Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-1670
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-17T06:54:44.861255Z
status: published
lens: g_space
sector: health
headline: Retrieval-to-abstraction LLM-assisted literature synthesis: Case demonstrations in digital health technology-enabled randomized controlled trial outcome reporting
dek: Systematic reviews of randomized controlled trials (RCTs) support evidence-based clinical advancement. Clinical trial reporting guidelines, including CONSORT, establish common data elements that support systematic synthesis of published trial outcomes. However, the rapid rise of digital health technologies presents new challenges for reporting guidelines, and substantial real-world reporting heterogeneity has been observed. Therefore, machine-assisted evaluation of reporting adherence offers a pathway for more a…
gain_title: An unmodified pretrained LLM performed screening and information extraction for digital health trial literature at F1 scores comparable to human inter-annotator agreement, suggesting faster evidence synthesis.
problem_title: (none)
trace_subject: (none)
gain_reading: An unmodified pretrained LLM performed screening and information extraction for digital health trial literature at F1 scores comparable to human inter-annotator agreement, suggesting faster evidence synthesis.
gain_evidence: Large language model performance was on par with inter-annotator agreement F1 scores (≥0.9) in both screening and information-extraction tasks.
problem_reading: (none)
problem_evidence: (none)
quick_read: Researchers proposed a framework that moves from retrieving articles to abstracting data using a pretrained large language model. It covers article screening, preprocessing of text and tables, and full-text review, without any model modification.

They tested it in two case studies of digital health technology-enabled randomized trials, checking reporting of technology and digital health equity guideline items. The results point to a route for more agile evidence synthesis as digital health trials grow more varied, though the work is shown only through two demonstrations.
limitation: 
tag: Evidence-backed gain
key_points: The framework combines article screening, text and table preprocessing, and full-text review in one pipeline. | It uses an unmodified pretrained large language model rather than a fine-tuned system. | Two case studies assessed adherence to technology reporting items and digital health equity guideline items. | Model F1 scores matched inter-annotator agreement at 0.9 or above for screening and extraction.
rundown: The study starts from the role of systematic reviews of randomized trials in evidence-based clinical advancement. Reporting guidelines such as CONSORT set common data elements, but digital health technologies strain these guidelines, and reporting heterogeneity has been observed in practice.

The authors argue that machine-assisted evaluation of reporting adherence can make information management more agile. Their retrieval-to-abstraction approach is meant to accelerate synthesis across diverse and disruptive digital health technology-enabled trials.

The authors report that the results support the scalability and rigor of accelerated evidence synthesis. They also say the results show why common reporting elements matter across digital health technology-enabled trials.
sources:
- peer_reviewed | PLOS Digital Health | https://doi.org/10.1371/journal.pdig.0001768 | 2026-10-09
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

dec920dbb8ba82188f80be8f94da78191e4be5f06190c8f52486efccea3605c6

previous

0000000000000000000000000000000000000000000000000000000000000000