Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-2246
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-25T06:45:28.271458Z
status: published
lens: trace
sector: health
headline: Readability of National Surgical Consent Forms: Artificial Intelligence Generation Versus Simplification
dek: Introduction Consent forms often exceed the reading ability of populations they serve. Large language models may improve accessibility of patient-facing information, but existing research has predominantly assessed de novo generation, with limited benefit. AI-assisted simplification of existing, clinically validated material represents a distinct workflow that remains underexplored. This study compared the readability of all 64 consent forms published by NHS England's Getting it right first time (GIRFT) with mat…
gain_title: Using an AI model to simplify existing, clinically validated NHS consent forms improved their readability compared with the original forms.
problem_title: Forms generated from scratch by AI were harder to read than the originals, and neither AI approach reached the nationally recommended reading age of 11 years or below.
trace_subject: Readability of NHS England GIRFT surgical consent forms under AI generation versus AI simplification
gain_reading: Using an AI model to simplify existing, clinically validated NHS consent forms improved their readability compared with the original forms.
gain_evidence: AI-simplification improved readability
problem_reading: Forms generated from scratch by AI were harder to read than the originals, and neither AI approach reached the nationally recommended reading age of 11 years or below.
problem_evidence: AI-generated forms were harder to read (17.65 ± 0.73 years | Neither approach achieved the recommended reading age.
quick_read: Researchers took all 64 national surgical consent forms published by NHS England's GIRFT programme across six specialties and produced two AI versions of each using Claude Opus 4.8. One was written from scratch and the other simplified from the original, giving 192 documents that were scored on five readability metrics.

The comparison suggests that AI is more useful for editing existing, validated consent material than for drafting it anew. Even the simplified forms fell short of the recommended reading age, and the authors say outputs need objective readability checks and clinical validation before any real-world use.
limitation: Neither AI approach achieved the recommended reading age, and the authors state that outputs need objective readability assessment and clinical validation before implementation.
tag: Dual reading
key_points: The study covered all 64 GIRFT consent forms across six surgical specialties, with matched AI counterparts for each procedure. | Original forms had a mean reading age of 16.39 ± 0.85 years, well above the recommended 11 years or below. | De novo AI-generated forms scored worse, at a mean reading age of 17.65 ± 0.73 years. | AI simplification improved readability relative to the originals, but did not reach the national recommended reading age.
rundown: The authors note that consent forms often exceed the reading ability of the populations they serve. Earlier research on large language models mostly tested de novo generation of patient information and found limited benefit, leaving simplification of existing material underexplored.

Each procedure had two AI counterparts produced with standardised prompts: a de novo form and a simplified version derived from the original. Outcomes included average reading age and grade level, plus Flesch reading ease and word count, and whether forms met the recommended reading age.

The conclusion is that AI may help improve accessibility of existing consent forms, but only with objective readability assessment and clinical validation before implementation.
sources:
- peer_reviewed | ANZ Journal of Surgery | https://doi.org/10.1111/ans.71030 | 2026-10-23
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

dab85f4967062e2ce1b3cf18ae7ed40750b9fc567d8b31edba0f87d7ea5afac3

previous

0000000000000000000000000000000000000000000000000000000000000000