SCRIBE RESULTS / 22 SEPTEMBER 2026
PatientNotes
Averages across 5 graded runs of 11 consultations, the same audio and instruction for every scribe.
2m 45sTime to a signable note · ranked 6 of 7
0%Notes ready to sign as written · ranked 7 of 7
21.6Significant note errors per run · ranked 7 of 7
9.2Hallucination flags per run · ranked 6 of 7
19.6%Word error rate · ranked 6 of 7
Summary
PatientNotes produced its prompted note 26s after stop, ranked 5 of 6, and used 28.8 MB of data per consult, ranked 5 of 7. Its word error rate was 19.6%, ranked 6 of 7.
None of PatientNotes' notes were ready to sign as written, ranked 7 of 7. Its editing effort was 37%, ranging from 32% to 44% between runs, and significant note errors averaged 21.6 per run, ranging from 15.0 to 27.0.
Every measure
| Measure | Average | Run range | Rank |
|---|---|---|---|
| Time to a signable noteWait from pressing stop to the graded note, plus review time: how long until the note can be signed if it is reviewed straight away. | 2m 45s | 2m 39s–2m 55s | 6 of 7 |
| Notes ready to sign as writtenShare of notes that captured every checklist fact in full with no hallucination flags, so nothing needed editing. | 0% | 0%–0% | 7 of 7 |
| Editing effort0% means no editing required; 100% means writing the full note by hand. Fixes are weighted by the work they take: partial fact 1, missing fact 2, hallucination flag 3, and +2 when a fact is captured wrongly. | 37% | 32%–44% | 7 of 7 |
| Review time per noteEstimated clinician time to proofread and fix one note (190 wpm reading, 5 s to recall and 40 wpm to type each missing or partial fact, 12 s to verify and delete each hallucination flag, including a wrong fact before it is retyped). | 2m 16s | 2m 6s–2m 30s | 6 of 7 |
| Worst note, any runThe single slowest note to review and fix across every case and run. | 4m 12s | — | 7 of 7 |
| Time saved per noteEstimated time to write the note yourself (5 s to recall each checklist fact, typing it at 40 wpm) minus review time. | 2m 9s | 1m 52s–2m 19s | 6 of 7 |
| Significant note errors per runMissed facts, partial facts and hallucinations that a blind AI rater judged could change diagnosis, treatment, follow-up, safety or the legal record, across 11 cases. Significance is a judgement call; the flat count is in Hallucination flags. | 21.6 | 15.0–27.0 | 7 of 7 |
| Hallucination flags per runEvery claim in the note with no support in the consultation, counted flat: each one counts, however minor. No severity judgement. Across 11 cases. | 9.2 | 8.0–11.0 | 6 of 7 |
| Edit actions per noteMissing items, partial items and hallucination flags a clinician would need to fix, per note. | 5.16 | 4.36–5.90 | 7 of 7 |
| Clinical coverageShare of each case checklist captured by the note (partial counts as half). | 71% | 67%–75% | 7 of 7 |
| Significant transcript errors per runMis-heard words that change clinical meaning, rated blind, across 11 cases. | 32.3 | 30.0–35.0 | 7 of 7 |
| Word error ratePooled word error rate per run against the human reference. Every word counts equally, including fillers. | 19.6% | 19.4%–19.7% | 6 of 7 |
| Stop to prompted noteMedian time from pressing stop to the note written to our instruction, per run. | 26s | 24s–28s | 5 of 6 |
| Data per consultAverage network data per consultation, per run. | 28.8 MB | 28.5 MB–29.0 MB | 5 of 7 |
Definitions and weights are in the methodology. Compare PatientNotes head to head ↗
Consultations
CONSULTATION