What about court-ordered treatment and drug-court reporting? Scriber generates the underlying clinical note. Status reports and updates to drug courts that re-disclose Part 2-protected information remain a clinician decision and require either valid patient consent that names the court as the recipient, or a court order that meets the 42 CFR Part 2 standard. Scriber does not auto-transmit to courts and does not maintain a court-reporting workflow that would create a new disclosure pathway.

How does Scriber interact with state PDMPs? It does not query PDMPs. PDMP checks remain the clinician's workflow inside the state portal or the EHR's PDMP integration. Scriber documents that the check occurred when it is mentioned during the visit, with the result the clinician verbalized — supporting the auditable record without replacing the actual query.

Can Scriber accommodate program-specific templates? Yes. SUD programs vary — opioid treatment programs, office-based opioid treatment, IOP, PHP, residential — and each has documentation conventions. Templates are tunable per organization. The shipped templates are starting points, not the only options.

What about co-occurring disorders? Most SUD clients have co-occurring mental health conditions. Scriber generates dual-diagnosis-aware notes when the visit covers both, and it can switch between SUD-specific and general behavioral health templates per visit.

How long until a note is ready? With xAI fast transcription, the draft is typically ready within seconds of the visit ending, not minutes. For a clinician running a back-to-back schedule, this is the difference between leaving on time and going home with an open note backlog.

8. Try Scriber for SUD treatment

Scriber is built for the clinicians the rest of behavioral health software treats as an afterthought — addiction medicine prescribers, MAT clinicians, SUD counselors, IOP and PHP teams, opioid treatment programs, and dual-diagnosis programs. The note formats are SUD-specific. The privacy posture is conservative by default. The Doxy.me integration is real, the BAA is signed at signup, the audio is not retained, and the encryption posture is what an SUD program's compliance officer expects to see.

The trial is structured so a clinician can validate fit on real visits within a single clinic day. Sign up, sign the BAA inline, and run Scriber on the next intake — the ASAM-criteria template will populate from a 75- to 90-minute interview without the clinician transcribing a word. Run it on the next MAT follow-up and watch the dose, COWS or craving rating, and PDMP-acknowledgment fields populate from a four-minute conversation. Run it on the next group and review the per-participant attribution before signing each note. If the output is not better than what the program produces today, the trial costs nothing.

For multi-clinician SUD organizations, the demo covers the operational pieces a single trial cannot: organization-level template tuning so every clinician's MAT note matches the program's conventions, supervision and co-signature workflow for trainees and counselors-in-training, group-administration features for IOP and PHP rosters, and a walk-through of the 42 CFR Part 2 boundary — what Scriber sees, what it never retains, and where the practice's record-of-truth and consent ledger remain in control. Programs that have done this demo typically come away with a clearer picture of what an AI scribe should and should not do in a Part 2 context than they had going in.

Start a free trial · Book a SUD-program demo · Explore the Doxy.me integration