Review Goal
The review asked a narrow question:
Does the spoken audio stay faithful to the source papers, especially where clinical interpretation could be overstated?
The review covered medication-specific liver testing, three outpatient decisions, treatment burden, adequate medication trials and comprehensive epilepsy referral.
Evidence Standard
The episode was reviewed against the full text of the source articles — not abstracts or summaries — to catch places where the audio could misrepresent scope, effect size, or clinical implications.
- Xcopri (cenobamate) US prescribing information, revised August 2025
- NICE NG217 chapter 4: Treatment, safety, monitoring and withdrawal
- NICE NG217 chapter 7: Status, repeated or cluster, and prolonged seizures
- NICE NG217 chapter 10: Epilepsy-related death and SUDEP
- Marson A, et al. SANAD II focal epilepsy. Lancet (2021). PMID33838757; PMCIDPMC8047799
- Jehi L, et al. Timing of referral to evaluate for epilepsy surgery. Epilepsia (2022). PMID35842919; PMCIDPMC9562030
Process
The episode was generated directly from the source papers, with explicit instructions to preserve study scope and avoid overstating findings.
The audio was converted to text so spoken claims could be checked line by line against the papers.
Two independent AI-assisted reviews checked spoken claims against the source text and flagged material overstatement, misrepresentation, or unsupported interpretation.
Each complete draft received two independent reviews and one combined editorial verdict: cleared, cleared with minor notes, needs revision, or re-record required. A draft could not advance until accuracy issues were resolved.
Verdict History
| Draft | Verdict | What changed |
|---|---|---|
| First draft | Needs revision | Referral denominator, trial interpretation and unsupported explanations. |
| Second draft | Needs revision | Liver-warning opening and trial-analysis explanation. |
| Third draft | Needs revision | Unsupported liver-injury and drug-resistance mechanisms. |
| Fourth draft | Cleared with minor notes | Core claims faithful; local-practice caveats and ending polish. |
| Fourth draft, ending trimmed | Needs revision | Later full review found an overcompressed red-flag action in the original recap. |
| Fourth draft, targeted corrections | Cleared with minor notes | Faulty recap and localized overstatements removed; correct detailed clinical pathway and sign-off retained; both reviews repeated. |
How the Language Changed
These are the specific spoken claims that failed review, shown alongside the corrected language in the final version. Quoted or closely paraphrased from the transcripts.
Liver-warning framing
Earlier draft
The second draft incorrectly said no patient injuries prompted the regulatory update.
Final version
The final version reports clinically significant postmarketing liver injury, including acute liver failure requiring transplantation, without an incidence estimate.
Trial scope
Earlier draft
Earlier drafts blurred remission analyses and supplied unsupported causal explanations.
Final version
The final version uses observed adverse-reaction proportions as a tolerability example and explicitly rejects automatic switching of stable patients.
Referral counseling
Earlier draft
The first draft treated an original-cohort response percentage as an individual next-drug probability.
Final version
The final version explains declining response qualitatively, separates referral from surgery and continues medication care.
Final Result
Verdict: Cleared with minor notes for clinician listening
The exact corrected listening file completed both AI-assisted source reviews; the clinician editor accepted the audio and minor notes and authorized publication on September 30, 2026. The episode was published on October 1, 2026.
Minor caveats carried into the show notes
- Early DRESS can occur without evident rash. The written companion provides the complete label symptom examples for liver injury.
- Individualize emergency plans within local protocols and prescribing standards. The overbroad from-scratch passage was removed.
- The label specifies ALT, AST and total bilirubin. Additional urgent tests are clinician-directed rather than universally label-mandated. Clinically indicated follow-up extends beyond symptoms alone.
- Jehi’s surgical-referral consensus distinguishes offering evaluation up to age 70 from considering older patients without surgical contraindications. Comprehensive care remains individualized.
- Drug and trial names were included in the clinician listening review; automatic transcription spelling alone does not establish a spoken error.
What Changed
- Corrected the liver-warning opening and retained the full baseline test triad and three-month window.
- Kept SANAD II within a tolerability example, without unsupported efficacy or biological explanations.
- Removed unsupported injury and drug-resistance mechanisms.
- Retained referral as evaluation, continued medical care, and individualized planning.
- Removed the overcompressed red-flag recap and other localized overstatements; retained the correct detailed assessment/testing pathway and full sign-off.
Production Learning
Four generated recordings and two edited assemblies were checked. Each completed candidate received two isolated AI review runs using the same model, followed by editorial adjudication. This provides independent checks within one model family, not cross-model validation or independent expert peer review.
Boundaries
This review does not certify clinical recommendations, replace human editorial judgment, replace independent expert review, transfer responsibility to the authors of the source studies, or make the episode a clinical guideline.