Legal & Compliance · 2026-06-12 · 2,250 words
Presurgical psychological evaluations and cloud AI scribes: the formal report flows to the surgical team, but the vendor keeps the session audio
Bariatric surgery, spinal cord stimulator implantation, and solid organ transplant evaluations involve a documentation structure unlike ordinary therapy: the psychologist's formal report is explicitly meant to reach the surgical team, the insurer, and the referring physician. That expected disclosure is authorized, documented, and structured. What is neither expected nor authorized is the cloud AI scribe vendor retaining verbatim audio from the full clinical interview — a separately held archive that no surgical authorization covers and that third parties can reach independently.
- Presurgical evaluations (bariatric, SCS, transplant) have a structured two-document reality: formal report → surgical team and insurer; clinical interview audio → stays with the evaluating psychologist.
- A cloud AI scribe creates a third location: vendor server. The vendor's session audio archive is not covered by the patient's surgical authorization and is independently subpoenable.
- The gap between the formal report and the verbatim clinical interview is precisely what plaintiff attorneys, licensing boards, and disability insurers are looking for in adverse outcome cases.
- Licensing board investigations of presurgical evaluation practice specifically target the clinical interview — the content the psychologist did not put in the formal report.
- On-device processing eliminates the vendor as an independent custodian of the clinical interview session while leaving the formal report's expected disclosure pathway unchanged.
What presurgical psychological evaluations are and where AI scribes enter the workflow
Presurgical psychological evaluations are standardized assessments conducted by a licensed psychologist or mental health professional before certain elective or high-risk procedures where psychological factors are predictive of surgical outcomes. Three procedure categories have established evaluation requirements used across the United States in 2026:
Bariatric surgery. The American Society for Metabolic and Bariatric Surgery (ASMBS) clinical practice guidelines recommend a presurgical behavioral health evaluation for candidates undergoing sleeve gastrectomy, Roux-en-Y gastric bypass, or duodenal switch procedures. The evaluation assesses psychiatric comorbidities, substance use, eating disorder pathology, surgical motivation and expectations, social support, and capacity to comply with the significant post-surgical lifestyle and dietary modifications required for a durable outcome. Insurers including most commercial carriers and Medicare Advantage plans require the evaluation report as part of the prior authorization package. The evaluating psychologist submits a structured written report to the surgical team and the insurer.
Spinal cord stimulator (SCS) implantation. Major payers — Medicare and most commercial insurers — require a presurgical psychological evaluation before approving SCS implantation for chronic pain. Centers for Medicare & Medicaid Services coverage policy for SCS specifies that candidates must undergo psychological assessment to screen for psychiatric contraindications including active major depression, psychosis, untreated anxiety disorders, personality disorders affecting treatment adherence, and ongoing substance use — each of which predicts significantly worse outcomes in randomized trials. The evaluation report becomes part of the prior authorization documentation and is reviewed by the insurer's medical director. Denial of SCS coverage based on psychiatric contraindication findings triggers an appeal process in which the evaluation report is the central contested document.
Solid organ transplant. The United Network for Organ Sharing (UNOS) policies and transplant center accreditation standards require a psychosocial evaluation of transplant candidates as part of the listing process for kidney, liver, heart, lung, and pancreas transplantation. The evaluation covers psychiatric history, substance use and sobriety (with specific timelines required for alcohol-related liver disease listings), treatment adherence history, social support, understanding of the post-transplant medical regimen, and ability to manage the complex immunosuppression and monitoring requirements that begin immediately after surgery. The evaluation report reaches the transplant team's multidisciplinary committee, the listing committee, and in cases of combined organ transplants, sometimes the programs at multiple institutions.
In each of these contexts, the evaluation involves a clinical interview lasting 60 to 90 minutes, administration of standardized assessment instruments, and collateral contacts in some cases. A psychologist who conducts presurgical evaluations as a meaningful part of their practice — common in hospital-contracted or surgical-practice-contracted private practice — may see five to fifteen evaluation clients per week, and an AI scribe is a natural efficiency tool for transcribing and structuring the clinical interview notes that underlie the formal report.
The two-document structure: what the formal report contains and what the clinical interview contains
The formal presurgical evaluation report is a professional document. It is structured, typically five to ten pages, and organized around the specific referral question: is this candidate psychologically cleared for the procedure, with what recommendations and with what concerns? A well-constructed bariatric evaluation report covers psychiatric diagnoses and treatment history, substance use history and current status, eating disorder assessment, surgical motivation and expectations, social support assessment, understanding of post-surgical lifestyle requirements, and the psychologist's recommendation — cleared, cleared with conditions, or not cleared at this time.
What the formal report contains is the product of the psychologist's clinical judgment about what information is material to the surgical candidacy question. It is organized, professional, and deliberately selective. It does not include the full verbatim transcript of the clinical interview.
The clinical interview itself is substantially different from the formal report. Over 60 to 90 minutes of structured and semi-structured interview, a presurgical evaluation client typically discloses:
- Complete psychiatric history including treatment episodes, hospitalizations, diagnoses, and medications — including episodes the client may qualify or minimize in the formal report's summary
- Detailed substance use history with timing, quantities, treatment attempts, and relapse history — including admissions made with caveats ("I know I'm supposed to be six months sober, but I did have a few glasses at my daughter's wedding last month") that the psychologist may note as a condition in the report but that appear verbatim in the session audio
- Eating behavior history for bariatric candidates — binge episodes, purging history, night eating patterns, emotional eating triggers, and the client's own assessment of their relationship to food, often in their own unguarded words
- Social support assessment content including family dynamics, marital conflict, financial stress, and housing instability that may be summarized clinically in the report but captured in detail in the session audio
- Responses to standardized screening instruments including PHQ-9 scores, substance use screening results, and cognitive screening results — with any contradictions between instrument responses and verbal responses during the interview captured in the audio but not always documented in detail in the formal report
- Spontaneous disclosures that arise during rapport-building portions of the interview: current stressors, recent events, relationship conflicts, and concerns about the procedure that the client volunteered outside the formal question structure
When a psychologist uses a cloud AI scribe during the clinical interview, the vendor captures all of this. The formal report that subsequently flows to the surgical team and insurer reflects the psychologist's professional distillation of that material. The vendor's session audio archive contains the raw source.
Who receives the formal report — and what authorization that covers
The formal evaluation report's disclosure to the surgical team, insurer, and referring physician is authorized under HIPAA in two ways depending on the practice structure. For evaluations performed as part of a treatment team relationship — where the psychologist is part of the surgical program's coordinated care team — the disclosure falls under HIPAA's treatment operations exception at 45 CFR § 164.506, which permits covered entities to use and disclose PHI for treatment without specific authorization. For evaluations performed by independently contracted psychologists outside the surgical program, most practices use a specific HIPAA authorization signed by the patient at the start of the evaluation, which identifies the surgical team and the insurer as the designated recipients of the formal evaluation report.
In either structure, what is authorized is the formal evaluation report — the document the psychologist prepared. The authorization does not cover the AI scribe vendor's independently retained session audio and transcript. The vendor's retention of that audio is governed by the BAA between the psychologist and the vendor. A business associate agreement governs the vendor's security obligations and data handling practices — but it does not constitute authorization for the vendor's session audio to reach third parties, and it does not restrict what legal process can obtain from the vendor independently of the psychologist's authorization framework.
The patient who signed a HIPAA authorization for the formal report to go to their surgical team did not sign an authorization for the cloud AI scribe vendor to retain the full 90-minute clinical interview audio indefinitely. The vendor's data retention policy — often 30, 60, or 90 days after the session, or longer in some vendor terms of service — is a separate document the patient encountered in the vendor's terms of service acknowledgment, not in the surgical authorization they signed in the psychologist's office.
The vendor's independently retained audio and how it reaches discovery
The cloud AI scribe vendor's session audio archive is independently accessible through legal process directed at the vendor — not at the evaluating psychologist. This independence from the psychologist's own records is the specific architecture problem. A subpoena directed at the vendor runs to the vendor's own legal obligations, not to the psychologist's professional privilege claims or HIPAA framework.
Several specific litigation and regulatory contexts create motivation to reach the vendor's presurgical evaluation audio:
Surgical malpractice cases. When a surgical outcome is poor — a bariatric patient experiences serious post-surgical complications, a spinal cord stimulator patient reports worsened pain or neurological injury, an organ transplant recipient experiences early graft failure — litigation frequently follows. In surgical malpractice cases, both plaintiff and defense attorneys seek the full preoperative record: what did the candidate disclose during evaluation, was the candidate appropriately screened, were material risk factors identified and communicated? The formal evaluation report is one record. The plaintiff's attorney, having obtained that report, may issue a third-party subpoena under Rule 45 to the AI scribe vendor for the full session audio — specifically looking for content the patient disclosed during the clinical interview that does not appear in the formal report. Discrepancies between what the patient said and what the formal report documented are probative evidence in claims about whether the surgical clearance was appropriate.
Insurance prior authorization appeals. When an insurer denies prior authorization for bariatric surgery or SCS implantation based on the psychologist's evaluation report — citing identified psychiatric contraindications — the patient may appeal. In complex appeals proceeding through external independent review, the reviewing organization requests the complete evaluation record. If the AI scribe vendor's session audio surfaces through the appeals process, it may reveal content the patient disclosed that supported the contraindication finding, or content that the patient believes was mischaracterized in the formal report. Either direction can intensify the dispute.
Disability insurance proceedings. Disability insurance proceedings — SSDI and long-term disability — frequently involve prior records of psychological treatment and evaluation. A patient who had a poor surgical outcome and is now seeking disability benefits may have the presurgical evaluation used against them: the evaluation documented psychiatric comorbidities that insurers argue contributed to functional limitations predating the disability onset date. The vendor's session audio, if obtained in discovery, provides more complete content about the patient's psychiatric history and pre-surgical functional status than the formal report.
Licensing board investigations. If a patient files a licensing board complaint against the evaluating psychologist — alleging that the evaluation was conducted incompetently, that the clearance was inappropriate, or that the psychologist failed to identify contraindications — the licensing board has broad investigatory authority. State psychology boards have subpoena power to compel production of all records related to the evaluation, including records held by third parties. A board investigation that learns a cloud AI scribe was used during the evaluation can subpoena the vendor directly for the session audio and transcript. The board's review of the gap between what the patient said during the interview and what appeared in the formal report is the specific inquiry that the vendor's audio enables — and that the psychologist's formal records alone might not reveal.
What the verbatim gap reveals: why the clinical interview content is the target
In each of these legal and regulatory contexts, the target is not the formal evaluation report — that document is already part of the patient's medical record and is routinely disclosed under HIPAA's treatment and payment operations exceptions. The target is the gap between the formal report and the full clinical interview.
That gap is not evidence of misconduct. It is the normal product of professional documentation judgment. Psychologists conducting presurgical evaluations make deliberate decisions about what clinical information is material to the surgical candidacy question and what falls outside the scope of the formal report. A bariatric evaluation client who discloses a distant history of cannabis use that has been in sustained remission for eight years does not need that history prominently featured in a clearance report — but it appears in the verbatim clinical interview audio. An SCS candidate who describes significant marital conflict and financial stress may receive a report that notes "adequate social support with some reported stressors" — while the clinical interview audio captures an hour of detailed description of the specific conflicts.
The distinction between psychotherapy notes and progress notes under HIPAA does not map neatly onto presurgical evaluation content. The formal evaluation report is clearly a progress note equivalent — a record created for the purpose of transmitting information to treating providers. The clinical interview itself, conducted by a mental health professional and involving significant personal disclosure, occupies a gray zone. But the AI scribe vendor's retained audio is unambiguously a vendor business record — the vendor's own data, held under the vendor's retention policy, and subject to the vendor's legal response to process — without the specific HIPAA protections that apply to the psychologist's own records.
What cloud AI scribes actually transmit to vendor servers includes the full session audio, not just the text the psychologist ultimately uses in the report. Every disclosure, every spontaneous comment, every screening response qualified with an explanation, every emotional moment in the interview — all of it is in the vendor's archive, organized by session timestamp and client identifier, accessible through the vendor's API and producible in response to legal process directed at the vendor.
On-device processing: eliminating the vendor as an independent custodian
The architecture problem in presurgical evaluation documentation is specific: there are two legitimate pathways by which presurgical evaluation content reaches third parties — the formal report's authorized disclosure to the surgical team and insurer, and the psychologist's own records subject to subpoena and board investigation in the same way all clinical records are. A third pathway — the vendor's independently retained session audio, accessible through legal process directed at the vendor without the psychologist's involvement — exists only when a cloud AI scribe is used.
On-device processing eliminates that third pathway architecturally. If the clinical interview audio, transcript, and AI-generated note content are processed on the evaluating psychologist's Mac and never transmitted to vendor infrastructure, the vendor holds no session content. A subpoena directed at the vendor produces nothing, because there is nothing to produce. A board investigation that learns the psychologist uses an on-device AI scribe finds that the vendor's data holdings are limited to the psychologist's licensing and billing relationship — not to session content.
The formal report's disclosure pathway — the content that is supposed to reach the surgical team and insurer — is entirely unaffected. On-device processing changes only the vendor's custody of raw session content. The report the psychologist writes is still produced, still reviewed, still transmitted to the surgical team and insurer under exactly the same HIPAA framework as before. The change is that the full clinical interview audio, which was never intended to reach third parties and which the patient's surgical authorization does not cover, remains within the one custodial relationship that does apply — the psychologist's own records — rather than creating an independent second custodian with its own legal exposure.
For psychologists whose practice includes presurgical evaluations, this architectural distinction is the most relevant dimension of the cloud AI scribe privacy question. The evaluation workflow involves a structured disclosure by design — the formal report going exactly where it is supposed to go. What that structure should not involve is an unintended second custodian retaining the raw material from which the formal report was produced, available to be reached through legal process in precisely the adverse-outcome scenarios where the gap between the full clinical interview and the formal report is most consequential.
The formal report flows to the surgical team. The session audio stays on your Mac.
TherapyDraft processes presurgical evaluation session audio entirely on your Mac. Nothing reaches vendor infrastructure — so there is no vendor archive for third parties to reach through legal process.
Start free — 10 sessionsFrequently asked questions
Is a presurgical psychological evaluation covered by HIPAA?
Yes. A psychologist or licensed mental health professional performing a presurgical psychological evaluation is a HIPAA covered entity when they transmit health information electronically in connection with standard transactions — which includes billing insurance for the evaluation. The formal evaluation report, the clinical interview session, and the AI-generated note content are all protected health information. The cloud AI scribe vendor who records and processes the clinical interview is a business associate under HIPAA, required to sign a BAA with the evaluating psychologist. The patient's authorization for the formal report to flow to the surgical team and insurer does not cover the vendor's independently retained session audio — that is a separate custody relationship governed by the BAA, not by the surgical authorization.
When the surgical team receives my formal evaluation report, do they also get access to the AI scribe vendor's session audio?
No — but they can potentially reach it through legal process. The formal evaluation report flows to the surgical team and insurer under HIPAA's treatment and payment operations exceptions, or under a specific patient authorization. That authorization covers the formal report the psychologist prepared. It does not cover the AI scribe vendor's independently retained audio recording and transcript of the clinical interview. If litigation arises — a malpractice case against the surgical team, a licensing board complaint, a disability claim — the parties involved can issue third-party subpoenas to the vendor for the session audio. The vendor's legal obligation in response to a court-directed subpoena is independent of the psychologist's HIPAA authorization framework.
Can a plaintiff attorney in a surgical malpractice case subpoena my AI scribe vendor for the presurgical evaluation session audio?
Yes. In surgical malpractice litigation, plaintiff attorneys have broad discovery rights to records bearing on the patient's preoperative condition and surgical clearance. A third-party subpoena under Rule 45 directed at the AI scribe vendor is a straightforward mechanism for obtaining the vendor's retained audio and transcript from the clinical interview. The vendor's response is governed by its own legal obligations, not by the evaluating psychologist's HIPAA framework. The psychologist can provide notice of the subpoena and raise patient privacy interests, but the vendor — as an independent custodian of the session audio — must respond to valid court process independently. This is a materially different legal posture than a subpoena directed at the psychologist, which the psychologist can contest directly through applicable privilege arguments.
What does the vendor's session audio contain that the formal evaluation report does not?
Formal presurgical evaluation reports are professional documents — structured around the specific surgical candidacy question and the product of clinical judgment about what information is material. Clinical interviews for these evaluations typically last 60 to 90 minutes and cover full psychiatric and medical history, substance use history, trauma history, current relationships and support, surgical motivation and expectations, and structured screening instruments. The clinical interview also captures the patient's spontaneous disclosures, emotional reactions, and qualifications to formal answers — the content the psychologist assessed but may not have included verbatim in the formal report. In litigation and regulatory contexts, the gap between what the patient said during the interview and what appears in the formal report is often the specific target: plaintiff attorneys look for disclosures that contradict the clearance recommendation, and licensing boards examine whether material clinical information that appeared in the interview was appropriately evaluated and documented.
Does on-device processing protect presurgical evaluation session content from third-party access?
On-device processing eliminates the vendor as an independent custodian of the clinical interview audio — which is the specific architecture problem that creates third-party exposure. If session audio, transcript, and AI-generated note content are processed on the psychologist's Mac and never transmitted to external vendor infrastructure, the vendor holds no session content to produce in response to a third-party subpoena. The formal evaluation report still flows to the surgical team and insurer as before — that disclosure pathway is unchanged. The psychologist's own records remain subject to subpoena and licensing board process as they always were. The architectural change is the elimination of the vendor as an independent custodian of raw clinical interview content reachable through legal process without the psychologist's involvement.