We re-read every high-impact trial in the top journals

A paper in NEJM or The Lancet has already cleared the most demanding peer review in medicine. That is exactly why it is a fair test. We ran the full Adcurare rigor engine over recent high-impact randomized trials from the leading medical journals, then published every review. Here is what a careful second read turns up, and, more importantly, which papers it flags first.

What we did

We took recent, high-impact randomized controlled trials from the leading general-medical and clinical journals, NEJM, The Lancet, Nature Medicine, The BMJ and peers, and ran each through the same rigor engine the product ships, in full audit mode with nothing skipped. Each paper gets recomputed statistics, protocol-to-publication and claim-to-evidence checks, citation-integrity and reporting-completeness passes, and a 1-to-5-star rating built from itemized deductions. The complete review for every paper is published in the Rigor Showcase, so nothing here is a headline without a record behind it.

The point isn’t how many. It’s which ones.

Almost every trial carries at least one candidate issue somewhere, which is unsurprising: no paper is perfect, and a thorough read of anything long enough will find something worth a second look. If the tool stopped there it would just be alert fatigue. The value is in the ranking. Of 101 trials, the engine rates 3 as critical and pushes them to the front of the queue. Those are the ones a diligence team should read first.

A critical rating is not a verdict of misconduct or a claim that the trial is wrong. It means the engine found a demonstrable, high-severity issue that a qualified reviewer should adjudicate. Decision impact is determined by that reviewer, not asserted by the score.

What tends to surface

Across the set, the findings cluster in a familiar handful of places: statistics that don’t recompute, citation-integrity issues, unshared data or code. The most common specific findings, counted by how many distinct papers show each one:

Why it matters

• Peer review is a filter, not a guarantee. Even the most selective journals publish trials with checkable issues in the numbers, the claims, or the reporting. A second, systematic read is additive to peer review, not redundant with it.
• Triage is the product. Finding something in almost every paper is easy and useless. Reliably surfacing the few that a diligence team should see first is the hard part, and the useful one.
• Every finding is traceable. The showcase publishes the full review for each paper, with each finding tied back to its source text, so the ranking can be checked rather than trusted.